Diagnosing and fixing amplifier fault (loud crackling and popping)

Hi,

I've got an amp here for repair, which produces loud crackling and popping in the right channel, ocasionally. It has been doing this for years (5 or perhaps even more) sometimes, but it is getting worse now. I've already determined the fault exists somewhere in the poweramp, right side (left and right are on seperate PCB's).

I know this can be caused by degrading electrolytic capacitors, but tests would seem to indicate that this is not the cause.

My main question is: can transistors, and perhaps MOSFETs, be damaged in such a way that they will sometimes short or go open?

Also, perhaps somebody has other suggestions as to what it could be?
 
Is there a reason why your guess is in the differential stage, and not the voltage or current amplification stages?
It definitely could be the VAS too. Probably not current gain stage. Probably a stage where voltage gain is happening. I've repaired a lot of noisy amps and 9 times out of 10 it's the diffamp. If it turns out not to be the diffamp, it still sounds like front-end. maybe current source for the diffs, etc. I'd try spraying the diff transistors in isolation (Through a drinking straw, etc) and let them come back to temp, checking if they get noisy with the temperature transistion. Be careful when testing JFET front end's. If the input isn't loaded with a decent resistance (47 kohm, or so), the freeze-spray moisture might cause enough leakage to make the DC move around quite a bit.
 
Tap the offending board and see if it makes a noise. This may smoke out a problem with an intermittent solder joint. Once in a while, one will encounter a problem with an intermittent resistor as well (loose end cap, for example). I had one of those that drove me nuts.
 
Administrator
Joined 2004
Paid Member
Hi halfgaar,
You have some good suggestions here. I tend to think the noise is in the diff amp or Vas. Possibly a current source - maybe.

Diff pairs can be reverse biased. This is the best way to make a low noise transistor noisy. Don't forget the dies on these are smaller and therefore more fragile possibly.

-Chris
 
One other suggestion - don't run the the problematic channel into a loudspeaker, as troubleshooting may cause the amp to produce a fair amount of power output (I.e. a loud crackle, followed by golden silence is not nice!!) .

An incandescent bulb at the output may suffice to allow a visual if the channel in question pops/cracks. Or, a sacrificable resistor of a few ohms, 1W rating in series with the loudspeaker at least.

Cheers!
 
(first, I really wonder why on some forums I only occaisonally get e-mail notifictions. It's not trapped in spamfiltering, so I really don't understand where it could have gone (this time)).

Check also the solders point.

I reheated all the solder joints, adding some fresh solder to them. This does not fix pads which are coming off the PCB, so I will try the suggested tapping with a plastic tool. The only problem is that I can't reach the underside very well when the amp PCB is plugged into it's socket...

I did test all the pads BTW, with a multimeter. They seemed to be OK, but of course, intermittent failure is hard to detect this way.

If you do not have a multimeter whith a capacitor leakage measurement,you cannot measure well.
I can advise the substitution of the suspect cap.

I indeed don't have that.

I have taken out some of the caps, and when testing, they did retain their charge quite long. Waiting for them to discharge took too long. Also, there is no cap directly in the output path (at least, none that I can see), so a faulty cap is not what strikes me as the most probable.

What is the model of your amplifier?

It's an Erres HiFi sound project TA 8000. An oldie, I guess about 20-25 years.

The output stage is one I can't identify, since it makes use of two identical power transistors (or perhaps MOSFETS) per channel (typenr C1051).

It definitely could be the VAS too. Probably not current gain stage. Probably a stage where voltage gain is happening. I've repaired a lot of noisy amps and 9 times out of 10 it's the diffamp. If it turns out not to be the diffamp, it still sounds like front-end. maybe current source for the diffs, etc. I'd try spraying the diff transistors in isolation (Through a drinking straw, etc) and let them come back to temp, checking if they get noisy with the temperature transistion. Be careful when testing JFET front end's. If the input isn't loaded with a decent resistance (47 kohm, or so), the freeze-spray moisture might cause enough leakage to make the DC move around quite a bit.

The problem here is that I can't be sure the VAS transistor is the one I think it is. There are only two medium power transistors (TO-126 package), and those are the output drivers. There is only one that can be the VAS, but it's an ordinary TO92 (typenr C1439). And if this is the VAS, there must be a passive current source for the diff stage, since there is no transistor left for it.

Tap the offending board and see if it makes a noise. This may smoke out a problem with an intermittent solder joint. Once in a while, one will encounter a problem with an intermittent resistor as well (loose end cap, for example). I had one of those that drove me nuts.

I have been tapping and shaking the board itself, but not individual parts of joints, as described above. The tapping of the board didn't change anything, but further testing is required.

Yeah, I don't know what the failure mechanism is with differential amps either. I don't know if it's silicon contamination or bond wire fails, but it definitely seems to follow "low-noise" transistors.

.....for the most part.

So it can happen that transistors fail without being completely bust? I guess so, judging by the suggestions so far :)

One other suggestion - don't run the the problematic channel into a loudspeaker, as troubleshooting may cause the amp to produce a fair amount of power output (I.e. a loud crackle, followed by golden silence is not nice!!) .

An incandescent bulb at the output may suffice to allow a visual if the channel in question pops/cracks. Or, a sacrificable resistor of a few ohms, 1W rating in series with the loudspeaker at least.

I have a test speaker for that, for which no one will shed tears if it's bust :) But you're right, the crackles can get quite loud. Sounds like firecrackers often...

Anyway, I have a lot to go on now, thanks.
 
Administrator
Joined 2004
Paid Member
Hi halfgaar,
The outputs (C1051) are most likely 2SC1051 Japanese parts.

The load for the Vas and tail current generator are probably just resistors. The Vas may be a TO-92 part.

Tapping a transistor that is "cracking" normally will not show a problem. Changing the temperature is a better way. Freeze spray and wait (give it time to work) or heating with your iron. Don't overheat the part. I'm still thinking differential pair.

-Chris
 
Changing the two LTP transistors is of course quite easy. Now to send out the owner of this amp to get some parts which are identified as:

T.7B
A942
GR

I can't find a datasheets for it, so if they're not available anymore, I hope alternatives are. And otherwise, how critical is the transistor type in the diff stage, concerning exchangability with other types? Perhaps I could simply use BC546 or BC556. I used those in the input stage of my home made amp, and it works fine, sound quality wise.