The original Honda AVR circuit was not reparable (not even close). That was the most tenacious potting material I have even experienced. As mentioned, just trying to get to the defective part would have certainly destroyed probably half of the circuitry on the board. I never did do a failure analysis due to the de-pot being a destructive effort. However, in looking at the circuitry, I could formulate a failure theory. Most likely, the switching transistor shorted out.

In coming up with a replacement circuit, I tried a simple circuit first, that basically fed the voltage regulator supply voltage to the field winding. The result was unsatisfactory due to very poor generator voltage regulation. For round 2, I decided to mimic the operation of the original circuit but, use newer parts that were more suitable for the job, in addition to using parts that are easily sourced, now and in the future. This seemed like the minimum risk approach. Basically, the regulator circuit pulses the field winding during each positive and negative cycle, with the duty cycle varying as necessary to keep the generator output voltage constant. I wasn't happy with Honda's original AVR circuit, it appeared to be an unstable circuit that was fixed by throwing various parts at the circuit until it behaved satisfactory.

Working with James was a learning experience for both of us. We did all of the work using E-mail since we live about 1,000 miles apart. I designed and documented, while James attempted to build and test the proposed circuit, along with taking test data on the generator and Honda AVR. I assumed James had prior circuit board building experience but, the little bit he had was so long ago, most was forgotten, however we got through. I actually appreciated the less than perfect board construction because, if the board functioned properly, it increased my confidence level the board would perform well in real world environments.

The scope plot of the switching FET was connected to the drain. The N-channel FET is being used to ground the field winding when it is pulsing the field. The other field coil winding is connected to approximately +18Vdc. However, the confusing part is, the scope plot is inverted so, what looks like a positive pulse, is actually the FET turning on and grounding this line. There is a "catch" diode used to short out the field winding discharge when the FET turns off. The somewhat goofy/non pulse shape of the response is due to the field winding ringing as it discharges.

Miscellaneous - I was planning to re-layout the board using surface mount components, the proto board used in the video was for Proof of Concept only. I don't plan to use a 10 turn pot, I want a 3/4 turn pot with a very limited adjustment range. The pot also needs to be moisture and fuel resistant so, not your typical pot. Yes, I do plan to look at control loop stability under various inductive and capacitor loads, along with surges, load disconnects/connects, and whatever other torture methods I can come up with to make sure the product is reliable out in the field.

Most generators fail because of improper maintenance. Taking care of a generator or any small engine( most of the time) will make it last a long time.

@@fcdrummer1806 yeah that generator was well cared for and the forty years I owned it before I donated it to the channel. It always started as easily for me as it did in this video.

@@louisbigelow1866 I can see that. It doesn’t take much to clog a carburetor. It’s easier to do a bit of preventative maintenance then having something go wrong in the future.

Do you have any resources to figure out how to reverse engineer a simple AVR like this?

Hey Tom, that smoke sneaks out every so often. Problem is, I have never been able to find a component smoke refiller. It's probably on the back shelf at Walmart next to that spray that keeps amplifiers from oscillating and oscillators from amplifying.

Dust off the old PhD (post hole digger) and please give me feedback on the idea of using the sense voltage tap AC, bridge rectified to provide signal input to a transistor output optocoupler, or ( even a vactrol could work), with the output then used to modulate an adjustable PWM. Hardware Hackers Rule !

When I was in school for Computer/Electronics Engineering Technology, my first experience with this was when I was building a variable benchtop power supply and I put the op-amp in upside down. I flipped the power on and SNAP I saw a spark and smoke rising from that op-amp. Pencils have erasers for this reason.

@@billharris6886 My Walmart doesn't have anything in electronics pieces, I have to rely on Digi-Key, Element 14 or Sayal for that kind of thing.

You beat me to it.

I was going to suggest checking sources for Japanese motorcycle parts, because it looks like the same style used for switchgear connectors (headlight+turn signal switch etc), but that part number pins it right down. Thanks!

Well, my high school physics teacher told us essentially what Ms. Frizzle has been saying on the Magic School Bus: "Take chances, make mistakes, get messy!" His was more in the way of, "we learn from our and other people's failures, as in order to learn, we have to fail first."

@@WJCTechyman good grief.... I used to watch that show and remember that line fondly.

The "body diode" inside a MOSFET can be used sometimes, but they're notoriously slow, and a great many applications will use an additional external diode in that spot to deal with that issue. Not having a clue as to what the circuit is actually doing, I couldn't say whether this will make a differene or not.

@@Roy_Tellason Not only are in die diodes slow, they are unreliable too. There may be some flyback voltage as the OP said and over time this does for the diode. An external one would be a good choice in any event for an inductive load.

Note that the oscilloscope plots are actually showing an inverted signal. We are using an N-channel power FET with the source grounded. The voltage is about +18Vdc when the FET is switched off, 0Vdc with the FET turned on. I did include a high speed switch mode "catch diode" to discharge the field winding when the FET turns off. The somewhat goofy trace is due to the flybacking of the field winding. I made sure the FET is not switching fast to minimize radio interference. Since the oscilloscope is set to display a 60 Hz trace, the FET appears to be switching fast. I considered using a full linear regulator to feed generator but, there is no good way for the circuit to dissipate the substantially higher wattage this regulation method produces. Also, I tried to model the AVR operation after the original circuit to minimize developement risks but, use newer available parts.

Legacy automotive regulators used output voltage of the alternator (rectified) applied across a solenoid that attracted an armature piece that had adjustable spring tension like a relay armature to operate a contact slider as a linear potentiometer or rheostat. Same as for a fuel gauge sending unit, a fine gauge resistance wire coil, the position of a wiper contact would allow more or less current, exactly like a potentiometer. The simplicity and ruggedness of those devices made them extremely reliable and easy to adjust or troubleshoot or replace. I understand the advantages of technology but not all of the alternative legacy technology is in any way defective. A problem is in most cases it was more expensive to manufacture, easier for anyone to repair, lasted longer, and provided the consumer more in return for their purchase expense. What is actual progress may depend on a more subjective valuation, other than what is trending now.

@@artpatronforever Yes, the older cars I owned used electro-mechanical voltage regulators for the alternator and dashboard instrumentation. The nice thing about the mechanical stuff, it gave an indication it was dying instead of the electronic version's abrupt failure. The premise with going electronic is, with the lack of moving parts, reliability should be better - that is, if it is designed right (the BIG IF). Any electronic control circuitry can use either digital or analog design and end up with the same level of precision but, the question is: at the expense of how much circuitry? In this AVR redesign, I'm actually using a combination of the two to maximize performance with a minimum parts count.

@@billharris6886 A useful advantage for the mechanical regulators is adjustability for people who use their cars for many short trips. Alternator output is easy to adjust upward slightly to keep a battery charged. It is also preferable for having diode blocks not inside the alternator as an integrated module. With separate modules alternators run cooler and last longer. And a failed discrete part is more easily replaced. Most commonly the bridge rectifier fails or the regulator fails while the alternator itself is still working. Unit replacement designs are meant to be unserviceable and are built accordingly.

@@artpatronforever I never have encountered an automotive mechanical voltage regulator, either DC generator or alternator types. Adjusting these was a trial and error balancing act, you had to bend contact arms and return springs to adjust the regulation voltage. Then you had to go back and fine tune these settings to produce a constant output voltage versus load. Mechanical voltage regulators also constantly generated voltage transients on the 12 volt line. This was due to the rotor being fed excitation power from either ground (no charge), 12 volts through a resistor (half charge), or a full 12 volts (full charge) by a set of bouncing contacts. I completely agree about using a remote mounted 3-phase bridge rectifier for the alternators AC output (with good heatsinking of course). Living in Florida most of my life demonstrated how easily the alternator assembly (with internal rectifier and electronic voltage regulator) can overheat in the summer with stop and go traffic with the air conditioning running "full tilt".

@@billharris6886 Hello Neighbor in Florida. Just by coincidence a Honda EM500 has a 12V DC output rated 8A. How convenient and thoughtful of the manufacturer. There is more than one way to skin a cat. Since all of the main winding outputs are taps, an AVR only regulating one of the taps is done and the other taps follow that regulation with themselves regulated also. Any AVR variant that is suitable for 12V regulation could be associated with the 12V DC output. The lucky coincidence is that 12V vehicle regulators are cheap and common and might be easily adaptable to function as an AVR, in a variant configuration. There are many custom design variable adjustable output vehicle regulators and modifications in the public domain from car audio enthusiasts that could be helpful. I am certain that many of the vehicle regulators also have pulse width modulation outputs and that sense signal inputs can be incrementally changed by use of 1 or more diodes, that steps the regulator output response to match the fraction of a volt of each diode voltage drop. So even a a sealed and potted regulator can be adjusted incrementally for voltage increase by a string of series diodes. (Edit: I did review and this diode adjustment works on 3 wire electronic type automotive regulators but is not applicable to the PCM computer controlled regulators that have a PWM output. ) The diode blocks are separate on some of these modules. But the Honda EM500 needs 16V DC at peak demand for the Stator Exciter for Rotor, and a power MOSFET follower would work just fine there, to mirror the PWM output of the 12V nominal vehicle regulator, as a current sink for the Stator Exciter for Rotor coil, supplied by bridge rectified and capacitor filtered basic local power supply from the former EW Exciter Winding that earlier powered the AVR. The adapted 12VDC vehicle regulator would likely best be powered from the 12VDC output of the tapped main. But it could possibly be powered by a different local source conditioned and regulated to fit its requirement. Anyway, sorry for stream of consciousness pedantics. But where there is a will there is a way, and an automotive regulator could adapt as an AVR for any generator having a tapped main with a 12VDC output. Maybe it is ghetto engineering, but when it comes to Afro Engineering during a power outage, it may be good to improvise anything to keep a man out of the dog house. On the old automotive mechanical regulators, are you familiar with a Filco Crown Jewel Regulator ? ( Edit 2: After a review of the PCM controlled vehicle regulators that do have a PWM output function, I see the PWM signal is originating with the PCM (engine computer) where the actual voltage regulation sensing is done also) So it is only the 3 wire automotive electronic regulators that would be possible to adapt for use, and should work, but would not necessarily have a PWM output. Depending on the internal circuits of the electronic 3 wire regulator, it may be an analog output or it could be PWM, a detail I do not know. A follower power transistor output would be needed in either case because of the 16V maximum needed by the EW Exciter Winding for the rotor at full output. A 12VDC regulator could switch that 16V supplied separately but a 12VDC regulator could not directly source a needed 16V for full generator power output.

Not a power MOSFET though in those days!

I was going to make the exact same comment. Jlcpcb charge just $2 for 5 PCBs, plus shipping and taxes. If required they will even fit (most of) the components for you with just a small set up fee ($8) plus the cost of components, but be sure to choose basic rather than extended components wherever possible as there is a fee of about $3 tool change for each extended component. That may worthwhile if you want to offer (almost) complete boards while having to do the minimum work. If you were having 10 or even 50 boards made the set up fee and tool change fee is still the same so the cost per board is reduced. If you use surface mount components the board will easily fit into the required space. @James Condon If you want to send me the diagram I can soon knock up a pcb layout for you.

Thanks for the PC Board tips!

Hi Matthew, yes the most basic generator is a brushless type where the field winding is powered from an AC capacitor. This type has the poorest voltage regulation and is substantially affected by engine speed. While it is possible to convert one to an electronic regulator, that modification could be quite extensive. Agreed, a compact PC board layout is the next step, although this project is a spare time project so, it will probably another year before we can get to it.

The mini heat guns is my sons. I “borrowed” it,

Probably not a huge demand on and old 400 watt generator, but a fun project non the less.

@@jcondon1 absolutely! A few more tweaks with some community effort and Im sure it will be ready! Great video and project!

@@jcondon1 Well, never mind then.

Yes some do have a way to disable the auto off.

Thanks Adam, the circuit topology you mentioned is what I used for the AVR.

The 3D printing is an option but, very costly in comparison to someone already making the finished connector so, that's where we are looking at the moment. Thanks though for offering a possible solution!

Bill is the master. I am just an apprentice.

I actually used a basic comparator instead of an LM555 to keep costs lower. I didn't need the extra features the 555 has. Thanks for the PCB suggestions.

Spooner Tuner retrofitted an AVR to a non AVR equipped generator in this video...kzfaq.info/get/bejne/iJpyprdmqrnSaJs.html I could follow along well, but I'm not sure how the resistor on the sense wire comes into play or how many watts it needs to be.