DC Power Distribution - Egrouphub
You are at the eGroupHub wiki! group search main site Site Training | | wiki text format how to organise wikis mediawiki web site

DC Power Distribution

From Egrouphub

Jump to: navigation, search



The goal of this document is to teach people how to best use dc power coming from solar cells and other supplies. Hopefully, by using a few tricks, we can cut the cost of our power systems in half both by using less power and by using power more efficiently. This would be a huge advantage for people paying the price of solar power systems, and I believe that it can very definitely be realised, so please pay attention to this document as it is full of tips and tricks to save on the cost of your system.


What I want to do with this document is to achieve several goals.

  • Help people to understand how they can minimise the cost of their solar power systems. It should be possible to reduce costs of solar power by as much as 50% as compared to standard installations.
  • Provide a technical reference document for dc power usage
  • Provide easy instructions for people who don't want to get into the complexities of dc power
  • Provide real world experience and test results for people who wish to see how the rubber meets the road.
  • Eventually we may want to design open source products to help drive the cost of dc power down.
  • Provide parts lists, schematics and images to help the public to visualise what we are doing.
  • Use the information in this document to engineer and create a solar power guinea pig project.

This document will be a work in progress for quite some time, so not all of these goals will be achieved in early development.

Our Example System

This link leads to a page describing our example system. Our_Example_system


Alternating current refers to the use of a continually changing voltage to deliver power. Typically voltage varies in a soft, smooth sin curve that just looks like a smooth wave like a ripple on water. That voltage in the case of 120 volts ac power peaks at around 170 volts positive and troughs at about -170 volts. The 120 volt rating is what we call an rms or root mean square rating. To get such a rating, we would sample the voltage many times over a period of time. We would take each sample and square its value (to make the values all positive). Next, we would average all of the squared results and finally, we would take the square root of that average value.

It is important to remember that alternating current does not always follow a smooth curve that peaks much higher than its rms value. The most inexpensive power inverters for example tend to output something like a square wave. A 120 volt square wave would not peak at 170 volts but at 120 volts. 120 volts square wave ac would jump at intervals from +120 volts to -120 volts, spending no time in between at other voltages. Such a power source would not be suitable to power all types of devices that can be powered by a smooth sin wave.

Direct current refers to a voltage that remains constant. Choose the voltage and hold it there.

Three phase power is like alternating current, only we use 3 separate waves that are timed differently so that while one voltage is high, the other two may be low. To conduct 3 phase power to a location, at least 3 conductors are required. It is common to use an additional ground or 0 volt wire for safety.

It is important to understand the compatibility issues between these various forms of power. Lets list some of the compatibility information of value.

  • Heating elements don't care if they are powered by ac or dc, or what the waveform is so long as the rms value of the voltage matches their requirements.
  • A switching power supply is a power supply that uses transistors, a small inductor or transformer, and advanced circuitry to convert power from the input to the output. Switching power supplies normally could also care less if they are operating on ac or dc or even how the dc is poled (+- or -+) because before they convert power, they first rectify it to dc internally. At that point, it makes no difference if the incoming power was dc to begin with. The concern for switching power supplies is what range of internal dc they can operate from. Some switching power supplies are very flexible on this issue, some are not, so it is important to understand the power supply in question. Figure a switching power supply runs on an internal bus voltage that is approximately 125% of the rms value of an incoming sin wave, and they will always be rated for an incoming sin wave. The wave peaks higher than that, and the rectifier tries to store the peak of the wave, but the peak does not last long, so the internal dc voltage quickly drops back down from the wave peak, forcing the device to operate at lower voltage almost all of the time.

So a computer power supply running off of 120 volts ac is most likely really running off of about 150 volts dc internally and should be able to consume 150 volts dc up front (+/- 10% or so). Computer power supplies are often switchable between 120 and 240 volts. This switching scheme may impact their rectification process in an unpredictable way, meaning that it is possible for this scheme to be upset.

  • most motors are very sensitive about the nature of the incoming power. On the good side, An ac motor should be able to run off of either a sin or square wave. Some ac motors are actually ac/dc and do not care much what form of power is coming in. The way to recognize these is to look for brushes and wires wound on the rotating part of the motor. If it has wires wound on the rotor, it is almost certainly an ac/dc motor, and you can power it either way, just try to get the voltage right. If a motor has a funny looking bulge on the side somewhere, then it almost certainly not an ac/dc motor, but a motor with a capacitor start. Do not attempt to power it with dc.
  • incandescent (filament) light bulbs could care less what wave form they are running on. Just be sure to get the rms voltage right.
  • many light bulbs are available that use flexible switching power supplies that allow both ac and dc and a wide range of voltage. Good light bulbs are easy to find for most voltages of interest. A good place to look is at a local solar power vendor. 12 volt, 24 volt and 120/240 volt (often rated at something like 90 to 260 volt ac/dc) are available. Unfortunately, finding a good 48 volt light bulb doesn't seem to be an option. There are two good ways to solve this problem. The first and simplest way is to wire two identical 24 volt light bulbs in series with each other into the circuit. This means that the power goes through one of them before it gets to the other one. The bulbs must be identical in order to balance the voltage use so that each one gets 24 volts out of the 48. If we use identical l.e.d. light bulbs for this, then we should never have to worry about having to replace one with a non-identical bulb. However, it would be wise to use a lot of the same bulbs and keep a few extras on hand. Use of non-identical bulbs will result in unequal voltage use and most likely the destruction of at least one of the bulbs.

  • non-switching transformers are usually heavier than their switching counterparts. Old-style wall warts always used copper and steel. These transformers are not tolerant of dc or of differing input voltages or even of different wave forms. They are for standard power only.
  • stereo amplifiers almost universally use two-pole rectification and then run their circuits off of the rectified power. Running a stereo amplifier or any stereo equipment off of dc or square waves is not likely to provide you with much satisfaction. You might get lucky and get some functionality, but it's probably not worth the risk of even plugging the things in.
  • if you have a dc motor and ac power, there is a good chance you can simply run the ac power through a bridge rectifier and convert it to dc for the motor. This is especially true if your ac power is a square wave as the voltage of the wave will match its rms value. A 120 volt square wave will rectify to 120 volts dc. A 120 volt sin wave will rectify at the peaks to 170 volts dc, but the peaks are short and weak, leaving the average rectified voltage under most circumstances to be about 150 volts. To speed control such a dc motor, try rectifying the output of a variac (a variable transformer). DC motor controllers based on triacs and scrs are also available commercially. Bridge rectifiers are cheap, but if they will be passing much current, they need to be heat-sink-greased and bolted to a piece of metal to dissipate heat. Without heat sink grease, such rectifiers will not match their rated current values.
  • to get audio amplification using 12 volt power, automobile amplifiers may be a good choice. Big ones will draw a lot of current (up to 90 amps), so high power auto amplifiers will be problematic while low power ones should not be a major challenge for power distribution.

DC Nominal

Battery banks are often named to produce power at a particular voltage. Automobile batteries are typically named at 12 volts, but in fact, 13.1 or 13.8 is more likely to be seen. In the case of Lead-acid batteries, the named or Nominal voltage is basically about what the batteries put out when they are getting a little low on charge, but still working. They are typically named in increments of 12 volts. Maximum voltages in a lead-acid system with a DC nominal voltage can be considerably higher (up to 20%) than the nominal voltage. A 48 volt system may reach as high as 58.5 during normal usage.

For small carbon or alkaline cells for small devices, the nominal voltage is typically more like the maximum voltage. As the battery is spent, it will typically reduce its output voltage slightly.

Volts, Amps and Watts

Voltage is basically the amount of force pushing on electrons. Amperage refers to the number of electrons that are moving. Watts is a power unit defined as Volts multiplied by amps. It just so happens that a watt is also a joule per second. A joule is a newton of force times a meter of distance pushed. A newton is almost exactly 1/10 the weight of a kilogram of mass on earth. A Kilogram is not a unit of force but rather a unit of mass while a newton is a unit of force.

so Power = Volts * Amps or P=VI and if we know resistance in a circuit,

amps = volts / resistance or I=V/R

We can use these equations to find out information we don't know when working out the details of a system.

Note that R can also be a "reactance" (there are two types, inductance and capacitance) which makes the math a little more complex, but we won't get into that here. Nevertheless, we use this fact to help us build a reactive current limiter for charging batteries with transformers. By adjusting the reactance, or in that case "inductance" of an inductor, we can adjust a current limit to match our system without wasting power. If we use a resistor in stead, the resistor will waste power, which is what we don't want. So a key element to understand here might be that reactances can pass current, but they deliver the power back to the system that goes into them. They store power and return it rather than using it up. Reactances are also only effective for alternating current, and their behavior depends on frequency.

Inversion vs DC Power

Inverting power means converting it from 12 to 60 volts direct current (or more under cases) to 120 or 240 volt ac (or other ac voltages under cases).

The problems with inverting power are:

  • You have to purchase the inverters
  • square wave inverters are relatively cheap, but their output wave form is unusable for many devices (see notes above)
  • sin wave inverters produce a more standard wave form, but they can be very expensive.
  • Inverters will consume around 10% of your power, meaning you have to add 10% more power production capacity up front.
  • In essence, inverting power can just about double the cost of your power because of the combined cost factors. Considering the up-front cost of power installations, this is a real problem, especially for people on a shoe-string.
  • an inverter is "more shit to go wrong". It has to be housed and protected, and is subject to damage by relatively low quantities of emp and various other factors. For example, if someone fails to connect cables to the inverter solidly enough, they could oxidize over time and destroy the equipment or even cause fires. (btw. always make sure all electrical connections are solid and tight. Low current connections are less risk. High current connections (over 2-3 amps) that are not solid and tight are dangerous.

Advantages of inverting power:

  • using the power is easier as most devices are designed to use 120 volt or 240 volt ac.
  • ac power is easy to transform up for long distance distribution.

The problems with using dc power are:

  • The market does not provide as many options for dc powered devices
  • dc power usually refers to low voltage. Low voltage implies high amperage. High amperage imposes heavy costs for using heavier, more expensive cable. Furthermore, heavier cable is more difficult to wire into a structure.
  • The number of dc voltage options available causes problems in consuming dc power because there is a need to choose a specific voltage or set of voltages for use. Fortunately, there are ways to at least partially alleviate this problem by routing multiple voltages into the dwelling. Also, dc-dc converters are available to help solve this problem, and they have become relatively inexpensive.
  • Low voltage power requires more careful attention to the quality of electrical connections.

The advantages of using dc power are:

  • you can eliminate the cost of the inverter and save the 10% or so efficiency that it loses. This can nearly half the cost of the system. (drop it by 30 to 40 percent) This is a huge advantage.
  • Many devices transform ac power back down to dc power anyways before using it. Using dc power, you can potentially bypass those transformers. Also, for switching power supplies, conversion from low voltage dc to low voltage dc is more efficient than converting from medium voltage ac to low voltage dc.
  • low voltage dc power is safer and less regulated. However, there are some things to know about its management in special cases.


There are basically 3 kinds of inverters. Square wave inverters produce voltage that jumps back and forth from a single specific positive voltage to a negative voltage of identical magnitude. Modified sin wave inverters are similar except that they jump a little higher and a little lower. To compensate for the extra voltage, they wait at 0 volts between jumps. Different devices respond differently to different wave forms.

Personally, I only properly trust sin wave inverters, and the reason is that it's too easy to plug something in that isn't compatible and have a problem. These are unfortunately quite expensive. However, we seem to have found a way around that as some chinese companies are now offering inexpensive solutions. PowerJack is the brand we chose. We purchased an 8 KW inverter from them for less than it would cost for a 2 kw inverter made in the states. One might worry about quality and maintenance regarding such a choice. In fact, PowerJack had quite a few really bad reviews initially, but our purchase date came several years after those bad reviews. I set up the inverter in our system and it has run flawlessly for around 3 weeks.

Regarding inverters, it is crucial to understand that the primary problem for inverters is really high current flow spikes. Transistors suck at delivering really high current. Motors for example tend to draw around 10 times their normal current at startup, so this is a real problem for inverters. Here is how I dealt with the inverter problem.

  • I purchased an over-sized inverter 8 kw
  • I made a rule for the system that the largest motor to be started on the inverter is 1 horse power
  • I put a breaker in front of the inverter that will trip at only 5 kilowatts throughput.

I'm still thinking of how to protect the inverter against short circuits. A short circuit situation could potentially ruin the inverter all in one shot. I'm starting to think that the proper solution is to place fuses on the ac output of the inverter that will blow on short circuit. Fuses rated low enough will usually blow before transistors burn, so fuses can protect an inverter from current spikes. Circuit breakers are unlikely to protect the inverter against short circuits. There are small, cheap fuses up to 30 amps available for ac power, so they may be the correct choice. Inverters may have internal short circuit protection, but it may be unwise to rely on that.

The multi-voltage problem

There is a wide assortment of consumer items that run on low voltage dc. Many of these items utilize transformers to transform power from common medium voltage (120/240) lines down to the voltage required by the device. There is a lot of information to consider when dealing with this issue, and being aware of various advantages that are available could be highly valuable. Lets discuss some of the details of power distribution and usage that can fall to your advantage if you properly plan a distribution system.

  • The higher the voltage we use, the less copper is required to conduct the same amount of power. This means that to whatever degree possible, we want to try and keep the voltage high. use of 12 volt power is to be avoided when possible. 24 volt is better. 48 volt is something of a sweet spot because 48 volt is below the 50 volt limit applying to many building codes. 96 volts has an interesting advantage as many switching power supplies designed to run on 120 volts should be able to operate at 96 volts. However, do not trust a dc power supply to do this unless it expressly claims that it can. A switching power supply that runs from 120 volt ac is actually operating internally on something close to 150 volts dc. This means for example that a common 120 volt computer power supply should run stable with an input of 150 volts dc because the power supply just rectifies the incoming power to dc up front to begin with. 150 volts dc may be difficult to achieve using typical dc power producing equipment.
  • It is possible to deliver multiple dc voltages to a dwelling, allowing the evasion of transforming power for many devices. The problem of delivering multiple dc voltages to a dwelling lies in maintaining balance at the source of the power. Since the usage of the various voltages can not be predicted in advance, in order to balance the power input, flexible switching must be built into the power delivery system. Lets make a goal of designing a simple, inexpensive circuit for accomplishing this. In any case, power cables with 3 conductors (including a ground) are common. A 3 conductor cable would allow for example a +24 volt, 0 volt and -24 volt. This would create the opportunity to power both 24 and 48 volt devices from the system.

OK, so using for example a dual-pole 24 volt supply enables both 24 and 48 volt devices. Without any other ado, 12 volt devices could be powered with 24 or 48, to 12 volt dc to dc converters. These can be relatively inexpensive. Try ebay for example. Adding a 12 volt power delivery line has problems. Not only does it require additional balancing tactics at the power production end of things, but it requires very heavy cabling which is expensive, and is likely to impose uncomfortable inefficiencies. It's probably best to dc-dc convert the power to 12 volts at the point of use in stead of producing it up front excepting in the single case that your system is very small and the power does not travel far before being used. This means that simple 12 volt systems are good for small boats, trailors, and automobiles, but impractical for homes. Running enough power for a home at 12 volts would require absolutely unmanagable power cables that we can't afford.

Advantages of a 2-pole system (+- 24 volts)

  • switch 24 volt lights one at a time in stead of being forced to wire them in series (market doesn't offer hardly any 48 volt lights)
  • operate 24 volt dc industrial equipment (relays, contactors, solenoids, solenoid valves signal conditioners, panel meters, controllers etc.) without purchasing power supplies or transforming. Industrial equipment almost never comes in 48 volt flavors. Only 120vac,220vac, 12vdc and 24vdc Not much 12v stuff is used. For an offgrid community that seeks to be independent, ability to use 24 volt industrial electrical straight off the supply will save a lot of money in power supplies and a lot of lost power in conversions.
  • operate two-pole circuit designs without transforming power and paying for the transformers and lost power (such as audio amplifiers that require two-pole power, reversing dc motor drives, etc.)
  • enjoy the efficiency of 48 volt power distribution and inversion
  • the option for simpler, cheaper, more efficient inverter and motor drive designs. It should be possible for example to design a simple d-class high frequency flyback inverter. Having dc power available as 2 poles allows for simpler switching schemes for achieving push/pull or circular activity. For example, a single high frequency flyback transformer/rf filter could be used with two transistor switches to produce 120/220vac pure sin wave
  • Freedom to power 24 or 48 volt motors and pumps, freedom to use 24 volt equipment made for semi trucks without giving up 48 volt efficiency
  • freedom to use 48 volt computer power supplies

The 380 volt idea

Data centers and industrial centers are considering the use of 380 volts DC. Why the choice of 380? Here's my best guess. Take 208 (a commonly used circuit voltage) and multiply that by 1.93. Why 1.93? Because when you rectify (convert to DC using a 12 pulse diode bridge rectifiers) 3 phase AC to DC, what you end up with is 1.93 times the nominal AC voltage. This means that 402 volts DC is basically the rectified version of 3 phase 220 volts. However, if you have done any rectification, you know that when in practical use, the dc voltage output is always somewhat lower than what the math says because the voltage peaks are very short lived and contribute very little current to the system. Drop a little off of 402 volts and 380 would be a reasonable voltage to expect on a capacitor bank (where current is being drawn) that lies behind a rectifier interfacing with 3 phase 208 volts on the other side. So, a computer power supply designed to operate with 3 phase 220 volts will already be capable of processing the 380 volts DC inside. This means that if we put 380 volts DC on the lines in stead of the 208 3 phase, it will rectify back to its self and the computer power supply won't know that you are using DC to power it in stead of AC. Store the DC power in battery banks, and now you can power the battery banks and power supplies with a single large rectifier that draws power from the grid. (and thereby skip some power distribution steps)

Technically, we should be able to drive a typical computer power supply or Compact Fluorescent lamp or other switch-mode device (a device that rectifies the incoming power into capacitors and then uses a high speed switching scheme coupled with an inductor to dc-dc convert the power down to a lower voltage) that runs off of 120 volts with about 150 volts DC, and the device would never know the difference.

One concern for this whole idea is that battery banks output substantially different voltage (as much as 20% difference) between empty and fully charged states. So a battery bank maxing out at 380 volts dc might have a minimum voltage of 310 volts, or a 150 volt battery bank might dip down to 120 volts. This complicates battery charging equipment and imposes flexibility on the power consuming devices. However, these are not particularly difficult problems to solve. Simple inductor schemes can limit current flow, and slight adjustments inside of devices can be used to allow for slightly lower voltage. Chances are good that if you got a 600 watt computer power supply designed for 120 volts AC and drove it with 150 DC, when your batteries dropped to 120, your power supply would still safely produce 500 watts, so just over-sizing the power supplies would be an option for solving the problem. A 600 watt power supply may only be trustable for 400 watts if driven by a 96 volt (nominal) dc battery bank.

more details here:


Storing Power

Power storage is really the most expensive component of using solar or wind power. Single batteries can easily cost $300, and if you use more than 20% of their capacity, then you are going to wear them out quickly, so you have to have a lot of battery just to keep powered up over night if you are using much power.

If you have plenty of money, then store power. If you don't, then simply set yourself up to only use minimal power during the night. Some L.E.D. lights should keep your night usable for study, etc. without costing a fortune. Forget using anything like an electric oven or running a typical refrigerator at night.

It is probably cheaper to generate power at intervals during the night than to try to store a lot of it.

Charge Controllers

http://www.solarseller.com/solar_converters__inc__mppt_charge_controllers___faq.htm external faq on charge controllers.

Open_Charge_Controller project

Charge controllers input power from generation sources and use it to load up batteries. They protect the batteries against overcharging, and usually switch off the load when the battery charge gets low. One of the most important things they do is to dc-dc convert the panel voltage to the battery voltage so that you get more current coming into a battery than what comes out of the solar cells. This raises the efficiency of the system.

There has been a lot of hype about what they call MPPT controllers. MPPT stands for maximum power point tracking. Basically, what that means is that the charge controller optimises its behavior for maximum power draw from the generation device. Unfortunately, that isn't saying a lot from a technical standpoint. After more research I discovered that mppt controllers are using a standard flyback technology to dc-dc convert power from higher voltages down to battery voltage without losing power. The way this works is that when they switch a transistor on, current starts to flow into the battery, but it also has to flow through an inductor to get there. Extra voltage in the circuit causes power to store up into the inductor. When they switch the transistor, the inductor continues to push current because of its stored energy. Using a flyback diode, the current is allowed to flow through the battery and back to the inductor. By the way, this is not a big deal to implement. Because of the cheap and easy dc-dc conversion, mppt controllers can efficiently use voltages that are considerably higher than your battery voltage. This reduces line losses and loss percentages inside of the controller as well.

What is being dubbed "pwm" or pulse width modulated controllers are similar to mppt controllers, only they do not do a dc-dc conversion. This means that extra voltage supplied by the panels can not be converted into extra charging current.

It took me quite some time to properly recognise that the usefulness of mppt controllers depends quite a bit on your setup. The most important clues I found relating to this issue were waiting for me on a spec. sheet for a 70 volt solar panel. I expected the panels maximum power voltage to fall with falling light input, but the curves showed something very different. The maximum power voltage for solar panels seems to remain pretty constant, regardless of how much sunshine is available. I also expected the power output of solar panels to fall off rapidly on both sides of that optimum voltage, but it turns out that there is about a 7% window on both sides of the maximum power voltage where the output power of the panel changes very little. Very little as in about only one percent of output power. What this tells me is that if your panels are matched to your batteries and your cables are short, there is very little value in either mppt or even simple pwm charge control. Simple relay control may perform just as well, or even outperform complex controllers. Remember that this is only true if your batteries are matched to your panels and your cables are short. More details about that below.

If you are charging batteries with a wind mill, a mppt charge controller will be crucial because windmills put out such a wide variance in voltage/current as they operate. Often, wind mills will stall and fail to produce because a power system drags at them too hard. An mppt controller will naturally come to the rescue in this case.

If you examine some of the devices out there, you will notice that some of the mppt controllers are essentially identical in appearance to the cheaper pwm (less intelligent) controllers. The obvious difference is the price. MPPT controllers typically cost about 2.5 times what the pwm controllers cost. So the question is, what do they have inside of them that is so much more expensive since some of the units don't even look any different? My response to that question is that there is probably very little difference aside from some altered processing code and maybe a few extra sensors,a coil of wire that stores tiny puffs of energy for short periods of time, and a healthy flyback diode. Fact is, mppt controllers are pwm in nature. pwm (pulse width modulation) control is capable of being optimised for maximum power point tracking without even hardly changing the hardware. I'm almost completely certain that the extra cost of MPPT controllers is often not well justified. They are charging you 2.5x for a little different code and a few additional parts.

So after checking it out carefully, I realised that because of the high cost of the controllers, MPPT tracking is likely to fail to pay for itself for solar power systems. For starters, the tracking only works when the batteries are low and the charger is operating at full charge rate. All things considered, unless you are grid-tied, this will only happen half of the time at most, so that shoots down half the value of their optimisation. Next problem is well, they claim up to 30% more power. I'm saying that the 30% extra is highly unlikely to show up. More likely you will be getting around 12% extra say 14 to give them the benefit of the doubt. If you pay attention to your design, then mppt controllers may not offer any extra power at all. Take a 60A charge controller at 24 volts with mppt operating only half the time and that difference turns out to be 100 watts which is $100 of panel while the charge controller costs $300 extra. Some of the same controllers can do the 60 amps for 48 volts, so if you use them for 48 volts, they are getting closer to payback as that would benefit you 200 watts for the $300 dollars extra. There are a few 90A MPPT controllers that might make up for their extra MPPT cost. In general, I must assert, unless you are grid-tied and dumping extra power to the grid. MPPT is likely not worth it.

There are some really cheap controllers out there claiming to be mppt controllers (like $150 for a 60A controller). Some reviews have asserted that they are not and couldn't be MPPT well maybe because they are too cheap. After looking at images of their insides and outsides, I noticed that they are very simple and don't offer strong configuration options. It may be difficult to get them to properly match your system. It is my opinion that the reason they are cheap is because they are low quality. Even if they generally do have MPPT functionality, I would not want to trust my system with them. But it may be worth trying one out to see how it actually performs. The complaints I heard against them were hardly scientific measurements, and the image of one I saw very interestingly showed a coil of wire that was smaller than I would have expected which indicates that (only) if it was running at high enough switching frequency, it may in fact be able to get some mppt advantage.

Now, before you throw up your hands regarding mppt controllers, I did find a few that I believe are reasonably priced for their performance. Also keep in mind that charge controllers do more than just optimise power. They also protect your batteries against over/under charge and some of them like the "midnite classic" have special charge modes that can help lengthen the life of your batteries. If a charge controller can add even 20 percent extra lifetime to your batteries, then it will pay for itself that way. Another issue is that solar panels tend to change over time. After 10 years, they may lose 10% of their capacity and after 25, they may lose 20% of it. If you carefully match your batteries to the panels, then as time goes by, the panels may not offer enough voltage to properly feed the batteries. For this reason, I suggest either choosing panels that offer just slightly higher voltage than required, or providing plenty of extra voltage and using an mppt controller.

I will post some links to suitable charge controllers right here for you. Below controller is only 45 amp which is not bad really, and only good for charging 12 or 24 volt batteries.


Its weakness is a maximum load current of 20 amps. So it can discharge only half as fast as it charges. However, we are planning to use our own battery taps anyways, so it may not matter at all. Also, this particular controller has ideal characteristics for cheating, so it could reasonably be used to charge a two-pole 48 volt supply by placing one battery bank between the controllers pv return line and the pv cells while using relays to swap some of the batteries from pole to pole.

You can also look up the pcm 6048 on ebay (search for 60 amp mppt charge controller and it will show up) This is a relatively inexpensive 60a 48 volt mppt charge controller. However, It has some uncomfortable limitations. It can only deliver the sixty amps to a 24 volt or lower battery. a 48 volt battery can only have 30 amps. (max. power of 1500 watts) It also expects you to deliver around 50 volts or more at the input in order to get the 60a, It only will allow 35 amps of input current.

The below listed controller can handle nearly 4 times more power, but is only cost effective for a 48 or higher volt power system. It is highly recommended for its ability to take good care of your batteries.

http://www.wholesalesolar.com/products.folder/controller-folder/MidniteClassicChargeController.html or the below link for the cheaper lite version that requires a computer for setup.


Building a 2-pole power supply will be tricky using a unit like this one, but I believe it is not impossible. We will discuss that again in a moment.

Outback seems to be offering some relatively pleasant options.


Both the midnite and the outback controllers have no control of current outflow from the batteries. You will have to create your own control or rely upon an inverter for that. The midnite controller has ground fault protection for the pv array which is required by NEC. What this means is that if the panel array somehow leaks current to ground, your charge controller will shut off the charging. The outback controller is missing this important function, and requires the use of a ground fault breaker to satisfy NEC. This may sound like points for the midnite controller, but if you are putting together a system with unusual characteristics, the midnite ground fault protection may not work properly or be of use.

There are some open source MPPT controller projects underway. I haven't found one that is adequately developed to recommend. I may develop just such a project here as I continue to develop solar power at egrouphub.

As it goes, we are making some minor sacrifices in order to have a two pole power supply available so that we can use both 24 and 48 volt devices without transforming power. This will add up to a major advantage as we go out in the market looking for dc powered devices to use. On the bad side, it forces us to make some adjustments on the power supply side. These adjustments preclude the use of 48 volt MPPT controllers (unless perhaps we expand to add an additional 48 volt pole so as to have 96 volts dc available). Because of this, what little advantage in MPPT controllers there might have been is lost.

The market pwm controller that is most ideal for this case seems to be the xantrex C60 available here http://www.altestore.com/store/Charge-Controllers/Solar-Charge-Controllers/PWM-Type-Solar-Charge-Controllers/Xantrex-Solar-Charge-Controllers-PWM/C60-Charge-Controller-60A-1224V-C60-Charge-Controller-60A-1224V/p2071/

and here


These controllers may be just as effective as mppt controllers, given that you match your batteries to your panels. Keep in mind that it costs around 1 to 1.5 volts to push a lot of current through a transistorised charge controller. Relays don't have this problem.

cheating the charge controllers

I am aware of three strategies to get more of what we want out of these charge controllers, thereby reducing the cost of our solar power charging systems by as much as 30 to 60 percent.

Trick number one ... achieve two balanced poles by swapping batteries back and forth:

The midnite classic mppt charge controller for example is an excellent value. My primary complaint for it is that to get the value out of it, you have to run it charging a single pole of 48 volts. I want two pole 24 volts. This is close, but no cigar because two separate 24 volt poles can get out of balance and cause failures if charged equally.

solution: configure your batteries so that about half of them are in a static -24/0/24 volt configuration. Take the rest of them and make them float. You can float them by setting up four relays for each battery. two relays connect the battery to the plus pole when on. Two more connect them to the minus pole when on. Now, use a timer circuit or microcontroller to activate the relays. Regularly swap the batteries around so that the ones that used to be on the plus pole move to the minus pole and visa versa. This will keep both power supply poles balanced.

Trick number two ... set up your own relays to let the power out of the batteries for a two pole configuration.

Trick number three ... double or thinkably even triple the power management ability of your mppt: This trick is cool. MPPT controllers are capable of dc-dc converting extra voltage. That is where their efficiency rise mostly comes from. Problem with them is the cost of the units that provide the service. Well, here is how you can nearly halve that cost. (and it should work for pwm controllers as well). It's another battery swapping trick.

Use relays to swap batteries just like the pole swapping trick, only this time, you swap batteries into positions both before and after the charge controller. So the voltage first passes through a battery before it meets your charge controller. This will subtract voltage out of the power and protect the charge controller from over-voltage, allowing you to use a higher voltage pv array output. Next, the current passes into the charge controller and through the dc-dc conversion circuitry and into your main battery bank. Back into the charge controller and back out and potentially into a third battery bank (which also subtracts voltage out of the system so that the charge controller is not exposed to it, and back out before returning to your solar cells/wind generator, whatever. This effectively doubles your ability to manage power. You can configure your solar cells for higher output voltage and cut line losses.

For swapping batteries during charging, if we use spdt relays, we can configure it so that there is a rest configuration where all relays are powered down to save power, and the batteries are configured in a reasonable configuration. This way, no relays will have to drain the batteries all night while they are doing nothing.

Which ones of these am I going to do? None. The reason why is that to do it I have to set up a microcontroller to manage the control of the battery charging. If I'm going to do that, I might as well go the next step and let that microcontroller take care of the whole process exactly how I want it to. That's what I'm going to do. But I included this section just so the people are aware of the possibility.

Value of the Midnite Solar Classic unit

I was particularly impressed with the Midnite solar classic charge controller, so I decided to write up some information describing its value so that the reader can consider. After we realised we can get by without mppt by properly matching batteries to panels, we realised that the added values of this controller still make it an attractive option.

feature added value common/uncommon
80a output common
60a input common
web server ? uncommon
150v input ? uncommon
12,24,48,60,72 volt charging ? 60 and 72v is uncommon but probably not of interest
arc flash protection $100 uncommon
ground fault protection $100 uncommon
mppt function 0-20% value of panels common
multiple charge stages 0-20% value of batteries common
modbus, usb  ? uncommon
battery temp sensor $70 or 5% of battery value usually an add on
Ac to DC Conversion Option: Using the 150 volt model with a rectified (with bridge and capacitor) boost/buck transformer powered by an ungrounded generator should be able to feed this unit as if it were a pv array, for the 200 volt model, no boost/buck transformer required because 200 volts is above the 170 volt rectified peak voltage. However, in both cases, generator/transformer power system must absolutely not be grounded $200-$400 requires relatively high voltage tolerance .. uncommon, must be properly set up by user, makes generator charging depend on charge controller

Cheap MPPT for wind

Considering the cost of mppt for wind turbines, it occurred to me that wind turbines OUTPUTTING ALTERNATING CURRENT could be fit with reasonably accurate mppt charge control quite cheaply. If you visit our Quad_Power_Rectifier page, you will see a circuit drawing on the top of the page. Part of that circuit is a hybrid rectifier that is a cross between a 2-pole rectifier and a bridge rectifier. It consists of the components C1,C2,D2,D3,D4,D5 . The behavior of this circuit can be drastically adjusted by changing the values of the capacitors C1 and C2. The rectifier automatically switches between 2-pole and bridge rectification modes, (and inbetween modes) as the voltage/current capacity of your power source rises. While operating in 2-pole mode, the circuit doubles the voltage output of your power source while halving the final current flow. While operating in Bridge mode, it does not double the voltage output, but rather leaves it untouched. As the current flow rises, the circuit automatically switches itself from 2-pole to bridge.

If you can design your power source to output a peak voltage that is around 15% to 20% higher than the battery voltage, then all you have to do is experiment with the size of the capacitors until your unit yields maximum power. The sizes of the capacitors should be roughly equal to each other. They may need to be in the thousands of microfarads depending on the design, and each capacitor should be able to hold the voltage of the battery/storage unit plus about 20%.

I don't have a special solution like this for turbines outputting dc current.

Switching High Current DC

One of the biggest problems when trying to handle low voltage dc current is switching the high amperages on and off. The only reasonably priced switches on the market are either transistors or relays. Transistors damage your efficiency, so we need relays when reasonably possible, but when you make and break high current through a relay, the energy in the current tends to destroy the relay contacts. There are two solutions to this problem, an expensive one and a cheap one. The expensive one is to get expensive mercury contactors at about $100 each. The cheap one is to use relays from China, or aggressively protect your relays from damage. The way to do this is add some extra but cheap circuitry that stops high current from hammering the relays. You can put (small) capacitors from one side of the switch to the other for example as well as free-wheeling diodes that allow the current to continue to flow in the circuit after a switch is turned off. Just add diodes into your circuit for example that allow current to flow from the minus poles to the plus poles and small capacitors across your relays to disable inductive arcing when the switch contacts open up. An example circuit diagram will be provided in time.

This is a link to a collection of example 24 volt dc high current switches from the automotive industry. The best one is going to be the part number 30M9189. If you look further down the page, you will find relays rated for 80 amps, but when you look at the actual data on the relays, you will discover that when switching a 24 volt load, their maximum current is actually lower than the 60 amp switches. The 60 and 80 are ratings for switching 12 volts, not 24. But if you aggressively protect the switches as outlined above, there should be no problem switching the same current for 24 volts.


Chinese are offering high current relays that can be found on ebay or ali-express. Look for JQX-62F . This line of relays offers dual 80 amp contacts as well as single 120 amp contacts. The contacts are rated for 28vDC, so to break 56 volts, you will need two of them in series, or one on opposite poles of your dc power.

If you need to switch lower quantities of dc current (for a dc pump for example), like in the 10-20 amp range, dc solid state relays are also a good option as single dc solid state relays can handle the full 56 volts.

Making Connections

Electrical connections are extremely important for low voltage dc power systems. The reason they are so important is because these low voltage systems pass a lot of current. Current stresses connections, and faulty connections can create damage and electrical fires. Faulty connections are especially dangerous over long periods of time, so let's mention some pointers about making connections that should be helpful.

  • Generally speaking, use a real crimper, don't crimp a terminal with a pair of pliers or a hammer. The only convincing exception to this that I have seen is an example where someone used copper pipe to make home made terminals. The pipe was relative long and strong, so there was plenty of connection surface area. They banged on it with a hammer, flattened out the end and drilled a hole through it. Keep in mind if you are going to do this, you have to deburr that hole because you can't allow a burr to create an air space that won't conduct. Crimpers are designed to make solid connections. Your other hand tools aren't. If your connection isn't solid enough, it's dangerous.
  • bolts/screws/etc. have to be kept tight. How tight? Well in simple terms, about as tight as you dare tighten something without damaging it. No loosey gooseys cause they are fires waiting to happen.
  • try using grease. I know it sounds ridiculous because grease doesn't conduct, but when carefully consider the real physics of a connection, you will realize that grease is an advantage. When we tighten down a connection, grease squishes out of the way and metal meets metal about the same as when we just have air. Flat metal surfaces aren't really flat by the way. On the micro scale where conduction happens, they are a porous, crystalline fractal. Only the little tiny mountain peaks do the actual connecting. Grease will squish into the pores. Where the metal wants to connect, there will be pressure that pushes it out of the way. The advantage of grease is that it doesn't allow your connection to oxidize. An oxidized passivation layer on top of your connectors resists current flow and causes heat build-up which causes more oxidation, so you can nip this spiral in the bud with grease. What kind of grease? Well, I saw an advertisement once for "dielectric grease" for electrical connections. Wow, I thought. What a con job. Pretty much every nonmetallic material you can lay your hands on can be considered a dielectric. Furthermore, the dielectric property of the grease is nearly completely deactivated (like 99.999%)by an electrical connection because there is no voltage across the grease because the connection is connected. There are even more reasons to realise that calling a grease a "dielectric grease" so you can sell it to an unsuspecting electrician at a wildly inflated price is a complete con job. Don't buy into this bullshit. You can use any reasonably stiff grease. Preferably a high temperature grease, so high-temp baring grease for automobiles is probably what I would use. Grease may also be very valuable for protecting battery terminals against acids and oxidation. Don't waste a cheap opportunity to make your system safer. Use grease on your connections.

Matching Panels To Batteries

Powering batteries directly with panels is not commonly done. The reasons for this are that there is no protection for the batteries against overcharging (and maybe also not for undercharging) that way, and furthermore, to have maximum efficiency, the voltage of the battery and the voltage of the panel need to be carefully matched.

Our discussion here will relate to a panels maximum power voltage. This is the voltage that a panel is best suited to operate at. It can produce lower and higher voltages if our system needs it to, but its power output will then not be at maximum.

Keep in mind that for a given wattage of pv panel, a panel of higher voltage produces lower current. If we are directly charging a battery with a pv panel however, the panel has a current limit so that if we can't use all of the voltage available, it gets wasted because our batteries are storing and returning power at a lower voltage. If we choose a panel of lower voltage, we will get more current from the panel in trade for the lower voltage and that gets us more of the power available from the panel. If however, the panel voltage is too low, it may not be high enough to charge the battery, causing partial to utter failure.

PV panels generally have 4 ratings. Maximum power, open circuit voltage, short circuit current, and the voltage of maximum power. Maximum power is only achieved at the maximum power voltage and at full sun. The open circuit voltage is the voltage produced by the panel when we draw no current (power). Short circuit current means we connect the plus and minus poles of the panel together at full sun and measure the current flow. Open circuit voltages tend to be around 20% higher than maximum power voltages. This means that there is a little bit of head room if the batteries need a little extra voltage to charge. Batteries will not charge at all until the pv voltage rises above the battery voltage.

One might imagine pv panels to have an internal resistance rating. The only problem with that is that p-n junctions are very nonlinear, so you wouldn't be able to assign a specific value to that rating as it would change depending on circumstances. You could create a mathematical model with some details to serve as such a rating, but getting information out of it would require more than just a few calculator button pushes.

If we examine the selection of pv panels, we discover that some of them appear to be designed to charge batteries directly. For a 24 volt system, the voltage of the system at full charge will be around 25.2 volts. The voltage of the system while being charged will be around 28.2 volts. A panel that has a maximum output power at around 29.5 volts will have just a little voltage to spare for line losses and for a 0.4 volt reverse-charge protection diode. However, as the panel ages, this extra voltage may fade, so an additional volt could prove valuable. Also, as the sunshine reduces, the maximum power voltage can drop by as much as 3 volts. I would therefore recommend a panel with maximum power voltage of close to 30.5 volts. There are many such pv panels that appear to be designed to charge a battery without a transistor based charge controller. If we now attempt to charge a battery through a charge controller with the same pv panel, we run into a problem. A transistor based charge controller is going to sap a little voltage off of the top. If it saps too much, it stands a chance of damaging the panels ability to charge the battery because it will require the panel to produce a higher voltage than it is designed for. Fortunately, a panel that optimally produces at 29.5 volts will still be doing a good (just not quite optimal) if the controller takes a volt and it has to produce 30.5 volts.

If we are using an mppt controller, we want to be sure to have a good 5 to 100 extra volts (depending on what the controller allows) coming in off of the pv panels. This ensures that we will always have enough voltage to charge the batteries optimally. To charge a 24 volt battery, we will be looking for a pv panel outputting around 35 volts or more.

If we are using a pwm controller, we don't want a lot of extra voltage because it will be wasted. For a 24 volt system, we want a pv cell that operates optimally at around 32 volts. The reason for the extra voltage over directly charging the panel is that the pwm controller will lose some of the voltage in its switching transistors.

To directly charge a battery, we can use relays to turn the charging on and off. Relays have the advantage that they use very little voltage, so you can use a pv panel that is rated for lower voltage and higher current.

If the panels you have in mind are plus or minus 1 volt (at maximum power) from the optimums I noted here, there is no need for great concern. A 30 volt panel can produce very close to full power at between 28 and 32 volts output.

charge strategyoptimal max power ratingmax power rating range
PWM 32V/64V 31-36V/62-71V
MPPT 0.6 * max controller voltage (going higher can ruin the controller with open circuit voltage)33-0.6*max/70-0.6*max
Relay 31.5V/63V30-35V/60-70V
Wind/Mppt Always use mppt for wind design hybrid rectifier suggested use voltage limiter/clipper to protect mppt controller

Charging Batteries With AC

Should you choose to charge batteries with an ac generator, you are going to have to solve the ac/dc conversion problem. Let's look at some of the options for solving this problem.

option +description cost advantages Disadvantages
done by inverter usually expensive, and implies the use of an expensive inverter, however, do some shopping around and a good option may appear most efficient and easiest option. Inverter may also allow direct feed-through of ac generator power to your building expensive, and dependant upon functionality of the inverter
Special, purpose specific battery charger around $250 (chinese) to $400 per kilowatt independent expensive
switching dc power supply mpja.com offers a 28 volt 150w unit for around $30 -> (approx) $200/kw is isolated and safe such units usually don't tolerate overcurrent. They will "fold back" if too much current is drawn. so a strategy to control the current flow is required (like using a small resistance in the flow such as 0.1 to 0.2 ohm)
rectify ungrounded (floating) generator into an mppt charge controller as cheap as $10/kilowatt as expensive as $35/kilowatt inexpensive and intelligent requires charge controller that accepts 150volts or higher, depends on functionality of charge controller, risky if someone foolishly grounds the generator power. This could destroy a lot of stuff 150 volt charge controller requires approx. 1kw boost-buck transformer (around $120 at automationdirect.com). 200 volt or higher charge controller requires no such transformer. If charge controller fails, boost-buck transformer can be rewired to provide isolated dc power to batteries (see next option). Depends on charge controller.
rectify isolation ac transformer into batteries around $150/kw using automation direct boost-buck transformersinexpensive and safe, potentially independent of charge controller requires current control strategy like wiring an inductance in series with the transformer Requires at least 2 transformers for 2-pole system unless there is a two-pole balancing system in place that switches the power from 24 volt pole to 24 volt pole

OK, so obviously the most interesting options are the last two. So let's talk about how to make them work.

First of all, we must absolutely understand that if we rectify grounded power into a charge controller that already has a ground reference, .. well ... This is a real problem because when we rectify grounded power, the ground reference jumps from plus pole to minus pole of the rectifier and back 60 times a second. This obviously creates a huge conflict with any ground reference already in the system, regardless of where it is placed. The likely result is rapid destruction of multiple expensive system components, fireworks, and magic smoke. Or at bear minimum, tripped breakers. Make absolutely certain that the in-flowing power is floating (not grounded at either line) and place a notice on the generator notifying any future electricians not to ever ground the generator so that no one unknowingly makes this mistake.

The next thing we must absolutely understand is that we can only use an mppt charge controller for gating of higher voltages (like 140 to 170 volts dc) to the batteries. Attempts to use relays or pwm controllers will result in destruction because the relays and pwm controllers do not transform the incoming voltage down to the battery voltage. The transformer will directly couple to the batteries (at low voltage) and instantly pop a breaker as well as likely destroying the charge controller. We can still feed dc directly from a transformer to a battery without an mppt controller, but we will have to carefully match the transformer output to the battery. This means we need a dc voltage of around 29 or 56 volts.

This strategy will require a bridge rectifier. If we are charging through the charge controller, it will also require a sturdy capacitor (several thousand microfarad dc) on the dc side of the rectifier. The capacitor is required to help hold the rectified voltage steady so that the charge controller is not too confused by a dramatically wiggling input voltage. If we are charging batteries directly with low voltage dc, the capacitor is not required. Once this is solved, the generator can pretend that it is a pv array and feed into your system as such. A slow-charge system may be required for the capacitor in order to protect the rectifier diodes from a giant inrush current when first starting up.

OK, so let's talk about the boost-buck transformer issue. First of all, a 120 volt ac power generator rectified will deliver a peak voltage of 170 volts. This voltage is no problem for a charge controller rated at 200 volts or higher. However, it is an absolute no-no for a charge controller rated below 180 volts dc. If we are using a 150 volt charge controller such as the midnite classic however, this problem can be cheaply solved by using a buck-boost transformer.

A buck-boost transformer is basically a bunch of copper wire coiled around a hunk of silicon steel. However, with some smarts, the transformer is designed for flexible use. The coil of wire is cut into four separate pieces. Two of them represent 12 volt coils, two of them represent 120 volt coils. If we put the 120 volt coils in series, the transformer can couple to 240 volts across the two coils. If we put them in parallel, the transformer is ready to operate on 120 volts. In both cases, the in-going power can be extracted from the 12 volt coils. If we put the 12 volt coils in series, they will add up to 24 volts. If we put them in parallel, they will stay at 12 volts. In both cases, the 12/24 volts and the 120/240 volts are isolated. That means that they are not connected together, so they can each have their own separate ground reference or float (remain disconnected from ground altogether).

If we put the 120 volt coils in parallel, then we can connect the 12/24 volt coils in series with them. When we do that, we have created a buck-boost configuration. The low voltage coils will in this case no longer be isolated from the high voltage coils. If our incoming voltage is 120 volts, then we will want to use the 12/24 volt coils in their series configuration of 24 volts. If our incoming voltage is 110 volts, then we will want to use the 12/24 volt coils in parallel. In both cases, if we place the 110/120 volts across the entire series coil configuration, we will be able to tap out 100 volts ac from the location where the low voltage and high voltage coils are connected together. Now if we rectify the 100 volts, the peak output voltage will be just 140 volts, which is compatible with for example the midnite solar classic controller.

The charge controller may then manage the current limiting problem, so that may be solved. However, that would be controller dependent, and if we want to make sure that there is current limiting in place, we will have to put a large inductance into the circuit somewhere. To do this, get a really large nut for a bolt, or a toroidal core, wrap it with a lot of wraps of wire, and place it in the ac line side of the power flow. Measure the current, and if it's too low, then back off some of the windings on the core. If the feed voltage is 115 volts, then we use the 12/24 volt coils in series with each other and with the high voltage coil to drop our voltage down.

About the power rating of buck/boost transformers. Buck/boost transformers are rated for their power output in isolation mode. That means with the low voltage coils disconnected from the high voltage coils. They have to be rated this way because the buck/boost mode power rating will change dramatically depending on the configuration. A buck-boost transformer can transform 5 to 10 times as much power in buck-boost mode as in isolation mode. If we put the low voltage coils in series with the high voltage coils wired for 120 volts (parallel) then we will be able to transform about 5 times the name plate rating of the transformer. That means that a 1kw transformer can transform 5kw for us. If we put the low voltage coils in parallel in stead (12 volt configuration) we will be able to transform 10 times the name plate rating of the transformer.

The buck-boost transformer could also be used to transform 220 volts down to 100 volts by putting all of the coils in series and tapping off of one of the 120 volt coils. In this case however, a 1kw buck boost transformer can only transform 1 kw (kilowatt) of power. This transformer has no option to bring 240 volts down to 100 volts, so 240 volt input power is out of the question for this purpose as is 230 volt power unless we are operating with a 200+ volt charge controller.

So you can see that the buck/boost operation can be done more cheaply if you can get your input power down to 110 volts ac. So examine your generator for options allowing you to adjust the voltage down. In fact, if you can adjust it down to 100 volts, there is no longer a need for a buck-boost transformer at all. Just make sure to leave a notice for later electricians as to how the generator has to be configured.

OK, so let's talk about direct charging the batteries. If we look back at the buck boost transformers, we see that they can operate in isolation mode and output the nameplate rating of power out of the 24 volt coils. This means we have to pay the full price per kilowatt of the transformer to do this operation. Currently at automationdirect.com this price is about $120 per kilowatt. This is still considerably cheaper than $300 to $400 per kilowatt price tag of battery chargers. If we do this, then we can ground the generator however we like because the output and input of the transformer are now isolated.

OK, so let's look at how the voltage issue works out. If we input 120 volts into the input coils we will get 24 volts out. If we input 110 volts into the input coils, we will get only about 22 volts ac out. 24 volt output peak rectifies to 24 * 1.4 which is 34 volts minus 1 volt for the rectifer, so 33. 22 volts rectifies to 30 volts doing the same calculation. Looking at these straight numbers, it looks like the 22 volt output is nearly perfect for charging our 24 volt batteries. However, that turns out to be wrong because the peak voltage of the rectification is held only for about 1 millisecond during the wave form of the power cycle. That's enough time for a peak voltage to crash through transistors of a charge controller, but it isn't enough time to charge the battery. The battery will force the peak voltage down to battery voltage because they will draw a lot of current. When transformers power a heavy load through a rectifier, we can figure the rectified voltage will drive the load at about 1.25 times the ac voltage rating ( - 1 volt for rectification) . That means that before the 22 volt transformer can drive much power, its rectified voltage will be very close to 27 volts. This will only slowly charge a lead-acid battery that is nearly full. It should charge a nearly empty battery only slightly faster. If we only have a small, say 500 watt transformer, this may be a great way to go because we only have 500 watts anyways. If we have larger transformers such as 1 kw, we are going to need the 24/33 volt option which will operate very close to 29 or 30 volts dc, so we need an input voltage of 120 volts ac. In this case however, we may need an inductor to limit the current flow so as not to draw too much current out of the transformer. This could happen easily if the battery is in a low state of charge.

OK, so lets assume we need the output voltage of that transformer to drop by 1 volt while delivering 1 kw of power. We may discover that this happens automatically, but just for the sake of going through the hoops, lets calculate what we have to do to solve this problem. First of all, let's assume that we are going to solve the problem by putting an inductor (coil of wire) in the low voltage side of the circuit? Why the low voltage side? Well, because we won't need much inductance to do that. We'll need a lot more if we do it on the high voltage side. 1 kw at 30 volts is 33 amps of current (safely carried by 8 or 10 guage wire) That means that we need 1/33 of an ohm of reactance to use up the volt. For inductors, R = 2 * pi * frequency * inductance , so for a 60 hz system, .03 = 2 * 3.14 * 60 * L, so L = .00008 henries, or .08 mH. That is quite a small inductance. We could resort to transformer equations to figure out what to do here, but for various reasons, those equations don't work well here. So from personal experience building inductors, I am going to say that we can accomplish that with around 5 feet of wire wrapped in any size of circles. (the resistance of the wire will add to the current limiting) If we want to wrap it around iron, we can probably do it with about 3 feet of wire. (say 8 or 10 guage). If we want to wrap it onto a toroidal piece of iron or a toroid core, it could probably be done with 1 foot of wire or less. This is no big deal. We will still be putting breakers on the transformer to limit the current in case of emergency. We probably won't need any current limiting as this is almost a perfect match, but just in case we do (the transformer pops breakers when turned on) this is the solution. Oh, by the way, make sure you wrap the wire in circles, not oblongs or ellipses. Squeezing the shape flat will defeat the purpose by dramatically reducing the inductance. Also, the wire wrappings need to be close to each other for the same reason. Start with double or triple the amount of wire I just suggested, and back off some windings to allow more current to flow.

Now, there is an additional issue with large transformers, and that is nuisance tripping. We can protect a transformer with a breaker gauged to 125% of the breakers output power (say a 10 amp breaker for a 1 kw unit) on the input, but then the transformer may draw a lot of inrush current and trip the breaker. We can also secure the transformer by putting a 250% breaker on the input and a 125% breaker on the output. In the case of a 1 kw transformer, the output breaker would be 40 amps, and the input breaker would be 20 amps. An additional option to stop nuisance tripping would be to leave the original 125% breaker at the input, and go ahead and put our current limiting on the input side of the transformer. In stead of killing 1 volt, we would have to kill five, and we would have to make it happen with five times less current, so we would need 25 times higher inductance. Fortunately, inductance rises with the square of the wire length, so we would only need around 30 feet of wire, (20 wrapped on iron) and we could use much thinner wire, say 14 gauge. Interestingly, this is going to make the inductor roughly the same size (maybe 30% larger). The advantage of doing this is that we get to kill nuisance tripping at the same time. The disadvantage is that we may not need or want the current limiting, so it may not be a guaranteed strategy for stopping nuisance trips.

OK, finally, if we want to do the current limiting for say a 2 kw transformer, we can cut the length of our inductor wire while raising the wire gauge. Say from 5 feet, 10 gauge down to 4 feet 6 gauge. For a 500 watt transformer, change to 7 feet of 12 gauge.

If you are feeling cool, you might want to work out the details of a toroid core to limit the current. The toroidal core will need to be fairly large as cores go, but it will allow a much shorter quantity of wire to be wound. To make this work, we will probably need about a pound sized core, which will cost around $15 or so. The gain would be in a slight rise of efficiency for the charging system. There would probably not be a reduction in cost.

A two pole power supply will require two such current limited charging power supplies. (one for each pole :-) )

Still we have to have an automated means of starting and stopping the charging of the battery or we will overcharge. I will be designing an arduino based charge controller that will be able to do this, and open sourcing the code and design, but unless you are using a copy of that, you will have to come up with your own solution.


Our basic setup used 8 6 volt batteries from deka rated for 370 amp hours and weighing 113 pounds each. Half of this set of batteries storage capacity (the most that should be used daily) is around 9 KWH. 9 250 watt solar panels produced in the winter between 4 KWH and 12 KWH per day. Since a lot of the power is used without being stored during the day, this tells me that the set of batteries is a little more than is necessary for that many panels. They would be a little less than one might like to have for twice that many panels.



For power storage it is not recommended to use automobile batteries as they are not designed to be deep cycled. It is recommended to use batteries designed to deep cycle and to only drain approximately of their energy storage. This is the best way to assure that a bank of batteries will live for decades in stead of years.

There are various types of batteries even once you have decided to use deep cycle batteries. Probably the most important distinction at this point would be between using batteries that require maintenance and those that are maintenence free. Maintenence batteries require checking up on the water level.

Batteries need to be stored in a location with at least some air circulation as they can create hydrogen gas. The gas can potentially be ignited and cause explosions.


The next most critical issue regarding batteries is to understand that since we have to build series parallel circuits out of batteries, in many cases, the batteries have to match each other in behavior nearly perfectly. We can't put an old battery together with a new one or a small one together with a large one. We purchase batteries together at the same time. They remain together in specific groups for the duration of their lives, and we replace them all at the same time. That's how it works.

There are some slight exceptions to this rule. In the case of a two-pole 24-24 volt system, we will be charging each of the two poles with different timing. The charging of each of the two poles will be managed separately by a separate charge controller. For this reason, a group of batteries representing either of the two poles has to be kept together and always in the same configuration. However, you can piddle with each of the two poles separately. Replace all of the batteries on one pole at a time for example. This will not be a problem because they are acting as a separate storage group from the batteries of the other pole.


Putting batteries into series/parallel circuits is standard behavior. Suppose you want a 24 volt charging pole but you only have 12 volt batteries. So long as the batteries are identical, identically charged and purchased at the same time, there is no problem putting them together in a series circuit. To do this, simply connect the minus pole of one battery to the plus pole of the other battery and presto, you have a 24 volt battery. The 24 volt battery charges and discharges as one unit.

Now suppose that pair of batteries doesn't store enough energy for your taste. You can build a second pair of 12 volt batteries. Connect that pair in parallel with your first pair. To do the parallel connection, connect the plus pole of the first pair to the plus pole of the second pair and the minus pole of the first to minus pole of the second as well. Also, make sure that that pair of batteries is identical to and identically charged to the first pair when you do this operation. Now you have double the storage capacity.

The amount of energy you can store does not depend on how you configure the batteries. Two 12 volt batteries will store the same amount of energy whether they are connected in series or parallel. The difference between configurations will lie in the voltage to current ratio. Parallel configurations will have high current and low voltage. Series configurations will have low current and high voltage. For moving power any reasonable distance, it is important to use as high of voltage as reasonably possible. Moving a lot of power at 12 volts even 100 feet is asking for a serious distribution problem. Much of the power will be lost in the cable before it gets to the point of use. Current flow is what causes power loss. Voltage doesn't cause any power loss for direct current applications and it causes only minimal loss for low voltage alternating current. This is why we are not suggesting use of 12 volt power. To move it efficiently you will have to install gigantic, expensive cables that you can hardly fight into your walls cause they will be fat and stiff.

As time rolls by, we will be putting together some spread sheets online for doing calculations on things like batteries.

Desulfatizing (for lead acid batteries)

One way to help deal with the battery problem is "desulfatizing". What this basically refers to is using an electrical process to break up and re-dissolve sulfate crystals that form on (lead-acid) battery plates and cover over the active surface, therefore deactivating it. Basically, desulfators are designed to repeatedly pulse the batteries with very high currents. The high currents tend to break up the crystals, but running the current continuously will tend to overheat the batteries and draw a lot of power. So, that's why pulses are used. Desulfators may require days or even longer to do their job, but they are a known way to extend battery life. No doubt it is necessary to match a desulfator to a set of batteries for proper performance. Assume that large high current batteries will require large high current desulfators. You can find these items simply by looking them up on the net. Ebay and Amazon are good places to look. Since sulfate crystals are not the only problem that batteries have, desulfatizing will not make your batteries immortal. However, it may offer you years of extra service life in a case of fortune.

equalizing lead acid batteries

The procedure of equalizing lead acid batteries is intended to improve battery life. It may be done once per month to once per year. (or not done at all) To equalize the batteries, charge controllers may charge the batteries to an unusually high voltage. Unfortunately, the equalization voltage may exceed the acceptable voltages of devices on the power system, so the power draw needs to be disconnected from the batteries during equalization. Some equalization voltages for given battery voltages are given below.

  • 6v -> 7.8v
  • 12v -> 15.5v
  • 24v -> 31.0v
  • 36v -> 46.5v
  • 48v -> 62v

These are hard equalization voltages. It may be considered safer both from a hydrogen gas production standpoint and a standpoint of safety for your electrical appliances to drop 5% off of those voltages.

Generating Power

Gasoline and propane generators have a high use of fuel during idle and don't last as long as diesel generators. They tend to be a little cheaper. Continuous generation of power is not advisable for private, small-scale use as this will impose high cost for idling your system both in fuel and in wear on the generator. The best use of generators is to have them fire up during the night and times of minimal power just long enough to fill your battery storage.

Running a generator at full output is something we want to avoid as that results in inefficient power generation.

Reducing Power Consumption of a Refrigerator

Refrigerators are one of the most unpleasant power guzzlers. To solve the refrigeration problem, add extra insulation to the unit and don't open it at night. To additionally solve it, you can encase a refrigerator in a housing where damp cloth or matting of some type is kept and air allowed to circulate through so that the outer surface of the refrigerator is exposed to evaporatively cooled air. It may also be possible to house a refrigerator at an external wall sealed off and exposed to the outside air during winter times. Another option would be to encase the refrigerator in a housing and circulate winter air through the housing.

Single-component power management vs multi-component

There may be price advantages in purchasing single component does-it-all power management devices. However, I do not recommend this since:

  • such systems are not flexible
  • If something fails, it all has to be replaced, and an identical unit may be difficult to find. In other words, since it has multiple parameters defining its function, getting a new unit 10 years later with an identical set of parameters could be difficult to accomplish.

Voltage Regulators and Dc-Dc converters

Many devices will function perfectly on dc power, they only need a new transformer or power supply to deliver their dc power requirements. There are a few different ways of solving this problem.

By using a two 24 volt pole supply (+-24V)first of all, we can directly power both 24 and 48 volt devices so long as they do not require precise voltage inputs.

If a device requires dc power that can not be resolved that way, then we can use two different methods of solving the problem. Keep in mind that almost all devices can tolerate at least a 5% error in their voltage. Some can tolerate a lot more, so it helps to know something about the nature of the device so as to know how exact the supply power needs to be.


Voltage regulators are the cheap solution to this problem. If you have a device that uses a little power say at 18 volts, the cheap way to accomodate this is to purchase an 18 volt voltage regulator chip. These are usually around $1 on the market. The voltage regulator chip essentially dumps the excess voltage by wasting it to heat. It will usually consume a few volts for its own use, so it would not work to power a 23 volt device with a voltage regulator that has an input of 24 volts. However, it probably would work to power anything below 22 volts.

Voltage regulator chips are usually designed to output a specific voltage while accepting a range of input voltages. Some of them have an adjustable output voltage, so it may be useful to keep such items on hand.

The more your output voltage differs from the input power voltage to the regulator, the more power gets wasted, the hotter the voltage regulator gets and the more likely it is to fail on you. For this reason, it is not advisable to feed a 12 volt voltage regulator with a 24 volt power supply. A voltage regulator is best only for small steps down, probably around 2-6 volts.

Voltage regulator chips can generate quite a bit of heat, so if you are going to ask them to dissipate much power, you definitely have to mount the chips to a heat sink. Most voltage regulator chips are three legged square things with a metallic back side. Coat the back side with heat sink grease and mount to a piece of metal.

note ... heat sinking without heat sink grease is an ignoramus's mistake. Do not waste your time attempting to make that idea work. It doesn't hardly work, don't try it.

OK, so before we leave the regulator topic, there is one more interesting option. If your supply voltage is just slightly higher than the voltage you need, you can use a few diodes to consume the last tiny bit of voltage. Diodes consume between 0.4 and 0.7 volts and come with different voltages. If you wanted to knock 24 volts down to 23 for example, the easiest way would actually be to purchase a few diodes capable of passing the amount of current you require. Put the diodes in series with the item and presto, problem solved.

Diodes are super cheap. They can be purchased for as little as pennies. Of course, bigger ones capable of more current cost more, and once you get over about 6 amps, diodes start to require heat sinks and grease.

Dc-Dc converters

Dc-Dc converters are a more efficient but more expensive solution to changing dc voltage. On the good side, they waste less power and can push voltage either up or down. On the bad side, you can not purchase them for $1.

Dc-Dc converters use transistor switching and tiny transformers to transform power from one voltage to another. Like voltage regulators, they will have a single specified output voltage (that under special cases may be adjustable) and a range of accepted input voltages.

If you look around carefully for dc-dc converters, you will discover that they can be purchased cheaply enough. Ebay is a good place to look. At some time, we may open up an online store at egrouphub to supply people with items that will help them to solve problems like this.

Dc-dc converters are more efficient than ac/dc converters, so by using a dc-dc converter to get the voltage you are after, you will save up to 20% of your power, so these things will tend to pay for themselves by reducing the required power of your system.

Getting the polarity right

DC power has a polarity attribute. That means that the electrical current wants to flow in a single specific direction. There is a plus side and a minus side. If we foul this up, we are likely to destroy devices that we attempt to power by replacing their little transformers and such with dc-dc converters and voltage regulators.

It's an easy problem to solve. Just make sure you have a volt ohm meter and use it to read the polarity of a transformers output. Write that down if need be, and say, after clipping a foot (or more) of wire off the end of a transformer so as to preserve its connector and connecting that wire to a dc-dc converter or voltage regulator, check the polarity and reverse it if you need to in order to match the original polarity before soldering the wires.

Switching the polarity is as simple as pulling the wires back apart and swapping them so each wire is connected to the wire it used to not be connected to.

Be sure and solder any connections made ( or at least use good wire nuts or other good connectors) and wrap them so that metal with power on it is not exposed. You need a genuine set of crimpers if you want to use crimped connectors because pliers don't do a good job. If you don't have solid connections, your stuff is going to blinking on and off sooner or later. Also when you wrap a splice between wires, try to taper the wrapping a little so that you don't have a short space of highly inflexible wire transitioning rapidly to flexible wire as this will cause the wire to break inside its insulation at the transition between flexible and inflexible.

Also, if you can get some good rubber splicing tape that is self vulcanizing, you will be miles ahead of using cheap vinyl electrical tape. I think I might even prefer transparent packaging tape over vinyl electrical tape in a lot of cases.

Grounding .. Absolutely Crucial to Understand

In the case of solar power systems, grounding is very important to understand because there are some details and concerns that we did not have before.

Grounding a power system can serve several purposes. There are some dangers that lie in making mistakes in the grounding of our system. We need to take steps to avoid making mistakes of that type. One of the main problems lies in the fact that for ac power systems, system grounding is a commonly understood issue. AC power lines swing equally above and below ground, and cross over the ground voltage twice every time the voltage cycles up and down. Ground and neutral are kept basically identical, and they both represent the 0 volt reference. There is no such standard for dc power distributions. The entire power system may float with no reference at all to ground, or it may be grounded at low voltage or high voltage, or even in the center of two pairs of poles.

I personally advise against floating the system without reference to ground because a floating system could unexpectedly be referenced to ground without any obvious symptom. One might then incorrectly feel safe making a momentary reference to ground only to discover the bad way (by something shorting out and exploding in your face or being ruined) that an unexpected reference is already there.

Common charge controllers seem to assume grounding at system low voltage. Common charge controllers are not designed for two-pole dc power systems which I deeply regret. The flexibility of a two pole system is a major asset to the system. Common charge controllers often have means of allowing use with system grounds other than at low voltage, but sacrifices have to be made in order to accomplish this. Take for example the Midnite Classic controller which we will be using to build our reference system. If we ground other than at low voltage, we have to give up the units ground fault protection and re-implement that. A problem also arises with the ethernet and usb connections to the unit. To safely use them, we will need to use isolators for both connections because they are both referenced at low voltage. The manufacturers of the charge controller assumed that low voltage and ground are identical in their circuit design, and that probably simplified their design, but it caused a problem that can be unpleasant to have to deal with.

There won't typically be grounding issues using single devices. Potential problems arise when we start connecting two devices together. Two separate devices may have different ideas about what ground means. If one device assumes ground is at low voltage and connects ground to low voltage internally, and another one assumes ground is at high voltage and connects ground to high voltage internally, then inevitably, one of them is going to have a case at an active voltage. This may not be much threat for a 24 volt item, but if we then connect the two items together in some way, we directly short power to ground and are exposed to the ensuing fireworks.

Any well engineered dc device will not connect its case to either low voltage or high voltage, but rather only to a job-specific ground line. For safety, such a device may not make assumptions about the relative voltage of ground and power. The dc computer power supplies I have were done properly as well as I can tell. There is no reference from ground to power or from ground to any other voltage in the system. All the ground line does is drain any stray voltage off of the case to ground so that no one gets shocked. This is the official job of the ground line, and in the case of dc systems it should do nothing else. Devices should not define any relationship to ground other than that. Internal circuitry should remain completely disconnected from the ground line.

To double check any device for this property, you may use a volt/ohm meter and measure resistance from ground to power inputs. If the resistance is less than 100 thousand ohms or so, then something is not right.

To ensure maximal compatibility with market devices, pinning low voltage to ground will doubtless have the best success. If we do this, then for any device that we connect to the system, it will not be a threat for low voltage and ground to be connected together inside the device. We will have to reject any device that connects high voltage to ground unless it can be reliably kept in isolation from people and from other devices.

Ground for DIY DC Audio

Below we have a section on DIY audio amplifiers. Audio amplifiers can benefit greatly from using a two pole power supply. Such an audio amplifier will assume that the center of the two poles is the 0 volt reference. This imposes a conflict with the notion of pinning low voltage to ground. Audio equipment uses grounding extensively to protect signals from stray voltages in the environment. Should we connect such an amplifier to a dc powered signal source that assumes a low voltage ground, we could short power to ground accidentally. For this reason, in our DIY audio project, we will couple the input signals to the amplifier using capacitors that will block any flow of DC voltage. All signal lines (both signal and 0 volt reference) will need a 1-2 microfarad nonpolar capacitor capable of 50 to 100 volts. We will also fuse the signal ground line in case the capacitors fail so that we will have no risk of having two pieces of audio equipment disagreeing on their ground voltages and causing a problem. There will be a remaining issue as all output lines would rest at say 24 volts when using a low voltage ground system. This is not particularly risky, but under circumstances it could cause a problem, so appropriate warnings will have to be made.

For further protection on the DIY audio project, we will not connect the case to the center of the two input power poles. In stead, we will only connect it to ground. There is a final threat of sudden flow of charge out of the coupling capacitors as we connect the amplifier to something else overvolting the amplifier input. To solve this, we will place a pair of zener diodes across the amplifier inputs that will limit the voltage on the inputs. The combination of these tactics will allow the relative voltage of the circuitry inside the amplifier to be safely referenced to ground any way the user chooses.

These same grounding tactics will also work when using amplifier boards that require only one power supply pole.

Sparkless Plugs

Word has it that some organizations, including Taiwan's Rong Feng Industrial Co. are developing solutions for sparkless plugs. Sounds like a strange concept, but DC power has the problem that disconnecting flowing power creates much more electrical arcing through the air, and hence much more arc damage to the metal components of the plug than does AC. One of the strategies is for a disconnection data pin to alert a device that it is being unplugged before it actually disconnects so that it can intentionally stop drawing power before the disconnection happens.

Cable Sizes

Cable sizes may be defined by local codes. This section will discuss the issue.

As stated earlier, the lower the voltage we use, the heavier the cables we have to use, (and the less efficient power conversion is).

Power is equal to voltage times amperage. Amperage causes power loss and cable heating while voltage doesn't, so we try to use a lot of voltage and only a little amperage. To move a thousand watts at 120 volts, we only need some 7.5 amps. To move the same power at 12 volts, we need 75 amps. In the first case, we can get by on a cheap and cheesy 16 guage wire. In the second case, we will need something more like 4 guage, and it will be at least ten times more money and way harder to install. Furthermore, connectors will have to be bigger and more expensive. This is why we do not recommend 12 volts. Moving only to 24 volts cuts this cabling problem in half. Moving to 48 volts cuts into fourths, so we can still use a strong but not disastrously expensive or difficult to deal with 12 guage cable to deliver the 1 kw at 48 volts.

Here are some wire guage to current rating charts


I like the next one better


The amount of current passable by a wire of a given thickness is a debatable topic. Cables have different characteristics, and ambient temperatures vary. Enclosing a lot of cables in a conduit causes cables to heat other cables, so conduits require derating the current rating or ampacity of cables.

Standard, general purpose current ratings are:

18 guage: 10 amps 16 guage: 12 amps 14 guage: 15 amps 12 guage: 20 amps 10 guage: 30 amps

Two 20 watt LED light bulbs make a lot of light and draw only 40 watts. Using Power = Voltage * Current On a 48 volt system, that requires just under 0.9 amps. So an 18 guage cable could power 11 sets of two rather bright 20 watt led bulbs and light eleven rooms. However, it is not a bad idea to have extra current capacity as this enables unexpected use and decreases power loss of transmission. So I would probably use a 16 guage cable for this, or two sets of 18 guage cable sent to different halves of the building.

Now, if we use incandescent lamps in stead, we will likely draw five times more power and require five times more current. You can see how this requires a lot more expense in cabling as well as up front power availability. Using LED lighting basically pays for itself right up front by diminishing the power and distribution requirement of your system.


For cables traveling underground use plastic conduits to protect the wire. For cables traveling above-ground use metal. For cables inside a building, normally no conduits are required unless the cables are traveling inside of living space.


switches used for DC power products need to be rated for direct current. This pretty much precludes the use of standard light switches. You need to use t-rated switches like those listed here: http://www.sasanelectricals.com/cowisptactos.html or here: http://www.thesolarbiz.com/DC-Rated-On-Off-Wall-Switch-15Amp-UL-Light-Switch#gsc.tab=0 or you can go to a site like mouser.com, look up rocker or toggle switches and filter out anything that is not DC rated.

DC rated switches are much more of a problem than ac rated switches. T-rated switches may cost anywhere from $7 to $25 each. Reasonable rocker switches can be found for as little as $2.50, but mounting them may be a small challenge. In the real world, it would be easy to get a transistor to switch light bulbs, but code requires light switches to be listed, and I don't think the transistor idea works for that. So if you will not be scrapping with an inspector, you may consider something like the following circuit diagram to implement a dc rated light switch .. you can use a regular light switch for the switch.


To make this circuit work for 96vdc, double the high side resistor, i.e. 50k->100k, 100k->200k. Also, raise the transistor voltage capacity to 200vdc.


Don't use the same type of outlets for your ac power as for your dc power. This is too likely to cause trouble. You will plug a 48 volt dc item into a 120 or 240 volt power source by accident and that will be all she wrote.


Just so that we understand a short and simple issue. Plastic coatings for wire/cable are usually not designed to live outside. Ultraviolet light degrades the wire insulation and creates risk. Specially designed pv wire/cable is more expensive but it is designed to resist ultraviolet light. It may also be designed for "direct burial". This means that you can bury it without having to pull it through a conduit first.

While we are on the topic, let's mention that underground cables need plastic conduits. Above ground cables outside need metal conduits.

Breakers and Fuses

When running power cables in a house, we usually use a single breaker box. This forces us to run long cables throughout the house. If the cables are not extremely expensive, then this is not a big deal, but when we start using low voltage dc, our power cables get fatter and more expensive. You can begin to imagine how it would be important to minimise the length of the cable that we use. We can do this by powering two or three distribution or breaker boxes and running shorter runs to our various rooms.

That having been said, there is some crucial information to be understood regarding dc breakers.

  • Do not try using ac breakers for dc current. AC breakers rely on the natural 0-crossing of ac power to break the arc. This zero crossing doesn't exist for DC power.
  • DC breakers, excluding very expensive ones, are one way breakers only. That means that you have to define the current flow direction and it has to always be in that direction. This cuts out the idea of having a two-way current flow path from a charge control center to a battery bank. We have to think of dc charge control as a one way river for this reason. Current flows from the pv cells (or other source) through a charge controller to the batteries and then out a different set of conductors to be used. It's possible to do a little shortcutting by perhaps connecting some breakers together at the power panel, but the unidirectional breaker issue represents a genuine concern that has to be dealt with.
  • DC is more difficult to break, and many fuses are not rated for dc at all. It is something of an understood issue that glass agc or 3G fuses rated only for 250vac are used for systems up to 32vdc. Automobile fuses may also be rated for up to 32 volts dc. If we want to go higher than that (and we do), our options become limited. We can purchase very expensive fuses, or we can use a small number of 3G options. 3G is a size specifier for our standard glass cylinder fuses that are about 1/4 inch in diameter and around 1.25 inches long (approximate). 3G fuses are cheap, but limited in their capacity. Fortunately, there are some things that can be done to maximise the capacity of 3G fuses. Swapping glass for ceramic and fiberglass cylinders as well as filling the fuses with silica are examples of what can be done. Cooper Bussman makes the MDA-30-R and the GBB-30-R which are rated at 125vdc and are a perfect match for 10 guage pv wire. They are available at alliedelec.com. There is also a cooper bussman 40 amp 100vdc 3G fuse kicking around, but it is expensive and hard to find.
  • If your conductors are carrying more than 30 amps, You should (though the likelihood of having a related problem is quite small) fuse your pv wires before you connect them up to a higher current system because if your pv panel shorts out somehow, the entire arrays current could rush through that pv wire and burn it. Most pv wires are 10 guage and rated for 30 amp. MC4 branch connectors which congeal up to 4 panels in parallel are 25 to 30 amp units. The pv wires on the panels should be 10 guage in order to be compatible with the mc4 branch connector current rating. If not, you may have to fuse the branch connector at a lower current. The basic idea is you can congeal multiple branch-connected fused sets of panels to an even larger current flow. That larger current flow will have to have its own breaker.
  • One might think that if he has a breaker on the line flowing in to a charge controller that the current inflow is protected. However, if your charge controller has a problem, it could short out the battery bank from the other side. This could destroy the controller, destroy the batteries and even start a life threatening fire. For this reason, there has to be a breaker on the battery side of the charge controller. It is less important but may still be meaningful to have a breaker on the pv side of the charge controller. Code will usually require this. However, code may allow for omitting the breaker if the cable can pass something like 125% of the panels short circuit current. In the case, it would of course be pointless to have a breaker.
  • single pole systems only require one breaker for each meaningful current flow. The breaker belongs in the line that is not grounded. Dual pole or nongrounded systems require a breaker in both or all nongrounded lines. These breakers need to be 2-pole breakers that automatically trip both breakers at the same time. In order to do this, two breakers are connected permanently together, and they usually have a small metal rod connecting the handles as well as a trip propogation pin near the center of the breaker that lets all the breakers know that one of them has tripped. 2-pole dc breakers for two pole systems need to require that one of the breakers is designed for current to flow one way and the other is designed for current to flow the other way, so pay careful attention to this issue. This is different from ac 2-pole breakers in that the two breakers are actually different from one another.
  • panels have a maximum fuse value because of the supposed possibility of current that is moving through the bypass diodes over-reaching the capacity of the diodes (a very unlikely event) because of this, if panels are ganged in parallel, each panel or series group of panel has to have its own fuse to meet code. One way to solve this is to solder a fuse holder into the little components box on the panel, or just on one of a set of panels in series.
  • dc rated fuses can be hard to come by. However, reasonably cheap ones are available at alliedelec.com. We have two such models mentioned in our parts list. http://egrouphub.com/collab/node/136 When switching to different current values, note carefully that some fuses in the same series may be dc rated while others are not. For this reason, our 30 amp and 15 amp fuses are actually from different series.
  • the dc rated fuses we have listed will not handle more than about 90 volt mppt or 100 volt open circuit voltage. Higher voltages require much more expensive fuses, so that is a good reason to stay with 90/100 volt pv array strings.

Night Load Diminishing

One of the hardest problems for dc power distribution is the cost of battery storage. Because of this, we have brought together a set of ideas to help diminish the night time power usage. Load diminishing devices or devices that have their own energy storage abilities can be helpful.

  • we will be offering load diminishing light bulbs that will consume considerably less (as little as 1/5) power and operate more efficiently when your batteries begin to get down.
  • It is especially important to consider developing load diminishing refrigerators that store cold and/or operate at higher temperatures when batteries get low. Refrigerators that freeze ice during the day and allow the ice to thaw at night are an example of the right idea.


For 12 and 24 volt systems, lighting is easy to solve. There are 12 and 24 volt DC bulbs for lovers of both LED and CFLs. Be aware that some ebay led light bulbs rated for 24 volts really max out at 24 volts, which puts 24 volt lead acid batteries off of the menu because they will deliver up to 28.5 volts. For 48 volt systems, so far, the solution set is more complicated. Because 48 volt bulbs are very scarce and expensive, there are basically four options to solve this problem.

  • use a two pole power supply and use 24 volt lights (highest quality, safest, most flexible)
  • use pairs of identical 24v lights in series (lowest cost)
  • Use (dimmable) 120 vac compact flourescents. Nondimmable cfls may or may not work. Results are mixed. We tested a cheap bulb that lit up for about 10 minutes and a more expensive one that seemed to hold its own. In the end, you will have to use twice as many bulbs/fixtures because these lights will only put about half of their light. (easiest)
  • Make your own 48 bulbs out of LED components. This seems to be the best solution ... see below

Our experience with using 24 volt lights for 48v dc power systems is not good so far. We purchased 24v LED lights on ebay, expecting them to behave for us. Unfortunately, after looking them over carefully, 24v was the maximum rating. 24 volt systems will rise to as high as 28.5, even 29 volts during operation. Putting two 24 volt LED lights in series in a 48 volt circuit caused them to burn out. After checking out the bulbs, I discovered that the transistors in the chip that drives the power to the L.E.D.s had shorted through. The chip had been over-volted. Further investigation into light bulb voltage revealed that many LED bulbs are apparently designed to be driven by regulated dc power supplies and not by lead acid storage. To deal with this problem, continue reading.

After examining the configuration of the 13 watt corn cob bulbs, I discovered several things.

  • the L.E.D.s were wired in 8 sets of 6 and 1 set of 12 L.E.Ds.
  • those sets of L.E.D.s were further divided into separate sets of 3 L.E.D.s in series, each set of 3 being wired in parallel with the other sets of 3
  • The sets of 3 were effectively all wired in parallel with each other
  • the sets of 3 began conducting and lighting at 7.5 volts
  • the sets of 3 were effectively at full power at around 10.0 volts

After looking this data over, I was able to arrive at an easy solution to operate the burned out 24 volt 13 watt bulb with a 48 volt battery and maintain reasonable efficiency (as good as the original efficiency could have been expected to be). I was unable to resolve a similar solution for a 24 volt battery without severely damaging efficiency.

The solution looks something like this:



In other words, 4 of the 8 sets of six LEDs run in series with each other. The other 4 sets of 6 also are wired in series with each other. The two sets of 4 sets run in parallel with each other, and in series with the last set of 12. This creates a maximum voltage usage of about 50 volts. Putting a 2 watt 33 ohm resistor and diode in series with the whole circuit properly limits the current, and replaces the original drive circuit. This circuit will only operate when properly poled, I suggest using the center of the standard light socket as the plus pole. This circuit draws maximum power at around 58 to 59 battery volts, and less power as the batteries discharge, helping to maintain battery life. This circuit should also be slightly more efficient than the original bulb running at 12 volts, and maybe about the same efficiency as the bulb running at 24 volts.

I highly recommend purchasing the 13 watt, 24 volt corn cob bulbs from ebay and switching them over (re-soldering the wires, removing the drive circuit and replacing it with a 33 ohm 2 watt resistor and a diode) to run in this configuration. This creates really the ideal 48 volt light bulb. To do this, just pop the ring off of the end of the corn cob bulb and access the wires and drive circuit inside the bulb. I'm considering getting some people in a development nation to do this little job at a minimum cost. Direct driving the bulbs with dc will save power and help to avoid over-driving an inverter while adding independence from the inverter. This bulb should basically last forever, so once solved, there would be no need to replace it.

This circuit does not work for other similar versions of the corn cob light bulb, and is particularly designed for the 5050 LEDs.

We offer 48 volt dc circuit corrected corn cob LED light bulbs based on the 5730 LED with approx. 20 watt power rating and approx. 2400 lumens (equivalent to a 170 watt incandescent bulb) for $26/piece or $23/piece in larger quantities. The lights are specially designed for 48 volt lead-acid systems and will be bright when your system has full batteries and consume less energy (be less bright) when the batteries begin to wear down. This will help you to reduce your required power storage capacity.


Fabulously, there is another option. There are many 12 volt LED light strips and bulbs to be found on eBay/Amazon and other places. Some of them may use chips and transistors to process power. I do not recommend them. Others use resistors to process the power. Wire 5 identical such lights in series to handle 48 volt dc lead-acid power. They will shine brightly at battery high voltage (58 volts) and very dimly when the battery gets low (48 volts). The units will operate at battery high voltage with just a little under the rated power consumption. However, this promises you that if/when your system goes into battery balancing mode (around 60 volts), the units will not burn out.

12v LED lighting strips can be cut at different lengths while still only requiring 12 volts. If you cut such lighting strips, make sure that all five of your lights are the same length.

Another possible solution is to do a two-pole power supply. It will distribute power nearly as efficiently as a 48 volt supply, but you will have the freedom to switch single 24 volt bulbs (if you can find 24 volt bulbs that accept up to 28 or 29 volts) and drive them as they were intended to be driven. If we compare prices between 2-pole and 1-pole, it will cost about $5 for two cfls plus an extra fixture (maybe $5 to $10) that replace a $15 LED lamp. So the cost in bulbs will be similar. The LED system will pull out in front because it will then use about 10% less power (or produce that much more light) and the bulbs will not need to be replaced. However the cost of the two pole system is in the up-front cost. The two pole system will cost an extra $300 to $500 up front.

We experimented trying to run 120 VAC compact flourescent lights on 48 volts DC. Our success rate was rather dubious. Large cfls normally would not work at all(would not start). Even the dimmable ones would not start on 48 vdc. Smaller bulbs would often start, but most of them burned out quickly. There were a few that lasted a few days. Best hope might be small dimmable bulbs (we haven't tried that yet, and not more than 13 watts), forget the big ones. Unfortunately, even when these lights worked, they only put out about half as much light as they did on normal voltage. My final conclusion on this idea was to let it go because of its failure to perform consistently, and because of the large number of bulbs you would have to use and then wonder seriously about their reliability and ease of replacability.

Topping off the charts might be the use of LED light strings that you can purchase on ebay and perhaps elsewhere. These LED strings are purchased in reels of LEDs, typically 300 LEDs per reel. They are a very interesting option because with a little bit of work, they could be configured to run on a lot of different voltages. The diodes are actually configured into sets of 3, like the bulbs mentioned above. They are wired in a series/parallel circuit where all of the sets of 3 are wired in parallel with each other. You can cut the diode reels every three diodes and wind up with a set of three diodes and a current limiting resistor. Each set of 3 diodes/resistor is designed to operate on 12 volts. However, once again, be aware that they are designed to run from a regulated power supply and not from a lead-acid storage system. If you over-volt the 12 volt string, it will burn out in short order. Unfortunately, trying to run them on 12,24, or 36 volt lead acid storage runs square into this problem, and the only way to solve it would be to size and include your own additional current limiting resistors. However, for a 48 volt system, once again, the stars line up, and 5 sets of 3 diodes in series will safely operate on a 48 volt lead acid storage system. Choose your length of diode string (some number of sets of 3 diodes) make 5 strings of the identical length (must be identical length) and wire them in series with each other. They will shine brightly when the system is at maximum voltage and dimly when the system is at minimum voltage, helping you to use your power when you have it and not use it when you don't have it. The LED light strings are considerably cheaper than the corn cob bulbs and will require less messing around. However, mounting them in your space will be more problematic. You will need to be certain to arrange them in a way that they won't be disturbed. These LEDs also come in two different sizes and colors. Even multi-color-channel strings are available.

Computer Power Supplies

Transdev electronics (shenzhen China) makes 48 volt dc input computer power supplies. Currently, you can find these on ebay. DC/DC conversion from 48 down to 5v/12v is more efficient than doing it from higher voltages, so these power supplies will save power. They should be expected to save 5 to 7 percent power, so at 300 watts, 6 % is 18 watts. This would save 40 to 70 dollars of power installation, so the dc power supplies will just about pay for themselves in reduced up-front power installation cost.


There is probably no need to consider the first item, but be aware that boards running off of different power supplies have the potential to have grounding problems. Sending an audio signal from one board to another can run into disagreements as to what the signal ground is. If this seems like it might happen to you, then put a single 1 microfarad bipolar capacitor of 50 volts or higher voltage capacity in series with both signal lines. This will knock any dc error out of the signal pathway.

If you plan to use one of these high power boards, you will want to have a guess at their current draw. That 600 watt board might draw as much as 15 amps from a 48 volt power supply. The 600 watt board that runs off of a single 24 volt pole might draw as much as 30 amps. So if you want to run something like that, you are going to have to run at least 10 guage cable to deliver the power. At least 8 guage would be recommended for the 24 volt unit.

If your 48 volt supply turns out to really be 51 volts, you will want to put a few high current diodes in series with your amplifier to shave the last volt off. A high current bridge rectifier on a heat sink from radio shack will shave about 1 volt off and get you down to the recommended 50 volts.

Specific boards have specific characteristics. Be sure to read up on the characteristics of your particular board before using it.

DIY Audio Amplifier


Very Important Note: see the grounding section for notes on how to safely ground a DC DIY audio system

It just so happens that a two pole 24 volt power supply is perfect to power an audio amplifier of reasonable power. Normally, an audio amplifier has to transform power down to a voltage similar to this, losing power in the process. The two pole power supply completely bypasses what is likely the most expensive process in home audio amplification.

Simply search around for amplifier circuit boards that accept a two-pole power supply of +- 24 volts up to +- 30 volts and presto, all you need is the board, no need for the expensive add-on power supply.

Now the next thing to keep in mind is that it is ideal to have a board that can power low resistance loudspeakers. This is because 24 volts is just a little bit low for a powerful amplifier, but it can deliver a lot of power if the loudspeakers are low resistance. The amplifier board, of course, must be able to pass the required current flow in order to power low resistance loud speakers.

As an example, many subwoofers have dual 4 ohm voice coils. If we arrange those coils in parallel, then the result is only 2 ohms. at 2 ohms, a standard amplifier circuit powered by 24 volt poles will be able to drive 160 watts of power.

Unfortunately, a standard amplifier circuit board will then waste an additional 160 watts while doing that, so the next concern we want to bring up is the option of class D amplifiers.

Class D amplifiers are nearly twice as efficient as standard audio amplifiers. They deliver almost all of the power going in to them to the loudspeaker. They accomplish this by drastically altering their power switching process to a high tech high frequency pulse width modulation scheme. Because of their efficiency, they don't require huge heat sinks, etc. They can be relatively small for the amount of power they deliver. This is very desirable on a home power system where you want to conserve power as much as possible.

Keep in mind, a class D amplifier connected to your dc power supply also saves power when not in use. A standard ac/dc rectifying amplifier has its transformer connected to your power line all the time. The transformer naturally wastes power while it is doing nothing, and the transistor biasing wastes even more power. Running a class D amplifier directly from your 48 volt dc solar power will cost you only about 1/4 the power or less in the end compared to a standard amplifier.

Below are some current options for ideal amplifier boards

this board (above) is unusual in that it is literally designed for a single pole 20 to 50 volt power supply. This is an option. On the other hand, a 48 volt solar power system is going to really produce around 56 volts maximum. This is too high for the 50 volt board. However, the board could still be driven by a single 24 volt pole (which would typically be around 27 volts). Unfortunately, this would contribute to the imbalance of a two pole 48 volt power system. It can drive from 2 ohm to 8 ohm load with up to 300 watts of power per channel in stereo. The price is a little high at $185, but it has ideal characteristics for our 48 volt system. It is a class D amplifier board, so it will only use about half the power or less of a standard amplifier circuit. This amplifier board uses only 3 watts of power at idle. A comparable high power home audio amplifier will draw at least 20 watts at idle, more likely 30 to 50.

If that is too expensive for you,

This board has pretty much identical characteristics at a cost of only $89. By the way, to get this kind of power in an amplifier for that price is a dream come true for audio buffs. Remember, half the cost of a power amplifier is the power supply, which you do not have to pay because you already have the dc power.

This board (above) would work well if you are dreaming of a six channel home audio system. It can not use a full 48 volts dc, so you would just power it off of one of your 24 volt poles. It wants 14 to 39 volts dc input. It costs only $60 wow, and each channel is rated at 100 watts .. That will be a heavy current draw on your single 24 volt pole if you ever max it out. It is also class D, so all the advantages are there.

here is a super cheap one with two 50 watt channels. That is enough power for most peoples taste. The cost plus shipping for this guy is just $45. It uses only a single 24 volt power input pole.

This item is superior to the above option most likely. Double the power and just about the same price. Also runs off of a single 24 volt power pole.

look here for more class D amplifier options. Some of them request two pole power supplies.

There are a lot of great class D subwoofer amplifiers available for car audio, but they pretty much max out at 14 volt input voltage so we can't use them. You can always look around for more options by searching for "class D amplifier board"

  • Also, be sure and search ebay for class d amplifier board as these things are showing up even cheaper on ebay. There are some options that use something like 14 to 28 volt ac power with 3 pin inputs. Don't let that fool you. These boards are likely the most optimal and least expensive of all. Unfortunately, I can't link to them because the auctions are always closing. A board that accepts 28 volt ac power will almost always rectify the 28 volts to two pole 40 volts DC. Look for two large capacitors (round battery looking things) near the input power terminals to verify this. If you only see a single large capacitor, then it's not a two pole power supply and not interesting. Two pole 27 volts is what we intend to have in our 48 volt system, so that is basically already rectified. The rectifier won't care if you give it two dc power lines and a ground half way between them. It won't even care which voltage line goes to which ac power input. It will rectify that power properly into its capacitors either way and do its job. An example board for this purpose is the YJ TDA8950 being sold on ebay for around $33.

Powering Pumps for Water

Powering pumps for water is one of the things that dc power sources like photovoltaics are best at. Typically, it is only necessary to pump water during the day. This can be accomplished by pumping water into a pressure tank or up onto a hill. A pressure tank is a sealed container with air inside. As we force water into the container, pressure rises trying to push the water back out. The pressure can continue to push water during the night when no power is available from the sun.

Most solar power outfits will offer a few lines of water pumps. There are of course, wide price and reliability differences between pumps. Make certain that your pump provides adequate pressure for your system. To help do that, a pressure conversion utility may be useful. http://www.onlineconversion.com/pressure.htm .

Pumps come with voltage ratings and so do solar cells. Use matching voltages. Keep in mind that solar cells can be wired in series or parallel to achieve different output voltages. Wiring cells in parallel will not change the voltage output. Wiring them in series will add their voltages up.

In the case of pumps, there is basically no need for power storage. You can wire pumps directly to solar cells. However, because we need reliable water, and sunshine varies greatly, it is wise to power pumps with at least double the power capacity as what they demand. This means that during times of cloudiness, pumps should be able to still run at around half power. With the extra power capacity available at high sun, it would be a waste not to use a charge controller and help power your system with it. If you do this, then you may want to connect the pump to the output side of your charge controller, but in my opinion, its best to just leave it on the input side so you don't have to worry about the pump using up your precious power storage during the night.

Most dc solar pumps run on 100 watts or less.

If you are dc powering an RV or boat, the most interesting dc pumps will be 12 or 24 volt units. These are very easy to find at just about any solar supply location. If you are dc powering a home or buildings, The most interesting dc solar pumps will be 48 volts. These are hard to find. One of the issues that seems to be governing this is that the need for 48 volt dc appliances has only recently arrived and the west typically does not have respective products. I had my best success looking for recent models from china, although some models exist from the west. For some interesting options, put the following search string into ebay "48 volt pump -48F -24 -12 -6 -240 -230 -110 -120 -115 -220" you have to have all of the minuses because for some reason, ebay barrages you with items that don't match the request. Try "48v dc pump" (also works) or other word combinations. There are currently 48 volt well pumps and 48 volt submersible pumps available on ebay, both from china. Flojet and Shurflo currently don't seem to be making 48 volt pumps. Dankoff, grundfos and sun pumps make (expensive) pumps compatible with 48 volt lead acid storage. Alibaba and aliexpress have many more options if you are ready to manage the haggling. Here are some more related web links:

How to Implement a Dryer or Heat Your House

One might imagine that something that needs as much power as a clothing dryer would be next to impossible to power on a solar power system. Hehe ..ok, yes, of course, that is very wrong.


That is a link to a manufacturer of twin-wall polycarbonate plastic. Simply build 2 to 3 4x8 foot boxes with a lid of twin wall polycarbonate plastic. put black metal ducting of some sort inside. Place the boxes in the sunlight and draw air through the boxes with a solar powered fan to feed your dryer with very warm air. Now, for your dryer, you have no more need to power the heater, only the motor to drive the dryer. And while you are at it, with a bunch more such boxes and about 40-50 50 gallon drums of water to store heat, you can store enough solar heat to heat your house in the winter time.

Of course, you could only dry clothing during the day, but hey, the power would be cheap.

Our suggested system for a single dwelling on a minimum budget

We set up a small utility shack with 9 250 watt panels to run power for a trailor using an inverter. I discovered early that the trailor buzzes the transformer for its battery charger all day all night, sapping around 1 kwh out of the power budget. This needs to be resolved by switching that charger on/off as needed rather than running it continuously.

The first test period was winter time (december) and it is great to have information from that time as it may be the roughest season for the power system. When cloudy days came, we definitely had problems getting enough power, even when we cut a lot of stuff out of use. Cloudy days seemed to offer only around 1/5 to 1/2 the normal power.

The trailor was 38 feet long and includes a somewhat small refrigerator, lights, and outlets. We ran laptop pcs on the outlets, and also ran electric heaters about 40 minutes at night and about 1 hour in the morning. We avoided using the dc powered tungsten lights and ran cfls from some light sockets that ran off of ac. The final verdict is that using this setup, the 9 panels is not quite enough power. 18 panels should do the trick in my opinion, and we will be setting the next 9 up when we get the chance.

We used a PowerJack 8 kw inverter, oversized for reliability and safety purposes. We used 8 6 volt 113 pound 370 ah DEKA batteries offering about 18 KWH storage capacity of which one only wants to use a maximum of 9 on a daily basis. We set the rule that the largest motor to be started from the system is 1 horse power. The powerjack inverter stopped producing power when the (combined) battery voltage dropped to around 47 volts.

For the 18 panel solution, the final price tag with DIY installation will be around $10000. It should produce around 28 KWH/day on average during the year, which calculates like this, using california's $.13/kwh power cost.

$.13 * 28 * 365 = $1328.6

This system should pay itself back in about 7.5 to 8 years, leaving around 20 more years to recover around double the cost of the system in power. It will at best only be about double because a: batteries will have to be replaced and b: the panel output will drop down appreciably in 30 years. Additional maintenance costs, should they arise, will be problematic. If you have to pay someone else to do the installation work, your final cost situation will be around $5000 higher, stretching your payback time to around 12 years.

To attempt to reduce the cost of this, one might use just 9 panels and refrain from any use of power by hungry devices such as heaters and desktop computers. One might double insulate a refrigerator, etc. You could cut the cost of the system by $2500 or so that way, but the final price/watt ratio would rise because many of the tasks that have to be done for the larger system, still have to be done for the smaller one.

  • Place your panels or wind mill as close to your building as possible.
  • Use two charge controllers to charge two batteries or two sets of two 24 volt batteries in series for 48 volts total voltage. Use an additional charge controller that jumps from pole to pole, always charging the weakest pole. For a 1500 watt system, 45 amp charge controllers will be more than adequate. 30 amp controllers should still work.
  • You will want 1000 to 1500 watts of production capacity. Approximately one third of the capacity should be connected to each charge controller.
  • Only the dwellings lights will be powered at night.
  • Use a double pole double throw relay and a system of comparing the battery voltages to swap some of the power back and forth from one battery to the other to keep their voltages identical. A circuit to control this floating charge generation capacity is easy to design and requires only components available at a location like Radio Shack. (design schematic coming up)
  • run 3 wires (with fuses or breakers designed for dc current) into the dwelling. One wire is connected to the plus pole of the high pole 24 volt battery. The second wire is connected to the minus pole of the same battery, and the plus pole of the second battery. This wire is your 0 volt line, and does not require a fuse or breaker. In fact, it would be better not to have it. It would also be good to ground this wire to the ground using a single copper or copper plated grounding pole, or a set of multiple (10 or so) iron rebars or grounding poles. Grounding for this low voltage dc is however not a necessity. The third wire is connected to the minus pole of the second 24 volt battery (or set of batteries in parallel)
  • purchase a low power (1-2 kw) expensive type true sin-wave converter that inputs 48 volts and connect it across both batteries.
  • run 3 more wires from the inverter output into the dwelling with fuses or breakers on all hot wires (wires with voltage). Be sure to ground the neutral output or ground of the inverter to your grounding pole(s).
  • If you can pump near the dwelling, try to use 48 volt dc pumps for efficient power dist. Connect the 48 volt pumps across the inputs of both charge controllers in series. If you can not pump near the dwelling, then set up photovoltaics near the water that power only the pumps. Set up at least double the power capacity that the pumps require.

The three dc voltage wires now provide two poles of dc power that can be used separately, providing 24 volts for 24 volt devices, or in tandem for higher efficiency and devices that can use 48 volts. dc-dc converters can convert down (ideally from 48) for 12 volt devices. Use 48 volts when possible. That is your most efficient and inexpensive use of power. Use 48 volts for all of your lights by connecting identical 24 volt l.e.d. lights in series with each other. There is no need to wire ac power to your lights. 14 guage wire is good for 15 amps. 16 guage wire is good for 10 amps. Even 16 guage wire is good enough to power l.e.d. lights with 48 volts. In fact, even 18 guage wire and 7 amps would be adequate for this if you use multiple circuits.

Wire outlets in each room with both power options. Use standard outlets for the inverter power. Use different outlets for the dc power so that the plugs can not be interchanged. 14 or 16 guage wire should do for ac power depending on local codes. If you are using 220 volts, 16 guage should be plenty.

If you wish to power more than lights at night time, you may add a generator to the system at additional cost either up front or at a later date.

Our suggested system for a single dwelling on a higher budget

Solar Power and the NEC (Regulation, or the National Electric Code A United States thing)

Adoption and enforcement of the NEC (national electric code) may be spotty. You will have to find out the details of the enforcement on your own. In any case, it would be good to be aware of what the NEC people might be concerned about regarding solar power systems. Fortunately, someone else has already created a great writeup for this issue, so we don't have to. Please refer to the following link:


more related links

Most solar power and dc distribution regulations are found in article 690 of the NEC. Many attributes of those systems are defined elsewhere because they are not special to dc systems.

The most notable issues regarding NEC and solar power seem to be that both arc flash and ground fault protection are required on the panel side. Arc flash protection only being required if the system voltage is above 80 volts. The midnite classic solar charge controllers implement both of these functions at a savings of approximately $200.

Another notable issue is that regarding pv and dc power connectors. DC power connectors have a special concern or shall we say: potential problem. Large dc currents are so determined to cause arc damage that if we repeatedly connect and disconnect a dc power connector under load, the contact surfaces of the connector may be damaged. A damaged power connector may not properly connect, causing oxidation. Oxidation causes the connection to degrade, and as it degrades, it begins to absorb energy and heat up. This can eventually result in electrical fires. For this reason, code asserts that pv connectors must require a tool to connect/disconnect and they must sport a notification that they must not be disconnected under load. Standard MC4 connectors as well as several others in the solar industry satisfy these constraints. It would be unwise to use anything else.

Solar panels are required in the US to have bypass diodes. The bypass diodes supposedly create safety by allowing a shaded cell to be bypassed rather than forcing it to draw power and heat up in stead of creating it. They also help maintain panel output when some cells are shaded. Panels have a maximum fuse value because of the supposed possibility of current that is moving through the bypass diodes over-reaching the capacity of the diodes (a very unlikely event) because of this, if panels are ganged in parallel, each panel or series group of panel has to have its own fuse to meet code. One way to solve this is to solder a fuse holder into the little components box on the panel, or just on one of a set of panels in series. Another, more expensive, but prettier solution is to fuse each string or set of panels separately in the combiner box. This costs a lot more as it requires each string of panels to have its own cable leading back to the combiner box, and pv cables are not cheap.

It may be wise, even if you do not live in the states, to implement these functions.

More complete, but still abridged list of NEC concerns for typical dc systems. Some important grid-tie concerns or concerns for unusual circumstances not included:

NEC Concerns [John Wiles] The NEC suggests, and most inspection officials require, that equipment identified, listed, labeled, or tested by an approved testing laboratory be used when available [90-7,100,110-3]. Two of the several national testing organizations commonly acceptable to most jurisdictions are the Underwriters Laboratories (UL), and ETL Testing Laboratories, Inc. Underwriters Laboratories and UL are registered trademarks of Underwriters Laboratories Inc., 333 Pfingsten Road, Northbrook, IL 60062. Note .. UL approved devices are not always available.

(690.5):ground fault Installations of PV arrays not at dwelling units do not require ground fault protection. ground/pole mount pv arrays with not more than 2 parallel circuits and dc-isolated from buildings require no ground fault protection

690.41:Conductor ground: 6 gauge minimum, and not smaller than, one of the two 2-wire outputs or the center tap of a 3 wire system for power conductors, color must be white or gray. For nonconductors, color must be green, green/yellow, or bare wire

690.43 exposed, non-current carrying conductive components shall be grounded. There must be direct grounding connection between pv array and equipment (earth not acceptible) in the same raceway with other pv cables. Bonding devices seem to be expected for metallic pv frames.

[John Wiles] CONDUCTOR COLOR CODES The NEC established color codes for electrical power systems many years before either the automobile or electronics industries had standardized color codes. PV systems are being installed in an arena covered by the NEC and, therefore, must comply with NEC standards that apply to both ac and dc power systems. In a system where one conductor is grounded, the insulation on all grounded conductors must be white or natural gray or be any color except green if marked with white plastic tape or paint at each termination (marking allowed only on conductors larger than 6 AWG). Conductors used for module frame grounding and other exposed metal equipment grounding must be bare (no insulation) or have green or green with yellow-striped insulation or identification [200-6, 7; 210-5; 250-119].

690.10(c): no multi-wire branch circuits on inverter output, and warning label at box "Warning Single 120 volt supply, do not connect multiwire branch circuits"

690.13 means shall be provided to disconnect all current carrying conductors of a pv power source from all other conductors in a building. Grounded conductors may as according to 690.17 be disconnected by means of connectors if they are polarized and unmistakable. 690.14(C)(2):photovoltaic disconnects shall be labelled as such 690.14(C)(4):max number of pv disconnects in a system is 6 690.14(C)(5):pv disconnects shall be grouped with other pv disconnects, disconnects don't have to be next to the array 690.15 system equipment must be disconnectable from ungrounded conductors

690.16 fuses must be disconnectable from all sources of supply (place fuses after breakers, check for back-feed)

690.17(1) breakers/switches are readily accessible

     (2) breakers .. operable without exposing operator to bare components carrying voltage,  
     (4) warning "Warning Electric shock hazard. Do Not Touch Terminals. Terminals on both the line and load sides may be energized in the open position"

690.31(A) specially designed pv connectors and if 30+ volts, raceway for readily accessible cables. (B)pv wire may be used in outdoor locations (non pv wire may not we assume)

690.74 Battery Interconnections: flexible cables must comply with article 400 all battery cables must be moisture resistant, welding cable not allowed, Fine stranded cables require lugs identified for such use.

690 definitions: the system voltage is the highest voltage between any two dc conductors

690.7(A) Max voltage is sum of rated open circuit voltage of series connected pv units (at lowest expected temperature) (this will affect choice of circuit breakers) 690.7(D) circuits above 150 volts to ground shall not be accessable to unqualified people. 690.7(E) Max voltage for 2-wire circuits in bipolar system shall be the 2-wire voltage if 1 conductor is grounded, each circuit connected to separate sub-array, and equipment is marked with warning: "warning, bipolor photovoltaic array, Disconnection of neutral or grounded conductors may result in overvoltage on array or inverter. 690.8(A)(1) max current shall be sum of parallel rated short circuit currents plus 125% 690.8(B)(1) sizing of conductors and overcurrent devices shall accomodate 125% of max current exception: a related assembly rated at a given current may allow overcurrent to be rated at 100% of the assembly rating.

690.6(C) multiple inverters may have a single output disconnect, single inverters must have an output disconnect (bolted, connector, or terminal type) if included in multiple inverter systems. 690.9(D) overcurrent devices require dc rating 690.9(E) series connected devices can use just one overcurrent protection

690.10(A) inverter must be rated higher than the largest single load 690.10(B) inverter outputs must be overcurrent protected. located at the output of the inverter.

690.45(A)grounding conductor size defined by table 250.122 .. (if you can find it) 690.45(B) for non-dwelling units without ground fault protection, ground conductors must have 2 times the circuit conductor ampacity

690.46 array grounding conductors comply with 250.120 (if you can find it)

690.53 pv power source permanently labelled at disconnect including mpp voltage, mpp current, max voltage, short circuit current, max charge controller output current (respecting max charge controller input current may be good too)

690.55 max storage voltage, including equalization voltage must be indicated. 690.56(A)for standalone, directory indicating location of disconnect and notification of standalone system

     (B)for grid tie, directory showing location of service and pv disconnection means

690.60 only grid tie inverters allowed to tie to grid 690.61 grid tie inverter shuts down when grid shuts down

Batteries: 690.71(B)(1)storage batteries for dwellings may not be over 50 volts nominal unless live parts of battery bank are not exposed during routine maintenance. (2)live parts guarded (C)battery short circuit current limiting required (E)for over 48 volt battery banks, banks must be separable into 48 or smaller volt components (G)more requirements for over 48 volt systems

690.72(B)(2)(2) ampacity of charge diversion cable shall be 150% that of the charge controller diversion ampacity

Because we brought up the dc light bulbs

NEC 410.6 requires listing for lampholders and luminaires, but not lamps (light bulbs)

NEC 410.5 requires lamps to have no live parts normally exposed to contact. While these bulbs are low voltage and current limited, and may be exempt from this, to be sure to comply with code, these bulbs (lamps) may require mounting inside of a light fixture which does not allow them to be touched during normal operation (either by hands or metallic lighting components).

NEC 410.16(3) allows LED lamps to be installed into closets if identified for the purpose. Our opinion: These (15 watt version of dc light that looks like the above photo but uses different diodes) lamps do not get dangerously hot. However, would recommend lower wattage lamps for closets (like maybe the 12 watt version with the schematic that shows up above), or once again, mounting this lamp inside of a glass or suitable plastic enclosure for a closet. The enclosure should be significantly larger than the lamp. additional notes: battery short circuit protection should protect against direction of current flow during a battery short circuit.

NEC 410.134 luminaires for dc shall be (properly equipped) and marked for dc operation

In my estimation, this means that if a dc light bulb requires current limiting (unusual), then the fixture contains the current limiting. So far, inspection has turned up for me that lamp holders tend to be generically rated for voltage. I.e. they do not state dc or ac. That generically means that they work fine for both. You may want to check out your own fixture though.

[John Wiles] PV systems may be required to have dc services with 60- to 100-amp capacities to meet the Code [230-79]. DC receptacles and lighting circuits may have to be as numerous as their ac counterparts [220, 422]. In a small one- to four-module system on a remote cabin where no utility extensions or local grids are possible, these requirements may be excessive, since the power source may be able to supply only a few hundred watts of power.

For a new dwelling, it seems appropriate to install a complete ac electrical system as required by the NEC. This will meet the requirements of the inspection authority, the mortgage company, and the insurance industry. Then the PV system and its dc distribution system can be added. If an inverter is used, it can be connected to the ac service entrance. NEC Section 690 elaborates on these requirements and allowances. DC branch circuits and outlets can be added where needed, and everyone will be happy. If or when grid power becomes available, it can be integrated into the system with minimum difficulty. If the building is sold at a later date, it will comply with the NEC if it has to be inspected. The use of a listed dc power center will facilitate the installation and the inspection.

some exerpts from:

Photovoltaic Power Systems and the National Electrical Code: Suggested Practices John Wiles Southwest Technology Development Institute New Mexico State University 1505 Payne Street Las Cruces, NM 88003




the transdev power supplies show up on both ebay and amazon

Personal tools