How to Build an Inverter Unit for a Gasoline Generator

A good start would be to track down the manual and then see if you can source the rectifier part or even buy a second non-working (cheap) generator you could pull the inverter out off.

The manual should also indicate how easy it would be to modify it and add your own inverter. It may be more complicated than you think if the inverter unit has a proprietary or weird design that makes it difficult to take a direct voltage out (positive feedback for generator voltage regulation for example).

Others with more experience repairing similar units may have other advice.

Thank you for you input! I have already tried to find another unit (that was the very first thought) but these original Honda's are hard to get. The market is overflown with cheap Chinese units and the owners of Honda's are reluctant to give away any- If you have it- you keep it . The user manual for Honda ex350 has a wiring diagram showing the three phase alternator with no center (common) wire, and yes, it does have additional winding (SW), while the inverter is simply a "box". I doubt that it is serviceable at all and anybody has succeed removing the black cover-layer of molded plastic over the board and do any repairs.

http://powerequipment.honda.com/pdf/manuals/31ZC3603.pdf

I have a shop manual covering Honda 700 and 1000 wats inverter generators and they are also show as boxes-no circuit or showing elements inside. It is the case with other types of AVR's- they are simply sealed-no repairs possible.

With grate help and patience of 67model we succeed building an AVR for a three phase generator which is still working , I hope with some help of the more experienced members I could build a device that would give a second life to this, otherwise grate unit.

http://cr4.globalspec.com/thread/67923?frmtrk=cr4sd#comment721956

I worked at a place that sold and repaired those things so I have some basic idea of what's in one.

The only thing I found good about their design was that once the inverter board fries in theory you could cut the factory PM alternator off and replace it with an automotive alternator and make something useful like a portable battery charger/booster out of it.

As you can see the whole problem is the inverter part of the whole generator is sealed and nearly impossible to open up to any degree and service.

Thank you all for sharing your opinions. All the mentioned options are in game but since the very first thought was to rectify the "wild" power coming out from the alternator I think that will be the first thing to do. If the results are not good the next might be to replace the alternator - if I found an automotive one with the same dimensions and power.

That means, put a "horse shoe" (that's how we call the diode bridge inside the automotive alternator), measure the output voltage and see what happens when a good and appropriate load is hooked.

I would have a question for RAMConsult, considering the pretty high voltage coming out from the alternator, shouldn't I put a board with a voltage regulator/DC stabilizer after the rectifying diodes so I have a stabilized 12-14VDC and than hook the UPS or the commercial power inverter 12/110/220? In that case, I assume, I wouldn't even have to use a car battery in this "system"?

thank you all again...

First I wanted to know what the inherent regulation was from the PMG under load, that was the purpose of the load test. If the voltage remains high then you would need to design an inverter that could accept that input range. Putting a DC regulator in to drop from say 200 VDC down to 15 V is a bit of a stretch, way too lossy for me, maybe a spare SMPS (Switch Mode Power Supply) from an old laptop will do the trick. Sooner or later your system will get too convoluted and/or you're going to hit a budget constraint.

The easier/cheaper solution may be the cannibalism route suggested by others, keep searching the web for a unit with a dead engine.

Yes, I will do the test as you said. I am looking for a good spare diode bridge from any old car alternator. I will inform you about the readings.

I have also searched the net about the "cannibalized" unit but couldn't find any, there is one inverter from a honda eu1000 in UK asking 150 Quines. The second problem about this option is the fact I live at Balkans, Ebay and Amazon don't usually send boxes here (I wonder why:)

The last option will be either rewind/alter the existing alternator for a lower voltage or swap it with a core of a car alternator.

Will be back soon with the results

Hi Mishel,

If want to try a rectifier, you will need to make it with say 6 x 1N5408 diode or equivalent. A 12 volt automotive alternator diode is unlikely to have a high enough voltage rating. For a load, 230V domestic incandescent lamps - 2 in series - would be needed to survive 400V. Pity the eurocrats are banning tungsten lamps bigger than 60W - you may need 8 lamps and lamp holders!

That rectifier and load set-up should show the engine and generator are working OK.

However, just to prove engine /gen function, load units of 3 lamps of 230V in "star" would work off the AC direct.

Or maybe, you have a crate of 230V incandescent "bar" electric heaters to hand (imagine smiling face here - CR4 "emoticon" button no work now!!!!).

Regards,

67model

Hi 67model,

thank you, good tip in a good time. I was about to use this rectifier pulled out from a car alternator. I assume it wouldn't survive. I will make a bridge using higher voltage diodes and make an appropriate load. Will be away couple of days, my orchard is under attack of a serious bug (Capnodis tenebrionis) and I must take some counter-measures.

Speaking about rectifier, since there is no common (- pole) wire from the alternator coming out should I use chassis as a minus pole?

Thanks for your contribution (again :)

Regards,

Misel

I meant like this...

The little diagram connects the bulbs direct to the AC.

I (finally) did the test according the instructions. I was out of town for some time.

These are the readings for a single phase load (one 60W E27 bulb) hooked on one phase of the alternator.

voltage under load 251VAC

current drawn 285 mA

3 phase test AC (star load of three bulbs 60W - one of each phase of the alternator)

Voltage (significant drop) at each phase 140 VAC

an

And finally, test with the rectifier (six 1N5408) and two 60w lamps - voltage measured at no load 290 VDC (at open circuit - according your drawing plus/minus) and with the load hooked 230 VDC.

I hope this will put some more light at the problem :)

greetings

Mishel

Thanks for info Mishel, the assumptions on bulb watts at reduced voltage I base on test of a 24V 1.2 watt bulb on DC smoothed supply.

1. You get maybe 251 volts x 0.285 amp = 71 watts with the single phase load.
2. With the star load, assuming you mean 140 volts on each lamp, getting 47% the watts at 60% rated bulb voltage, makes 60 watts x 0.47 x 3 bulbs = 85 watts.
3. 140 volts to star point would be 140 x 1.732 = 242 volts line to line for a 3 phase sinusoidal supply.
   
4. With the 6 diode 3 ph rectifier, 290V no-load, suggests 305 V peak, because average voltage for 3 ph full wave rectifier is ~95% peak voltage.
5. Per 3, 115 volts per lamp gives 33% rated watts = 2 x 60 watts x 0.33 = 40 watts. with 230V DC average.
6. Regarding 3, a 230 V rms sinusoidal mains supply, like the home wall socket we know, would have a peak voltage of 230 x sqroot 2 = 230 x 1.4142 = 325V. Probably, this supply is not pure sinusoidal.
7. The single phase and star loads at a higher power 71/85W have a voltage of 240-250, but 40W load is giving 230V DC - however, the usual AC voltmeter measures the average rectified DC voltage of the AC and includes a scaling of 1.1 which is the ratio rms/mean for a sine wave. So a reading of 250V AC has a mean value of 250/1.1 = 227 volts.

So I guess there is a supply of around 240V AC , line to line, here, probably with a flattened peak compared to a sine wave. The DC change of 290V to 230V, for 0 to 40 watts, with the diodes, suggests the regulation is large, as expected.

The DC output with diodes is about the level you get in many AC input supplies which rectify to DC and follow with a DC-DC chopping converter. Some of these supplies in TVs, computers etc might work OK direct from a 2 wire supply from this generator.

Most likely, fluorescent lights of the tube (particularly modern electronic ballast types) or "CFL" (incandescent bulb replacement) type will work direct from the AC windings - best if balanced on the 3 wires. Would make the genset useful for emergency lighting and battery charging, at least.

The data for this generator suggests continuous power is about 250 W at Full and 120W at Silent selection. The engine speeds are given as 3000/4500 rev/min, Silent/Full, I guess. So the generator is going to be 50/75 Hz minimum with two poles - more likely it is 4 poles or more, so 100/150 Hz or more.

Sorry, the Honda diagram does not reproduce well. But moving onward, there are some questions.

1. What does changing the speed selection, Silent/Full power, do to the generator voltage?? Change in proportion to speed is expected.
2. What does full load do to the output voltage?? Try say 3 x 200 watt incandescent floodlights in star. Does the speed fall with load??
3. What is the voltage of the two wire winding from generator to inverter?? Its wires are Yellow and probably Gray (unreadable) on the diagram. Silent/Full power settings please.
4. Were the figures you gave with the Silent or Full power selection??

67model

Hello 67model,

I should have addressed that part, too. I was going to but I simply forgot. I altered the carb adjustment and set rpm "somewhere in-between" :) The lower setting was to weak and the higher was to noisy and smelly (two stroker:) so the readings that I gave were done at the average mid-rpm ( I cannot tell for sure the exact rpm).

1. I've done the measurement today with the three star ac load (3x60W) and the voltage at the single phase fluctuated from 150VAC to 240VAC (min rpm/max rpm) Full power setting, pulling the lever to the Silent power at max rpm reduces the voltage from 240 to 175VAC.

2. I cannot use higher power bulbs (today is Sunday and the stores are closed) but disconnecting the load of these three lamps while engine running make the rpm rise. Again I cannot be specific since I don't have any possibility to measure RPM.

3.The two separate wires gave from 12VAC to 18 VAC (yes Yellow and Gray),

4.As I said , the figures given at Full power selection ( with reduced RPM)

My opinion is that in order to achieve the declared 300-350 W output power this small engine has to work on a higher rpms which is not quite a good thing. I would prefer 200-250 Watts with a quieter and cleaner environment. When I adjust the carb at RPM "on my own" I get 220 VAC with a load of 60W hooked and a 16VAC (Yellow/Gray) with no load.

I am aware that most of the lights and some of the TV's at my off-grid cottage wouldn't mind this power supply. On the other hand, I looked through the holes at the alternator and tried to see how many poles it had-did not count but it has many so I assume the frequency will be high for other applications?

regards

Mishel

Hello again, Mishel

I did not know you had to adjust the carburettor "fuel mixture"! I expected the engine would have some speed governor, which the Silent/Full power lever would adjust. Maybe "mixture" is just a another problem when you have an engine of unknown history.

Which makes me realize I never worried about why a two stroke motor bike does not rev-up to self-destruct if full throttle is used "out of gear". I found out the answer for four-strokes - the valves "bounce" and do not shut. No valves on a two stroke.......

Your measurements indicate the voltage range to be expected and that the Yellow-Gray wires would make a good source of low-voltage control power.

If you think a rev counter/frequency meter would be a useful tool, an electronic circuit using an analog multimeter as an indicator is not difficult to make. Operation from ignition coil or mains voltage triggering is possible.

67model

It is not just the "mixture" :), it is the "idling" screw that adjust the RPM at no load. And yes, it also has a small centrifugal adjuster with a lever to automatically change the engine speed according to the load hooked - and keeping the voltage to a certain level. That is the "motor part" that somehow stabilizes the voltage. I think this governor is the same as with the four stroke engines.

Yes, if you could direct me I am ready to build whatever you say :), shall I search the web by myself or you already have something on your mind?

Mishel

If you have some transistors and diodes scavenged from an old radio etc this works........

But if you need to buy new, a 5V low drop-out regulator LP2951 for stabilized supply with a 74HC4538 dual monostable would do well. Most analog multimeters would serve as indicator. See link below as a source example.

http://www.logingel.com/-1/browseProducts/1/byk/74hc4538

For supply frequency, a resistance-capacitance reducer from the high voltage would be needed as an add-on.

67model

Yep, can do - but it could take some time , I've been thinking, if the frequency of the alternator output voltage is the crucial information that we (you:) need, my Godfather has an oscilloscope and we could measure the frequency range propotional to the voltage at the output connector of the alternator? Wouldn't it be simpler and would it supply us with needed data or the RPM is more important so I need to build the circuit you have sent?

Mishel, I believe you are "overthinking" this situation and making it more complex than it is. The Honda engineers were very clever when they designed this system and use it in other products as well. First and foremost this is an open loop system, there is no feedback loop from the electrical load back to the engine, all that "magic" takes place in the electronics package.

The "load lever" is simply a mechanical rev limiter that sets two different steps on the carb's throttle plate, it does not know nor does it care about frequency and/or load.

The electronics package takes whatever AC voltage is input to it, rectifies and smooths it, then inputs the varying DC voltage into the inverter section which produces a fixed frequency AC output voltage/current for consumption. It is no different than an inverter package for a PV (Photo Voltaic) array which takes the DC Voltage that varies as the sunlight that strikes the panels, then provides a nicely regulated AC output.

You might even consider finding/winding your own 20:1 transformer, running the generator output through it, rectifying it and inputting that into one of these commercially available inverters (cost seems a bit high to me for the effort though). Otherwise just take the raw output of the alternator and use it to power a device that accepts any frequency and/or voltage that you throw at it.

Yes, I agree with you I just wanted to explain,,, I know most of the gensets have a centrifugal regulators (briggs, Tecumseh ,,,) while some have a vacum pumps that "hold" the engine at certain speed. I have not seen any fix two steps (like here 150W/250W) position button at the lever on any other generator ( and I've seen many) But no matter, I was going to ask 67model about using a transformer. It would be the most elegant, quick and cheep solution I guess. It just might be too big to fit inside the cavity, if let's say, we have 350W output power at 12VDC the trafo should be able to handle at least 20 A. I will take the 20Apms power supply from my Ham Rig and make a small experiment hooking all these together with a 400 W power inverter at the end. Will put some photos of this soon with the results. Maybe next week.

greetings

From the voltage reading you have indicated coming out of the PM Alternator, I would guess that the "regulator" or AVR is an 'H' bridge Mosfet or IGBT PWM controller. They would rectify and smooth the alternator's AC into a DC Bus then chop it to get the Sine wave output.

It occurs to me that most 12 VDC to 240 VAC inverters use a switch mode section to produce around 300 VAC at some highish frequency (16-24kHz) then rectify and smooth it then chop it with an "H" bridge to generate the Sine wave AC. This latter section being the same idea as that in the Honda

So you could feed the internal 300 V DC Bus of one of these inverters from the rectified and smoothed AC from your alternator. The smoothing capacitor (DC Bus) would have to be upsized depending on what frequency your Honda alternator produced. I'm guessing 400-1000 Hz. So the filter caps would probably have to be ten (or so) times larger.

Just a thought...

One of the options is to buy a secondhand previously-loved unit and sell the remains of this one via an internet auction site (name withheld). Internet stockists offer these models locally for £200-350GBP + delivery. One might get £50-100GBP for the remains of the old one, perhaps.

Another option would be to find one on an internet auction site with a knackered engine, and cannibalise 2 to make 1 good one. The remains would always come in handy for spares.

Interesting ideas for the most part but unfortunately due to the higher and highly variable working voltage that the current PM alternator puts out it creates a lot of rather basic problems for doing much of anything with without a dedicated voltage regulation system of some sort.

As I mentioned earlier just removing the factory PM alternator and replacing it with a common internally regulated one wire type automotive alternator would at least give a reasonable and fairly stable DC power source out of it from which an off the shelf power inverter of your choosing could be ran after that.

Given the internal space inside one of those gensets you could conceivably stick some gell cell UPS type batteries in it as well which would give you a bit of usable DC power without running the engine plus give the alternator an actual battery to work with when the engine is running.

Also if a power inverter was ran off of the unit the batteries would give it a bit of extra reserve power above and beyond what the alternator can produce.

Basically in rational terms if you are not knowledgeable in power electronics that higher variable voltage and frequency power you have now is not an easy thing to deal with whereas a standard 12 VDC source is.

That's my theory as to how to make this into a somewhat more practical and useful device again.

Yes, I agree with you. I understand the difference (now:) between the car alternator and the alternator within the inverter generator. Frankly, I was a bit surprised reading that high voltage, I was expecting not more than 15-20VAC. According the wiring diagram I thought it would be an easy job!! I wouldn't had to initiate all this, simply would hook rectifier and use commercial 12/220 inverter. But, this makes the challenge even bigger. The last option will be to swap the original alternator (I could maybe use it for my next wind generator) and put the automotive one with the battery cell as you said. Anyhow, firstly I intend to follow the instructions of 67model and do the "stress test" and collect some info about how the generator copes with load. During the day time I am out of town but next week, first thing to do will be to hook the AC load and record the currents/voltages and then try with rectifier and find appropriate DC load to the voltage gotten. It is also possible to try one or two old lap top power supplies and see if I can get a steady 19VDC.

regards

Mishel

If you are into wind power I would definitely say put a bit more effort into keeping it intact as you separate it from the engine!

Now as is once the three phase is rectified and turned into DC most any universal voltage power supply will work just fine off of it as long as the DC voltage is within the voltage working limits of the power supply.However you get a bit different upper and lower voltage limits using a DC input though.

On a DC input the voltage limits will be what ever its AC ratings are times 1.414 which means that if you have a universal 88 - 264 VAC rated power pack it will work just fine on DC anywhere from 120 VDC all the way up to around 400 VDC!

Good point from tcmtech about using generator for wind power, if rebuild is not economic.

But the real benefit of using a PMG generator with a rectifier and inverter to get 50Hz is that a high frequency generator, say 400 Hz, is much smaller than 50 Hz for the same power.

So what is the frequency of the generator??

Also, if AC - DC - inverter is used, a 9000 rev/min motor bike engine, rather than 3000 rev/min max for 50 Hz, is possible.

Seems as though Honda thought they could make the inverter small, cheap and reliable - but failed on the reliable.

So what is the engine governed rev/min??

Another thought, these Solar power inverters are made for high DC voltages.... [trouble is, the Grid-connect ones are designed to go off-frequency and shut-down if the Grid fails]

Thinking about it, if you do not want a pure sine wave, inverters are not so complicated...

67model

Yes, using this alternator would be grate for a wind generator, but I am not (still) giving up :) Today I've bought some things needed for the load test and hopefully I will have results before the end of this week.

Building up a wind generator from scratch is a bit challenging project here because we do not have well supplied stores with that sort of electronics. Permanent magnets are scarce, I've been using (various shapes) PM from old hard drives and have followed grate site http://www.scoraigwind.com/ for instructions and so. The point is, those home made generators are (again) low DC voltage ones for low RPM and this alternator would most probably cause problems in the later application of the output voltage.

I can't tell the rev/min now but I will investigate that and let you know.

will be back soon

regards