Fixing the ETI 477 100W Audio Amp module

ETI 477 Power Module Circuit Board showing the bipolar transistors heatsink assembly.

Back in 1981 there was a design for a power amp kit, published by Electronics Today International.  The design was a pretty good one, which had very low THD figures and plenty of power.  More interesting to me was the idea that it would use MOSFET output stage instead of Power Transistors, or Valves.  The FET acted more like a valve in anycase, so all was good.  I had built it that same year, but had pretty much only ever used one channel, since I had only one speaker built. [other things to worry about i guess]

Even prior to that in the late ’70s at the University of Sydney I had attended all the graduate lectures of the great Cyril Murray [Electronics] and Richard Small [Audio].  These guys were the total gurus of amplifier and speaker design at the time and were really pushing the envelope in Sydney for how to make very high quality audio equipment – out of the new stuff – transistors – better materials – MOSFETs for example.  So when the ETI design came out, I could see the potential value and have something of the designs I had learnt from Dr Murray built into something I could put together.  Many others probably thought the same.


My son had his 18th birthday recently, and we had aquired two old speakers, so the thought was, lets connect them up and use those for the party.  It was a great idea, except that only one channel actually worked, the one i had used for the past 30 years.  The other channel was just creating lots of hum and noise.  So it was back to the single channel single speaker again.  The whole AMP / SPEAKER arrangement was placed outside in anycase, so it could have been cockroaches, insects, termites or any other kind of small living thing that tends to make its home in strange electronic equipment here.

Then, it was a long weekend, being in honour of the Queen of England, we all Aussies got Monday off work.  It is still raining Tuesday, and a short break over lunch..Looked inside, found about 6 resistors burnt out.  Here was a problem, no circuit diagram, and it took me a while to remember how the whole circuit worked as well.

So David Tilbrook had designed the AMP for the ETI magazine, and since the magazine was now out of date, i searched all over for someone on the web who knew anything about it…. Yes in fact., there were whole threads about this little known amp and its redesign by David Tilbrook, and the subsequent more advanced amplifier, the AEM6000, and if you are interested then join up with DIYAUDIO which is a pretty useful forum.

So this last Saturday, I downloaded the circuit, established which components were broken, went down to the local electronics place and could not believe that I could still buy all of the damaged components off the shelf.  I mean this amp was designed more than 30 years ago boys !.

So… then a day or so of painstaking replacement of burnt out tracks, and burnt out components.

I switched it on.. !@#$!~ heaps of smoke started pouring out of one or two of the resistors.  But since I was watching I saw which resistor was going first.  It was one just connected to a capacitor, a very small feedback capacitor {C14}.  For that much power to burn out a resistor it had to be a shorted cap, never heard of that before.  I checked the resistors, black but still working, transistors should be ok.  Resistor values were out by around 10%.  I only had it on for around 3 seconds.

Replaced the resistors for the second time.

When I replaced the capacitor, a very small one, I noticed that it had been mounted quite close to a straight aluminium heatsink, which has all the bipolar power transistors mounted on it.  It may have been heat stress or something that caused the capacitor to fail.  When I replaced the capacitor, I made sure it was lying down on the PC board, well away from the the high temperature heatsink.  It heat had been the problem for the little capacitor , then it would not be a problem for this little guy in the future.

This time turned it on, it worked, but the sound was not good, it sounded like crossover distortion..

So I carefully poked simple multimeter probes into the high voltage part of the amp, connecting them to either side of the variable resistor RV1, and adjusted the bias on the MOSFET gates to the same as the working channel, about 1.4 volts dc.  It improved dramatically, and since the most sophisticated measuring equipment I have is my ears, and a Herbie Hancock electronic music record, I figured it would be ok.  I then double checked the steady state temperature on the large heatsink, and both channels were operating at around the same temperature [around 40C].

Now both channels sound identical and very clear.

A nice amp, a great designer, and made of useful components, which are still available. The board layout is simple, and there are many small capacitors put in to compensate the amp for parasitic noise and oscillations.  I don’t have any problem with the AMP, It had been built exactly to the instructions in the magazine at the time.  I think though if I wanted to build a new one I would probably built the AEM6000 with whatever other mods various enthusiasts have suggested on DIYAUDIO.

Annotated Circuit Diagram - showing failed capacitor