Trace Elliot Commando Boxer 65W - need expert help is it oscillating ?

[Steve Conner]

I agree, it is a terrible design. It's actually worse than the architecture Teemu is thinking of.

But to quote another Enzoism, we're here to repair things, not re-engineer them. Given that attitude, we have to assume that it worked when it left the Trace factory.

Of course it is just that, an assumption. Trace Elliot did go bust after all.

If we decided to start re-engineering, then I agree that sticking the driver transistors and bias diodes to the heatsink with thermal paste was a good start.

[chazpope]

bu**er.. I just restored the driver emitter resistors back from 100 Ohms to factory state, thinking that somehow the cooking problem must go away along with the other one.

I'd rather have it in factory state really and give it back to owner, but I don't want it to blow after 10 mins of use. Steve I was talking about another amp when I said I fitted heat skins - here is how this one is - you can't do that easily. I think the drivers cook at about 90C and they will probably blow, I have not let it run for more than a minute so far.

[soundguruman]

There are plenty of manufacturers these days that make amplifiers, to last long enough to survive the warrantee period, after that --POOF.
JUNK IT and accept the fact that it's not worth fixing.
YOU do not want to be responsible, in the future, when it blows up again, for the same reason, or another reason.

[J M Fahey]

I almost miss the spambots.
Their advice is always as useless as this one above, but at least they are shorter.
And they always agree/congratulate/whatever, but in general spread good karma
No danger of that happening here.

[chazpope]

ok so I did Fahey's mods, the drivers are still cooking though at half power, output transistors are cold during standby. I played it loud for 30 min - it did not blow. There is some new 'buzz' added to played notes, must be due to shifting the operating point. But I guess this won't be heard over the smash of cymbals in a live gig - actually this amp belongs to a drummer

Thanks all who helped! This was a weird one, at one point I considered putting a 3886 in but it would not take the voltage. I hope to have seen the last of its gubbins and plan test it at pub jam this weekend. May it live long enough to sell

[J M Fahey]

Cool !!
Congratulations.
The problem is that unfortunately this amp (like many Fenders) has no adjustable bias, but 4 diodes which, in theory, *should* match the 4 BE junctions (2 per Darlington) they must *just* turn on.
In practice it's not so, or too tight for comfort.
And pulling 2 of them would have been brutal.
The resistor values I suggested both lower driver dissipation and bias the Darlingtons with the equivalent of "1 1/2 diode"
Yes, you will have some crossover distortion, but in actual stage or rehearsal use it's bearable.
Not so in a bedroom late at night playing a piano.
Oh well.

[chazpope]

sorry to report this but it did blow up in the end We put it in its cab box and played it at a live jam at about medium volume for about 45 minnutes. It played and sounded very well - and then suddenly it died ... I would say this was 'light use' power-wise with breaks every 15 mins to change the line up, however heat build up in the closed cab probably killed the tiny drivers cooking at about 0.5W and that hammerred the rest. I went as far as to plug it in through a bulb and it looks like a complete short circuit again.

Now I can beat myself that maybe I could have reengineered the drivers completely by moving onto the large heat sink and conecting with wire extensions to the pcb - but I thought halving their factory set power consumption level would be enough - sadly it wasn't.

[J M Fahey]

Oh sh*t.
Not much can be done then, I think your friend has just got a nice extension cabinet for any other amp or head he has. Oh well.

[soundguruman]

sorry to report this but it did blow up in the end We put it in its cab box and played it at a live jam at about medium volume for about 45 minnutes. It played and sounded very well - and then suddenly it died ... I would say this was 'light use' power-wise with breaks every 15 mins to change the line up, however heat build up in the closed cab probably killed the tiny drivers cooking at about 0.5W and that hammerred the rest. I went as far as to plug it in through a bulb and it looks like a complete short circuit again.

Now I can beat myself that maybe I could have reengineered the drivers completely by moving onto the large heat sink and conecting with wire extensions to the pcb - but I thought halving their factory set power consumption level would be enough - sadly it wasn't.

A 150 CFM duct blower may do the trick.
http://www.amazon.com/Tjernlund-Duct.../dp/B0000AXFP3

[chazpope]

hm, contrary to what I expected, excisionary post mortem showed that the drivers are quite alive, while T1-3 are cooked and T4 is not. Here I give up trying to make this crap schematic work...

But I still intend to 'fix the amp' - by deleting all to the right of RV4 and plugging a new amp with LM3886 there.

So I will have to shut down IC2 somehow and also lower the psu to +-30V, so next I will see how I can unwind the transformer

[J M Fahey]

Forget unwinding the transformer. Can of worms.
A TDA7294 will work with your PSU and stand it.
What's your exact +/-B voltage?. Not indicated on schematic.
Load it with 8 ohms only; 4 ohms will be too much, so forget extension speakers.
Problem is the "all in a PCB" construction.
You *can* pull all Power amp parts but please post a couple full chassis pictures, to see what space you have available.
Absolute worst case, you can drill a couple 1/2" holes to let wiring through and mount a power amp *inside* the speaker cabinet, like Polytone does.
Having that much space will allow you to mount any discrete power amp you wish.
I have done it countless times to beef up good sounding but small amps, but itīs an option to save an unrepairable one too.

EDIT: at least you need to pull R21, C8, R20, and jumper C12.
That way, Ic1B becomes a unity gain buffer for the signal present at RV4 (master?) and its pin 7 becomes the preamp out to drive the new power amp.
Then remove everything to the right of the Leds plus R18 , C18.

[chazpope]

it is +-50V and over so it won't do without unwinding ? it needs to go down to 35 - too big a drop for a ic regulator (+cost). But I haven't tried to dismantle at the transformer yet - it's been 35 years since I last did that - it may be trouble

it is a 6 Ohm load and it needs to be 65W, it doesn't look that the TDA is any higher voltage than LM3886 and this one seems to be less parts to make and can be bolted on chassis. Plenty of room in the box.

[J M Fahey]

+/- 50V is crazy high.
Real world more than 100W into 8 ohms.
For 65W RMS it's around +/-40 , which is fine for a TDA7294.
Doesn't look that big a difference, but it is.
Where do you live?, what's the line voltage?.
I said forget unwinding because modern transformers use epoxy impregnation , you won't be able to separate laminations and *if* you can, unwinding will tear the wire enamel.
Rewinders "burn" the old transformer in an oven to burn epoxy, recover laminations and be able to measure wire and count turns for a total rewind with new wire and bobbins, but I don't think you want to go that far.
Worst case, you can add a 12V 2A transformer with reverse phase to lower line voltage from 120V to 108V , that's probably enough. Or a 24V 2A to lower it to 96V.
Or maybe you can get a cheap surplus transformer for peanuts.

[chazpope]

yes 50 is high, hence the need to unwind.

UK mains is 240V, traf disassembled, as usual had to separate the lamelae one by one with a stanley knife, wire is .4mm, a bit thin I would say. adding another lump transformer to lower voltage is in my view a quirky idea, hope to find cheap wire next or I can use the old one

[J M Fahey]

Well, you got lucky, not much of impregnation there, maybe some varnish only.
Anyway I wouldn't trust recovered wire insulation.
0.40mm is fine as primary (240V) wire, you will find 0.70 to 0.80 mm used in the secondary.
You can mount an extruded finned heatsink over that ventilation grid cut in the chassis bottom, with the chipamp mounted on it of course.
If you use an LM3886 do not go over +/-32V; with a TDA7294 you can use *up to* +/-42V.
Good luck.