Solid state amp repair - distortion problem

[dave cox]

PS, D30 / 31 are Schottky diodes, even though they are drawn as Zener diodes!

That makes more sense as they will protect the collector (at least to 30V!) and still (barely) leave enough headroom for the bias transistor to operate (V b-c ~ 350mV).

The out of balance LTP could account for the fact that the clipping is asymmetric. A quick calculation gives the input offset at ( 740mV-210mV ) divided by LTP gain. This gives a 53mV offset at the LTP input. This will be correcting a ‘notional’ output offset that is 53mV times the amplifier gain. That does seem quite large ( if I knew what the gain was ) but there are also component variations that come into the equation.

dc

[Doris The Diode]

Tests done.

There is no change in voltage across D20, a solid 10.1V under all conditions even when heated up by 15W for a few minutes to the point where the whole chassis was warm to the touch. It might need a bit more than that I expect to change state.

Dave,
I measure from stone cold i.e. first switch on was just now today and the voltages across the resistors were measured immediately after switch on.

R87 425mV
R88 345mV
R89 340mV
R90 360mV

The above readings were within literally 20 seconds of switch on.
Then the voltages started to drop rapidly to he point that I couldn't measure them as the change was vast between each reading.
After about 3 min the voltages stabilised and here are the readings.

R87 0mV
R88 0mV
R89 0mV
R90 2.5mV

Now there is a clue for someone who knows a bit more about these amps.

With a signal applied showing 8Vrms across the load, these are the DC readings:

R87 103mV
R88 53mV
R89 96mV
R90 42mV

All above voltages are DC.

Perhaps Q13 is leaking/faulty as why would it be conducting if there is no voltage drop across R87 & R89? Current could be flowing from top through R91, 92, 94, 95 and through the leaky Q13 junction and producing a voltage across R90 when at idle, no signal?

Yes, well earned Brownie points and a few beers too!!

All the best,
DD

[Silicon]

It's interesting that D35 and D37 have 100mV across them.

What kind of diodes are they?

When you give us voltage readings you don't say whether they are AC or DC readings.
You force us to make assumptions. It may seem pedantic but it has relevance.

When you stated that D28 had 40V across it, you didn't state the polarity.
If the cathode is 40V more positive than the anode, the diode is probably good.

However if the anode is 40V more positive than the cathode, the diode is probably open circuit.

[mhennessy]

OK,

I'm looking at the middle sets of results - the ones showing lots of zeros...

For all the world, the voltages across those resistors suggest that the tail current is being taken away by something - even if it's not the thermistor.

Could you get it back into the same state, such that there is 0V across R87 and R88, and measure the voltages across R83, R84 and R82?

Thanks in advance,

Mark

[mhennessy]

Originally Posted by Silicon

It's interesting that D35 and D37 have 100mV across them.

That's not a worry

There should only ever be a voltage across those diodes when the output current reaches a certain level - look at how they are across the 0.22R emitter resistors. The exact arrangement is a bit strange, but looks like a form of dual-slope limiting.

[dave cox]

Exactly as Mark says !

Even the first set of reading look a little on the low side ...
It's not really possible to measure dc withsignificant ac present, even the best meter will be telling fibs!

dc

[Doris The Diode]

Here we go, readings for R83, R84 & R82.

All readings are DC, to make it clear as pointed out rightly by Silicon.

The amp was switch on again and left for 5 min so that the voltages 'stabilised', no signal, amp cold.

R83 1.6mV
R84 0mV
R82 0.2mV

I then put a signal through for a minute which gave 7Vrms across the 4 ohm load which showed no distortion.
I then took the signal off and measured the resistors again out of curiosity and this is what I found:

R83 875mV
R84 74mV
R82 9.6V

I waited another 5 minutes and measured the voltages, no signal and they all remained the same.

DD

[mhennessy]

Weird!

The first (cold) reading make no sense. With 0.2mV across R82, Q8 is not providing any operating current to the long-tail pair (the lack of voltages across R83 and R84 confirm that). That is an amplifier that is switched off. For this to be the case, either there was 0V at the base of Q8, or Q8 is faulty.

But you were able to put a signal through it. After you'd done that, the next set of readings seem quite reasonable. We can see a sensible 9.6V across R82, which tells us that the current source is providing 2mA, and that 2mA is being split across the two collector resistors (albeit unequally).

Those cold readings seem completely spurious to me. I think I'd want to get to the bottom of this before doing anything else. They should be pretty much constant under all combinations of time and temperature - only changing when the thing has definitely over-heated.

I'll give this some thought...

[dave cox]

I think D35/36/37/38 are there to stop the VAS from swinging too far above / below the output (speaker) voltage. At 5.7V above (or below), Zener D34 (D36) conducts VAS current, via R103, into R107 (R110) making Q16 (Q17) turn on throttling the output even more. I don't think they are involved in this fault as it appears the VAS is just lacking enough push and pull in the first place !

dc

[Doris The Diode]

This is becoming a headache and it just might be quicker to replace the transistors.

I've had a look online with not much success, found a few only but has anyone got a book/DVD and are able to look up these equivalents please as some transistors are hard to come by or are ridiculously expensive, perhaps European equivalents?

2SC2362K
2SC2383
2SA1016K
2SA1013
2SD1857
2SB1236
2N4001
2N4003

Regards,
DD

[dave cox]

I would avoid doing any wholesale changes with a fault like this.

It is just as likely to be a passive part that has failed, and every piece of work you do (or anyone else for that matter!) carries a finite risk of introducing a defective part or installing it incorrectly, bridging tracks etc. etc. Having two faults is SOOOOO much harder!

Don't forget it's 'SO NEARLY' working and its not blowing the fuses or emitting smoke !

Go back and re-measure the dc conditions on R82/83/84 as the originals make no sense. Stick with it, if it was easy anyone would be able to do it

dc

[mhennessy]

I agree with Dave

We all know the pain with jobs that "linger", resisting all attempts at persuasion, but we've come so far, and it would be such a shame to not discover the root problem.

[Doris The Diode]

Yes I know what you mean, the chase is exciting and I'll stick with it.

So, today's DC measurements across R82/83/84 are, voltmeter had been on for 30min, no signal:

Within 10 seconds of switch on:
R82 9.5V
R83 805mV
R84 105mV

After 30min
R82 9.6V
R83 793mV
R84 133mV

DD

[dave cox]

Thats more like it!

Still well out of balance and consistent with running out of steam on +ve going output. I can't be sure that 'normal' component variations could cause that degree of imbalance but it seems unlikely. Assuming a fault is causing it ...

Q12 will be passing higher than normal current (same voltage on R88 as R83 ?) just to get the output to 0V (check!). Where does that extra current go ? R103 is a possibility - in normal operation I would expect to see virtually nothing passing there. Assuming nothing into R103 then either the 'lower bias half' is pulling to much current or the upper half not supplying enough ...

For the upper half R91/R92 supply current and R102 is sinking it away. For the lower half R99/R100 are sinking and R104 is supplying. The difference between those 2 halfs it what the VAS has to do.

dc

[Doris The Diode]

Chaps,

Update, I replaced Q8,9,10 and the problem is still there.

DD

[Doris The Diode]

Fixed!!

I decided to desolder the remaing transistors in the dc amp chain and measure the hfe.
All transistors showed the expected hfe but one transistor's hfe kept on decreasing as time went by a test of 30min.
The other trasistors stayed stable but this little blighter's hfe decresed.

Q11 was the problem.

I bought a new transistor, soldered it in and the rest and did extensive testing on the amp and all is well.

Thank you all for your efforts, suggestions and the will to live with this solid state monster.

I can now go back to my valve projects and enjoy life once again ��

All the best untill my next thread.

DD

[mhennessy]

Well done

And that is consistent with the reported conditions around R87/88, etc as well. It's good when it all makes sense at the end of the day.

[Doris The Diode]

I thought I’d test the amp again just to make sure and guess what, my TG200DMP oscillator stopped working!!! Why me!!!

I just wanted to get back to my valves and now I’m going to have to check more transistors, booo