Solid state amp repair - distortion problem

[dave cox]

The schematic is too small to read the component designations, can you post something larger? It looks like this amp has some output protection, it's the 2 transistors with their b-e connected across the output emitter current stabilising resistors (just to the left of output darlingtons). If these start to break down they could cause exactly the symptom you see ...

dc

[Biggles]

I personally would compare fault condition voltages with non-fault voltages and also compare scope readings. Particularly around the early stages. I still have a hunch towards the output transistors though, as these get the most stick. Would incorrect output bias be more likely to give crossover distortion than clipping? You need to make certain that the power supply rails are remaining rock steady too. If these are dipping dramatically, that would give exactly those fault conditions I reckon. Very carefully directed freezer spray could help solve this one too.
Alan.

[dave cox]

Guitar amps are not usually stereo!

dc

[dave cox]

I would measure the output stage bias current (no signal) across the two output emitter resistors when the amp is cold / operating normally. Then provoke the fault and repeat the measurement (again, after removing the test signal). This will check the bias current stability.

Given that the amp operates normally at first, and when warm the available output voltage swing is much reduced I think it’s possible that the output protection transistors are limiting the available drive voltage. Either (1) the bias current is going very high (2) the protection transistors are faulty / operating prematurely or (3) there is a short circuit in the load. I have been known to be wrong!

dc

EDIT:
It would be nice to measure the bias current WITH a signal present but it would require a decent differential probe with a high CMRR (common mode rejection ratio).

[Biggles]

Re post #23 Dave...?
Alan.

[McMurdo]

The distortion first starts on the positive peak

Does the distortion alter depending on whether there's a load or not?
Where bias is too low, it normally gets better as it warms up, distortion when cold being the normal symptom.

I've had assymetric distortion caused by all sorts of things such as drivers that can't source enough current, leaky ceramic capacitors causing earthing of the drive, and weird offsets caused by the differential pair drifting so far out of spec that the loop cant balance.

[dave cox]

Once the amp has been on for about a minute or two the sine wave on the output distorts, the top and bottom clip and go square wave.
If I turn it down to 4V rms, 4W it is only at this level or bellow that the sine wave is not clipped any more.

There is no output capasitor on this amp but it's possible the scope was on ac coupling ? I assumed symetic clipping around 0V ...

DC voltages on all semiconductors have already been measured and I compaired one side to the other symetrical side and all are the same on each side, the other side having negative voltages.

I guess this is refering to upper / lower halves of a mono amp ?

It’s a Fender FM212r guitar amp, nasty little thing, the valve amps are so much nicer in sound as well as to work on.

dc

[Al (astral highway)]

Originally Posted by Doris The Diode

I’m good with valve amps and fault finding and repair but this is the first time repairing an amp of this classic transistor nature.

Any tips would be appreciated.

Hi there,

Transistor amps can be a lot more complex to repair than valve amps. It is entirely possible that you will have to change the power amplifier stage transistors, as well as their drivers, as well as all the associated electrolytics. It can get very time-consuming and frustrating.

Maybe do take a good old look at the power supply as an early step. Remove the load and test it to see if it remains as it should be.


Al (Skywave) may be onto something with this thermal protection point. Of course, if a thermal protection device is shutting down the amp, it's because there's too much current passing through where it shouldn't be. So it doesn't fix the problem, just confirms the illness.


I can't read any component values, even with a magnifying glass....

but see where there are two sets of diode voltage dividers, top and bottom of the diagram? What is the quality of the signal into the next stage transistors at this point?

[Skywave]

Originally Posted by astral highway

Al (Skywave) may be onto something with this thermal protection point. Of course, if a thermal protection device is shutting down the amp, it's because there's too much current passing through where it shouldn't be. So it doesn't fix the problem, just confirms the illness.

There is a subtle difference between bias regulation and thermal protection.
I have not considered "thermal protection" as such; bias regulation against temperature variations, yes. And that type of bias regulation is very unlikely to "shut down the amp"; creating an distorted O/P waveform, yes. But if this amp. does indeed have specific thermal protection* (independent of bias regulation against temperature variation) and if that protection is defective, then that bias control cct. will obviously be unable to cope, since the thermal protection cct. will effectively 'call the shots'. In that scenario, there could be two independent faults leading to one problem that subsequently becomes manifest. However, in the case of the problem here, I think that is most unlikely.

* I'm thinking of the type of protector typically found in high-power transistor PSUs (and other applications). It's about the size of a 5p. coin, about 5 mm. thick, usually black in colour with a pair of sizeable lugs for wiring and two fixing holes in a small metal plate for it to be mounted on a heat-sink - or whatever gets hot. Unlike a thermistor, that device usually operates as a switch.

Al.

[mhennessy]

The thermal protection on this amplifier is provided by thermistor RT2, which is a PTC type. When its resistance rises, Q6 is able to switch on, which in turn turns off Q7. This deprives the input LTP current source Q8 of the base current it needs, shutting down the amplifier.

Therefore, a temporary short between the collector and emitter of Q7 will help you decide if this is the mechanism that's causing the problem.

Looking more broadly, the pre-amp section is rather complex with lots of JFETs that can alter the way the circuit works. With that in mind, I'd be troubleshooting by applying the input signal to J4 ("Poweramp In"). You might already be doing that, but I mention it just in case...

Another thing to check: when in fault mode, does the +/-16V op-amp supply remain good?

[Doris The Diode]

Hi everyone,

Thank you for the replies.
I’m away at the moment and will be back at the amp in the new year.
I will post a pdf clear schematic as there was a request for that and also do some well suggested tests and report back.
I haven’t done any tests with the load disconnected, I would have thought that that sort of thing is a no no, something which would blow the PA transistors.

I have in the mean time ordered the PA and driver transistors and will probabley just simply replace those when I get back and see what happens as there have been many suggestions to replace the chain of transistors, which looks like the way to go.

The +16/-16V rail is stable.

When I mention compairing both sides I’m not compairing a stereo amp as this is a mono amp, I mean both sides as in top and bottom of the schematic.
The PCB, when it comes to the output stage in question, is symetrically laid out and is easy to go between components on the positive side of the rail and negative side of the rail.

I will check the bias when cold and hot.

The tip of the probe is connected to the resistive load and the ground is connected to chasis, not directly across the load. The scope is on A/C input.

DD

[Doris The Diode]

PDF of the schematic Fender Frontman 212r

Needless to say, when I put a signal in at TP7 of 5mV I get the 2.05VAC and no distortion.
There is also clean signal up to TP19 of 33mV, after that it’s guess work.

DD

[dave cox]

That PDF is much easier on the eye

NOTE: I think the Fu designations, eg Fu47 means fusible 47R.

A distinctly non-trivial amplifier, now I can see it better!

Bias appears to be controlled by Q14 / Q15.

Q12 is doing most of the voltage gain, with current mirror load Q13 (fed from Q11 with unity gain). There are also 2 bootstraps, C55/56 helping to get the maximum swing.

Q16/17 monitor the emitter current in the output stage and throttle the drive from Q12/13 if the current is too high (If these activate, current is pulled through D32/33 limiting the drive).

Also, there is a current sense resistor (R114/115) in the low side of the output and that ac voltage is ‘summed’ with the main feedback loop (R79/78/77/69 with C44/48 reducing the gain at dc) – I assume this will reduce the gain if a lot of ac current is sensed in the load.

dc

[mhennessy]

On the schematic I'm looking at (from Elekro Tanya, Q11 is acting as a VAS as well. It's almost the classic Hitachi circuit that found its way into the Maplin MOS-FET kit. With the current mirror load, it's a push-pull VAS.

I suspect the current sense at the output is to modify the sound of the amplifier rather than act as a protection mechanism. Most likely, it's altering the output impedance of the amplifier, but hard to be sure without building or simulating it.

I'm waiting to hear what happens when Q7 is shorted

And if that changes anything, the next question will be: is it doing the intended job because the heat sink is too hot, or is it erroneously thinking there is an over-temperature situation?

But if that section is blameless, there are plenty of other rabbit holes to explore

[dave cox]

The previously attached PDF must be different to the one Mark is referring too, as I don’t see a Q7! Also the potential divider R108/107 mean that some serious current needs to flow in R109 to turn on Q16 (~6A)? Also, I think U6-A is a dc servo to adjust the dc level on the ‘inverting’ input to get the output to 0V dc.

My gut feeling is that the fault is AFTER Q9/10. There is so much (open loop) gain in this amplifier that the differential signal level (at Q9/10 base) needed to drive the amplifier to full output is so tiny that having a prematurely clipped signal in the driving source seems unlikely!

I would: -
(1) Check the signal at Q9 input with / without fault. If it is unchanged the fault is in the power amplifier. If so …
(2) Measure the output signal with dc coupling on the scope to detect any dc offset occurring under fault conditions. I doubt this is happening but safer to check.
(3) Measure the bias current (across R109 or R112) with and without fault. A dc voltmeter (with no signal present) is the most straight forward method.
(4) IF Q16/17 ARE turning on (for some reason) then current will start to flow through R103 and D34/35/36/37. Tricky to measure, if you have a 2 channel scope then either side of the Zener diodes should suffice.

dc

EDIT: I see Q7 now, doh.

[dave cox]

Voltage across R81 should be ~10V for normal operation, otherwise the tail current is getting throttled. I can't imagine this is for anything other than protection in a fault condition? The VAS current will follow at 1/2 the LTP current, via Q11 (unity gain) and Q13 (x10 current mirror). So, if the LTP current is 2mA (9.4V on R82), 1mA in each leg, 1mA in Q11 and 10mA in R88, Q12, Q13 & R90. Easy enough to check

dc

[mhennessy]

Yes, see post #30 for my best guess about the action of Q7 et al. While fishing about for ideas about what could cause symmetrical clipping, that one seemed reasonably plausible. Also, I did suspect U6-A (I agree that's it's a DC servo in some form), but that seems rather less likely given that the components around it are low-power passives that lead a quiet life.

At this stage, all other possibilities I can think of would cause unequal clipping. Despite appearances, this is not a particularly symmetrical design. There are also a few touches that are unusual to my eyes, as I'm used to hi-fi rather than musical instrument amplifiers. It makes an interesting change though

[dave cox]

It may be the case that one of the design goal is an amplifier that degrades progressively (by premature clipping) as it overheats rather than just ‘switching itself off'. To a musician, this would be quite important in a performance! If they had just reduced the gain as the amplifier overheated (assuming you manage to solve any resulting stability issues) then someone might just turn the volume up to compensate so the result would still be catastrophic failure!

Messing with the standing current must have made getting this amplifier stable a right PITA
dc

[Doris The Diode]

Hi everone,

Thanks again for the replies.

I'm back after 2 weeks off and in the mean time I have replaced the two PA and driver transistors which has helped slightly as I can drive it to 7V instead of 6V without clipping, still not much use!

I'm not concidering Q7 as the problem is now with only one side of the sine wave.

I measured across R109 & R112. With no signal the voltage across them both is 4.7mV.
With the sine wave just before clipping they are both equal in voltage.
When driven to 19.5Vrms across the load, R109 has 455mV across it and R112 has 553mV across it, around 100mV difference.

Photos attached show a clipped wave form at 19.5V which is 95W into a 4 Ohm load and only the top is distorted, the bottom is clean, this has now changed since the first diagnosis.
The other wave form is just before it distorts showing a clean wave form at 13W.

I also meassured the thermistor and it showed 180mV across it cold and rose to 330mV when driven just before distortion for about 10mins. The PA got hot and so did the mains transformer, this amp cannot possibley do 100W continous to answer a question above.

I also took DC measurements cold on top and bottom of the diagram and cold they were the same to within a few mV. When driven just before distortion both sides also had the same voltages.
When the output was driven to 50W out and well into clipping the top and bottom voltages differed significantly, as exptected.

So, I'm still very stuck and will more than likely just replace the 'top set' of transistor first to see if there will be a difference. This is based on the fact that the top of the diagram is responsible for the positive going peak of the signal, this is what I’m presumimg.

DD

PS The scope is a cheep one from a distant country from the Easterly direction and at this scale is not showing a nice sine wave. On other scales it is better.

[Radio Wrangler]

A couple of points I noticed in the circuit is that the VAS has diode clamps to keep it out of saturation.

It relies heavily on bootstrapping... I wonder about the bootstrap caps and the resistors it drives. They can cause weird trouble.

David