You can run this piece of chalk !

I do remember that time Bill Rogers and I were in a power shed out in the middle of the desert, maybe 25 miles from an oiled road.  Our mutual friend and excellent host had been using the gen set all winter, it had literally changed living in this remote spot from a serious burden to a joy, he had wired it all up and was quite happy. But there we were together wondering exactly why we heard that Generator Groan! It would have been easy to walk away, but to leave the riddle unsolved? It’s not like I was going to leave it alone, but Bill was like a bird dog on point! “George that just isn’t right!” And it wasn’t, but that’s another story.

Following is Bill’s account of a recent experiment in his PV Solar install, Bill doesn’t talk subsidies, he talks about real power, and his system is all about putting real power to work for himself.  Bill is a favorite friend, and of course he’s a DIYer…………………………………

Update Jan 30th 2016,Bill’s Book is now in digital form!  http://www.amazon.com/dp/B01B8IZOXA

(Bill Rogers)

Rectifier Loads on Small Generators

Do you have a battery-based solar system with a generator to provide power when the sun isn’t shining? Many folks who live off-grid do. Solar alone isn’t enough, and a generator by itself would have to be run such long hours that it would require constant maintenance and it would wear itself out in short order.

24 panels 225 watts each

24 panels 225 watts each 5400 watts total. 12 per FM60 charge controller in 4 strings of 3 each. That’s a buttload of power, but the load far exceeds that quite often, especially in the summer months.

A typical inverter-charger like the Outback FX series goes into battery charging mode whenever the AC input is hot. When AC input voltage and frequency are within tolerance, the load is transferred to the AC input and the battery charging sequence begins. And that’s the problem – power quality goes way down when that happens.

The typical small generator is an inferior source compared to modern sinewave inverters. In fact, a good inverter is the next best thing to the grid, and an argument could be made that’s it’s better in some circumstances. Voltage and frequency are rock solid. The lights don’t flicker with each power stroke of the engine and the frequency doesn’t droop way down when the well pump comes on.

It’s much better from a power quality standpoint to leave the load on the inverter and charge the batteries directly from the generator. If you’re lucky enough to have such a modest demand that it can be satisfied with a 12 volt inverter system, you’ve got it made. Just get a big truck or bus alternator, turn it with the prime mover of your choice, and connect the output directly to the battery bank. Even at 24 volts there are alternator options, but the price goes up considerably (think aircraft industry.) With the loads I have I’d be pushing the upper limits of a 24 volt system, so it’s 48 volt. There are no (affordable) 48v alternators that I can find anywhere.

By the way, a very skilled DIYer by the name of Bob Gayle – who is now the mayor of his hometown in KS – has had good success boosting the output voltage of the Leece-Neville 555 series (the triple nickel.) The alternator option at 24v may indeed be viable if you’re willing to experiment a bit (It would be interesting to compare the efficiency of that vs what’s presented here.)

The only downside to this whole arrangement is that it’s less efficient to make DC with the generator and invert it to AC. Some folks will consider the extra fuel consumption worth it for better power quality and surge capability, others won’t. An effort will be made to quantify the difference later so you can decide for yourself.

The heart of the system is a pair of Outback FX3048 inverters stacked to give 120/240 volt output. The well pump is the only 240v load normally, but there’s a 240v 1-ton minisplit heat pump that could be powered up if necessary. The battery is minimal since I’m on grid. It’s 230 amp-hours, which is a little over 11 KWH at 48v. At 50% depth of discharge there’s a bit over 5 KWH of useable capacity, enough to furnish the base load for a few hours. To go beyond that, plenty of sunshine is needed or a 48v power source of some kind.

Inverters

Outback FX3048 inverters 6000 watts capacity at 120/240 VAC

The generator has only 3 KW of capacity. It’s one of George Breckenridge’s outstanding permanent magnet generators (a PMG.) It’s driven by an air-cooled C186F. It’s rated 9 hp at 3600 rpm and 7.7 hp at 3000 rpm. It’s overdriven 5:4 so 60 Hz is at 2880 rpm. The power at that rpm should be ideal for 3 KW of output from the generator.

It’s an awesome little generator but it’s limited. It groans when the 1 hp well pump comes on. The lights dim and the frequency droops to 57 Hz. In comparison, the Outback inverters have 6000 watts of capacity with at least 12 KVA of motor starting capability. The well pump starting up doesn’t even blink the lights. The frequency doesn’t change with load. It’s rock solid at 60 Hz all the time.

Now, how do you go about making 48 volts DC? Sounds easy, doesn’t it? No problem, they’re called battery chargers, right? They’re everywhere. The telecom industry standardized on 48 volts DC many years ago. Most sites have 2 battery chargers – a primary and a backup. Older units are being scrapped during upgrades in favor of modern lightweight switch-mode units. My employer had several older rectifier/regulators that had been retired. I intercepted 2 of them on the way to the scrapyard. They’re made by C&D Power Systems and rated 50 amps at up to 60 VDC. That’s 3000 watts – same as the generator. Good fit, right?

No! The C&D charger is a battery-friendly source – that’s what it was engineered for, but it has to be a generator-friendly load too. That’s where it fails, big time. It’s all the little PMG can do just to energize the thing. It droops way down when you close the AC input breaker on the charger, then once it’s stabilized it draws over 6 amps at no load. That’s at 240 VAC. The PMG is rated for 12.5 amps at 240v, so it’s about 50% loaded with the charger on and NO LOAD.

Yes, it’s mostly reactive, but current is current to the stator windings. They don’t care if it’s in phase or out of phase – it’s load. And once it reaches 12.5 amps, that’s it. Doesn’t matter how much (if any) real or active power is being delivered to the load.

Other issues surfaced. The charger has a nasty way of “testing” the strength of the AC source when it’s first powered up. The output current ramps up to well above 50 amps for a few seconds before it falls back to the current limit setting. They’re made for the grid. And they’re of a ferroresonant design which has an appetite for reactive power. The grid has plenty of reactive. The charger can take all it wants. It’s also too smart for it’s own good. If battery voltage drops while it’s in current limit (eg, well pump comes on) it goes into alarm mode with a “rectifier fail” alarm. You have to power down to reset it.

Standard power factor correction techniques were applied to no avail. Some of the 6 amps at no load can be eliminated but not nearly all of it (and I didn’t understand why at this point.) All in all, it’s the wrong piece of gear for the job. It’s made for strong AC sources, something the PMG definitely is not. They were almost taken to the scrap yard (they weigh over 200 pounds each) but I’m so very glad I didn’t. I wound up re-purposing many of the components.

Now what?

What about a regulated current limited power supply? Fine, but where are the affordable ones? 3000 watts is a tall order. Older telecom rectifier/regulators are often available for scrap prices. Switching power supplies tend to have very poor power factor, and it’s the worst kind – it’s distortion power factor which is not correctable with standard techniques. The newest designs with active power factor correction are expensive, and the efficiency numbers are not all that impressive. Apparently the active PFC has a substantial tare load of it’s own.

Solar Charge Controllers

Ever since I’ve had a solar system I’ve thought about how nice it would be to be able to feed DC from a generator into a charge controller and let it do it’s thing. Modern charge controllers are highly efficient. The Outback FM60 approaches 99% efficiency under certain conditions. Regulation and current limiting – not to mention the other features that charge controllers offer – don’t have to cost much in terms of efficiency. And in terms of dollar cost, it’s already there. A charge controller is part of any battery-based solar system. It’s already wired up and ready to go. All it needs is a source of DC in place of the solar panels.

Raw DC

So what is the best way to make raw DC for a charge controller? The obvious thing to do is to rectify and filter the generator’s output directly. Feed the output into a full wave bridge rectifier and filter it with a capacitor. The problem with that is the FM60 has a maximum input voltage of 150 VDC that cannot be exceeded without damage. Maximum working voltage is 140vdc. It’s very important when sizing arrays to make sure that the open circuit voltage of the panel strings does not exceed 140. Other charge controllers may not have that limitation, but I’m not familiar with any of them. I have to use what I have.

When 120 volts AC is rectified and filtered, the DC voltage approaches 170 volts. If the generator’s speed is set to 62 hertz, the no-load output voltage is closer to 130 and the rectified DC then is well over 180 – way too high for the FM60.

A transformer is needed

A transformer to drop the voltage is the obvious thing, but that adds losses to the system. This turned out not to be a big deal at all, but I didn’t know it at the time. First thing I did was borrow the Outback X240 transformer that’s doing load balancing on the stacked FX inverter output. It’s a simple 2-winding transformer with a 1:1 turns ratio. Each winding is rated for 120 volts at 25 amps. It comes from the factory configured as an autotransformer – the 2 windings are in series. The goal is to get 60 volts AC which will be about 85 volts DC once it’s rectified and filtered. The FM60 will love that. It will be operating in the fattest part of the efficiency curve.

It can be used 2 different ways. You can apply 120 VAC across the 2 windings in series, and take  the output from the common to either hot leg. The 2-wire output would be at 60 VAC with a capacity of 50 amps. A 4-diode bridge is needed for rectification.

Initial circuit

Initial circuit. It worked but not well.

First attempt

Initial setup with X240 transformer 4-diode bridge and 8,700 MFD filter capacitor.

Another way is to apply 120 VAC the same way, but use all 3-wires for output instead. The output then is center-tapped 120 VAC which is two 60 VAC windings in series. It only takes a 2-diode rectifier for full wave rectification when you have a center-tapped transformer. Output voltage is the same. It should be slightly more efficient due to lower rectifier losses, but I wasn’t able to measure any difference. I built it both ways and it seemed to work the same each way.

Second circuit

This is functionally the same as the previous circuit

Revision 2

Revision 2 with X240 and 2-diode bridge. The rectifier module is one of the many goodies found inside the C&D chargers. The 70A DC breaker below the FM60 came from the charger as well. The results are identical to the previous setup but this rectifier has a higher rating and it runs much cooler.

A note about isolation

Autotransformers are great from an efficiency standpoint because the load current only flows through a portion of the windings – half of the windings in this circuit. But in every case, at least one of the output wires is connected directly to the input. The output is not isolated from the input. If the output has a ground connection – and it does when the load is a rectifier feeding a grounded battery bank – then it is absolutely essential that the source not be grounded in any way. For example, if I had attempted to do my testing with grid power (with it’s grounded neutral) instead of the isolated PMG, it would have tripped a breaker and hopefully not damaged the rectifiers. I never made that mistake I’m proud to say.

Results

This worked but not well. Power factor was awful – about 65%. The current from the generator had almost 100% total harmonic distortion (THD). That in turn ruined the generator voltage. It had over 20% THD. The transformer made a loud enough racket to be heard over the generator, and it was getting too hot too fast. The generator was fully loaded (25 amps) with just 1800 watts of load.

The answer I thought was to correct the power factor like I’ve done so many times in the past. Once again I felt like a kid without a clue just playing around. It didn’t work, in fact it made things worse. Time to stop and figure out what’s going on.

The problem

Any time the total current is greater than the active current – the part that actually does work – you have a sub-unity power factor. A phase difference between voltage and current does just that. The total current needed to furnish the load is higher than it would be if PF=1.

Another way that total current can exceed active current is if the current is loaded with harmonics. Distorted, in other words. The harmonics do no work but they generate their own copper and core losses and they add to the current demanded from the source. The current can be 100% in phase with the voltage, but if it’s distorted, PF will be less than unity. Substantially less when THD is nearly 100%!

This is called distortion power factor. The other kind that’s based on phase shifts is called displacement power factor. When you have both, the total PF is the 2 numbers multiplied together.

Rectifier loads cause distorted currents because rectifiers conduct only when the AC voltage exceeds the voltage across the input capacitor (first filter capacitor after the rectifier.) That’s only for brief moments at the peaks of the waveform. The current looks more like alternating impulses than a sine wave.

Current with no harmonic filters in place

Here’s what the current looks like with no harmonic filtering of any kind. Notice how much higher the negative peaks are compared to the positive peaks. THD = 91%.

The solution

The only way to reduce distortion PF is to reduce the distortion itself. Phase shifting capacitors won’t help, in fact they’ll make things worse because the current is already in phase with the voltage. There are 2 basic approaches to reducing distortion – harmonic filtering using passive components, and active circuits that attempt to force the current to follow the voltage waveform. If it did so faithfully, PF would be unity. Active power factor correction is preferred by manufacturers today because of cost, size, and weight restrictions. Passive filtering requires bulky heavy and sometimes very expensive components.

The testing setup

I got tired of running the generator every time I wanted to try or test something, so I decided to install a transformer so I can use grid power safely. I’ve had it for years. My employer sold it for scrap. It’s a type that’s very common in plants and factories with 480 volt distribution. A 2:1 transformer is needed to provide standard single phase 120/240. A large factory will have many of them in service, and they tend to get scrapped out during plant upgrades with a whole bunch of useful life left in them.

Transformer

Dry type transformer 240/480 high side, 120/240 low side. Very common in large factories and plants with 480 volt distribution.

This one is rated at 10 KVA, so the high voltage windings are good for 21 amps and the low voltage windings will carry 42 amps. That’s a good fit. The low side can be used as an autotransformer same as the X240, or it can be used as a regular transformer with a 120v high side and 60v low side. That’s how I did it initially because it provides the much needed isolation from the grid.

There is now a grid-powered transformer in place of the generator/X240 combo. The power factor improved by doing this (above 70% now) and that’s because the transformer itself serves as a harmonic filter. The current THD is 91% and voltage THD is only about 3%. Grid voltage tolerates distortion much better than a small generator.

Harmonic filter development

DC filtering

One of the first things learned is that you can reduce harmonics somewhat by filtering on the DC side of the rectifier. You need inductance though – a choke as it’s sometimes called. A filter capacitor alone won’t help harmonics, in fact it can work against you if it’s bigger than necessary. That’s because the peak current through the rectifiers (and thus from the source) is higher with a bigger capacitor.

You need inductance to reduce harmonics. Series inductance sustains current in much the same way that parallel capacitance sustains voltage. Working together they help fill the voids between rectifier firings. A good DC filter has a reservoir capacitor (the first one after the rectifier) followed by an inductor and another capacitor called a smoothing capacitor. The resonant frequency of the inductor and smoothing capacitor is what determines the effectiveness of the filter in reducing harmonics. Anything below the input frequency of the rectifier (120 Hz on a 60 Hz supply) will help, but for maximum benefit it needs to be 1/10 of that or less (12 Hz.) That’s very hard to do in practice.

Parts galore!

So where does this inductor come from? It has to carry the full output current so it won’t be insignificant. Well, it’s been on hand the whole time, 2 of them in fact. Each of the C&D chargers has a big power transformer plus some smaller transformers (single-winding inductors actually) that I never paid much attention to. One of them is wired in series with the DC output. And it’s followed by some really big electrolytics.

After extracting one of them, the first thing needed is to determine the inductance. Unlike capacitors, the inductance value is normally not marked on the unit. It has a C&D part # and nothing more. This is where a variable AC supply comes in handy. You can’t apply 120 VAC to such a low impedance without popping a breaker. I ran the voltage up (slowly!) until the inductor was drawing 10 amps AC. V/I then is the impedance of the inductor. The DC resistance is too low to measure so that makes the math easy. Dividing the impedance by 377 (2*pi*60 Hz) gives the inductance in Henries. This one is 0.92 milliHenries (mH.)

DC Filter

DC filter. Note the nice copper bus bars and the push button switch. The switch safely discharges the capacitors through a resistor. All of this was salvaged from the C&D chargers.

Now for the capacitance. C&D has 3 capacitors at 64,000 MFD each following the inductor, which is 192,000 MFD total. Well, guess what? A 0.92 mH inductor resonates with 192,000 MFD of capacitance at 11.98 Hz. They hit the design goal right on the money.

There’s another inductor in each charger that will come in very handy later. I had a wealth of parts the whole time and didn’t know it. I was so turned off by the behavior of the chargers when I tested them that I failed to recognize the value of what was inside each box. I came THAT close to taking them to the scrap yard.

DC filter configuration and results

There’s a problem though (isn’t there always?) The capacitors are rated for 75 volts DC. That’s fine for a charger with a 60 volt maximum output. The voltage is over 90 VDC at no load with the generator running at 62 Hz. The voltage usually drops below 75 under load but you can’t count on that. Too many variables.

To be safe I wired 2 caps in series. Two 64,000 MFD caps in series give 32,000 MFD total but at twice the voltage. To go all the way to 192,000 you’d need 6 times as much or 6 pairs of capacitors. That’s out of the question, besides the results were fairly good with just 32,000. The resonant frequency is 29 Hz, well enough below 120 Hz to do some good.

Current with DC filter only

Current with DC filter in place. It doesn’t look much different but the peaks are a little lower and the pulses are a little fatter. The DC filter alone reduces THD by about 10%

Another issue surfaced. The little 8,700 MFD capacitor was getting hot. The ripple current was way too high for it. It was replaced with another 32,000 MFD (2 more 64,000’s in series) and they’re fine. They get a little above room temperature and that’s it.

The filter with 32,000 MFD followed by the inductor and another 32,000 MFD raised the power factor to 80%. That’s a lot better but it’s not good enough, in fact if I don’t get solidly into the 90s I’ll consider the project a failure. The generator can now be loaded to about 2200 watts before the total current exceeds nameplate rating. I briefly ran the load up to 3000 watts and it took 33.5 amps from the source to do so. The generator is rated at 25 amps at 120 vac.

AC filters

Harmonic filtering on the AC side can be done several different ways. You can pass the current through an inductor and capacitor in series that resonate at 60 Hz. Currents at frequencies other than 60 Hz are attenuated. You can make a parallel LC circuit (a tank circuit) that resonates at 60 Hz and connect it across the AC line. Currents at frequencies other than 60 Hz are shunted and fail to reach the load. Finally, if necessary you can build harmonic traps that attenuate specific harmonics (like the 3rd which usually has the highest magnitude.) A trap is a series LC circuit that resonates at the harmonic frequency but it’s connected across the line as a shunt or through a resistor. All 3 methods can be combined if necessary.

AC filter configuration and results

The “other” inductor in each C&D charger is on the AC side. The charger’s AC input current passes through it, so it’s made for at least 20 amps. It’s about the same size as the DC choke, but it has many more turns of smaller wire. This time it took over 80 VAC to get 10 amps (vs just 3.5 volts for the choke) The math works out to 22.3 mH. To resonate at 60 Hz about 320 MFD is needed. That doesn’t sound like much but we’re on the AC side now. AC caps are much bigger and more expensive than similar sized DC caps.

Thanks to Joel Koch, I have quite a few 100 MFD caps rated at 370 VAC. He and is Dad used to build phase converters so he had quite an inventory at one time. I can use 300 MFD which would resonate at 61.6 Hz or 400 MFD which would give 53.3 Hz. I started with  400 thinking it’s better to be further away from 180 (the killer 3rd harmonic) rather than closer.

AC filter construction

AC filter construction. The 3 caps at the bottom are 100 MFD each and the one at the upper right is 20 mfd. I added a discharge switch to the caps same as the DC filter.

I started with a simple series resonant circuit. I’ll do the parallel shunts or traps later if needed. The results were startling! The scope is set up to watch the current as it comes on and the FM60 starts loading itself up. For the first time in all this, the current had a shape somewhat similar to a sine wave. It was ragged looking but it had the general shape. It’s been anything but that up until now. PF is 89% now.

Voltage readings across the series inductor and capacitor bank show that the actual resonant frequency is a little different than calculated. At resonance, the voltage across L and C will be equal. The readings indicated that the capacitance was too big or the inductor too small. Reducing capacitance to 300 MFD brought the voltages closer together and the PF went above 90 for the first time ever. The current waveform has some meat to it now.

Current with DC and AC filters

Current with DC and AC filtering. The peaks are lower and more symmetrical. Looks more like a sinewave now. THD is only 35%. This is comparable to a typical switching power supply with active PFC.

To determine the actual resonant frequency of the filter the generator was used instead of the grid so the frequency can be varied. At 57 Hz the voltages across L and C became equal. PF went up to a little over 91%. The system can now take a heck of a lot more load before the generator current exceeds its rating. I’m still not happy though. The filter carries quite a bit of current and so it gets rather warm in a hurry.

A note and warning about LC filters and stability

The bandwidth of a bandpass filter is determined by its Q or quality. The higher the Q, the narrower the bandwidth. In fact, the bandwidth is F/Q where F is the resonant frequency of the filter. We want a narrow bandwidth for maximum attenuation of the harmonics, but with higher Q comes the propensity for oscillation. If you’re designing a radio circuit, that may be exactly what you want. You build the highest Q circuit possible if you want it to oscillate with minimum excitation. But we don’t want that here. Oscillations can be very damaging. The FM60 could be subject to voltages way above normal. The last thing we want to do is blow it up!

The Q of a series RLC circuit is largely determined by the L/C ratio (actually the square root of L/C.) The higher the ratio, the higher the Q. It’s also inversely proportional to resistance. The higher the resistance the lower the Q. Resistance is low in this case. It’s the sum of the inductor resistance, PMG stator resistance, transformer winding resistance, and circuit resistance – all low numbers. So we’re left with the L/C ratio as the primary driver for Q.

This particular filter is fine. The Q is high enough to work but low enough for stability. The L/C ratio is about 70 and the Q of the filter with R=2 ohms is about 4. The bandwidth is 60/4=15 Hz. But what if the inductor was 100 mH instead of 22.3? To resonate at 60 Hz it would need to be followed by 72 MFD instead of 320. The L/C ratio is now 1,389, Q is 18, and the bandwidth is a mere 3 Hz. We certainly don’t need that tight of a bandwidth to do the job. It will be a lot less stable for sure, and could very well cause problems. Any Q value over 1/2 is underdamped and will oscillate in response to transients (like initial circuit energization.)

The good news is that you’re unlikely to come across a 100 mH inductor capable of carrying the currents we’re dealing with here. The best advice is to harvest your parts from equipment that utilizes the components in a manner similar to what you intend. The concerns outlined here will have already been addressed by the manufacturer.

More voltage, less current

The AC filtering would carry half the current if the voltage is raised to 240. The transformer is capable, so is the generator. There’s really no reason not to, so that’s what I did. It made a big difference. The results were the same but with half the input current. The filter is quite happy now. It gets warm after several hours but never hot. Full load on the generator is now 12.5 amps instead of 25.

Displacement power factor

After adding the series resonant circuit, I began to notice a bit of a phase shift in the current. How much of the remaining PF is due to a phase shift rather than distortion? The AC meter I’m using is a dandy. It’s made by Satec and it has the ability to separate distortion PF from displacement PF. I set the meter to read displacement PF only. It’s now 94%. It was about 99% before the filter was added.

AC metering

AC metering with Satec PM172E (upper right.) The Satec will show both displacement power factor and distortion power factor. The green-top cubicle things are current transformers (CTs) There are 2 on the grid side and 1 on the generator. The generator needs just 1 because it’s 2-wire only now. They can be switched in and out as needed. The generator breaker is directly below the Satec.

There’s room for correction there. The meter says we need to kill about 400 VARS of reactive. That calculates out to be 18.5 MFD across a 240vac line. I tried 20 mfd. The results once again were startling. Displacement PF went to 99, total PF is now 94-95%. The generator will now take over 2800 watts of load before the current exceeds its rating. We’re there, finally!

PMG voltage and current full load with DC and AC filters

Here’s what the voltage (yellow) and current (blue) look like from the generator at full load with all filters in place. Current THD is under 30% and voltage THD is under 10%.

Looking Good! says George

Full load with all filters in place. Top number is KVA, bottom is KW, PF in middle

Current THD

Current THD is under 30%

Voltage THD

Voltage THD is under 10%.

Final circuit

Here’s the final circuit

Efficiency

All the while I’ve been keeping an eye on the amount of AC power from the generator vs the amount of DC power coming out of the FM60. The generator has good AC metering and the FM60 has good DC metering on it’s output. The input metering on the FM60 is suspect but that’s irrelevant. The power coming out of the FM60 going to the batteries and inverters is what matters, and that’s proven to be very accurate.

This will be done much better later, in fact the best way is to accumulate watt-hours each way and compare those numbers over time. That’ll tell the full story. For now the instantaneous readings are very encouraging. For example, with 2380 watts output from the FM60, 2540 watts was being produced by the generator. That’s 93.7% efficiency from AC to regulated DC. That’s as good as a PFC corrected power supply, in fact most are in the 89-91% range. The few advertised at 93% are extremely expensive.

Engine room January 2016

Conclusion

You can’t imagine how happy I am to see these results. It looked like a bust early on. It’s been fun, if you find bumping up against your limitations, fun. I love having the FM60 in the picture. It gives a great deal of control and it has good metering and data logging. And if worse comes to worst, I still have an AC generator that can power the loads directly. The transformer serves as a load balancer. No need for a neutral from the generator. I can still backfeed 3-wire 120/240 even though the generator is 2-wire 240 only.

What I built is big and bulky but it works. It’s efficient and very generator-friendly. The current THD is under 30% which is outstanding for a rectifier load, and the voltage THD is well under 10% (compare that to an ST5 at no load) The components I used (assuming you already have a charge controller) are generally available for scrap prices.

I still need to quantify the efficiency at various load levels, and I want to do some fuel consumption testing. How many grams of red diesel does it take to deliver a KWH of DC to the batteries and inverters? I’ll be finding out soon.

Leftover scrap sourced parts

There are enough parts left over to do all that again and then some.

And there’s still something to sell at the scrapyard. Glad I didn’t do it too soon!

Disclaimer

This is for informational purposes only. It’s not a how-to guide. If you injure yourself doing any of this, it’s your own fault. Series resonant circuits can generate voltages MUCH higher than the source voltage. Be careful!

Copyright 2016. All rights reserved. This document may be freely disseminated as long as it remains intact.

———————————————-

(George B) My thanks to Bill for sharing his hands on experiment.  I will always remember a customer of mine who lived in a very remote mountain site. When he changed from a generator he had to an utterpower PMG, he charged batteries more quickly and with less fuel. It’s a fact that some generators are running lower voltages, I’ve seen many set or made to put out 105 volts! I also have a habit of plugging a Kill A Watt into friends’ generators and it’s rather amazing what I find, some I’ve found running at 55 hz. It’s always good to check, never assume, and normally a charger is happy to receive 126 volts if you can make it, and 60hz is generally a good idea too. Built in inverter chargers get the work done in less time when E is higher and still meeting ANSI standards. With the Kill A Watt so cheap, poor voltage and Frequency should never be a surprise.  Getting it right can save you money. Of course from there, you need remember Bill’s lesson… the method you choose to rectify it to DC can in deed be quite bad….. you need to check it, and the first step is knowing of the pitfalls.

All the best

George B.