Help with DC on volume pot

[Mooly]

We have come this far so lets try and nail this if possible or at least try and understand why.
(if you use solder braid it's easy to isolate parts as you just need desolder one leg making sure it's "free" in the hole in the PCB)

1. Leave C518 removed

2. Also remove C530

Doing this totally isolates the filter stage. Also for safety disconnect the speakers.

Now repeat the scope measurement across R529 and also measure across R530. Any spike or jump as switch operated ? Anything that appears across R530 is passed directly to the power amp.


If the problem is still there we have to isolate things further. Working on the theory that the only way the switch can be noisy is if there is an abrubt change in voltage (even if tiny) as the switch is operated lets isolate some more areas of the circuit. We are going to isolate the switch from points in the circuit that have voltage.

3. Remove C522

4. Remove C524

Now test again across R529 and R530.


If there is still a spike and jumping then we are looking at something odd such as electrical leakage on the PCB but we have to prove it.

5. Remove R538

6. Remove R540

Retest again.

It's also worth checking R530 and R536 to make sure neither is open circuit... I'd be amazed if they were but we can't assume anything now.

 

[mkusan]

We have come this far so lets try and nail this if possible or at least try and understand why.
(if you use solder braid it's easy to isolate parts as you just need desolder one leg making sure it's "free" in the hole in the PCB)

I agree and I am willing to continue

Your advice about desoldering only one leg is excellent, this will save me a lot of time.

I am using cheap car speakers for tests which I don't care about so if you want some of these test to be checked on them there is no problem to do so.

About checking resistors R530 and R536, can I measure them in circuit and if they measure higher than they should I can be certain they are open or have gone up in value. Or the only meaningful test is to desolder them and measure them out of the circuit, in which case I would rather just replace them with the new part?

I did not mention, and maybe it is important, that the spike I saw was faced downwards, voltage drop perhaps?

 

[Mooly]

Best checked with one leg desoldered. DVM's do starange things on resistance measurements if there is even a tiny voltage (charge on caps etc) present.

The spike polarity just shows the switch pulls the voltage down momentarily at the point of measurement. It could have been the other way round Just the way this problem is.

 

[mkusan]

We have come this far so lets try and nail this if possible or at least try and understand why.
(if you use solder braid it's easy to isolate parts as you just need desolder one leg making sure it's "free" in the hole in the PCB)

1. Leave C518 removed

2. Also remove C530

Doing this totally isolates the filter stage. Also for safety disconnect the speakers.

Now repeat the scope measurement across R529 and also measure across R530. Any spike or jump as switch operated ? Anything that appears across R530 is passed directly to the power amp.

It's also worth checking R530 and R536 to make sure neither is open circuit... I'd be amazed if they were but we can't assume anything now.

After removal of C530 there is no more jump across R530, same for R529. Just a beautiful straight line

I checked and even replaced R536, but the old one checked OK out of circuit.
Here is what I don't understand with my limited knowledge, why R535 in circuit measures 324 Ohm, R529 99 Ohm and their RIGHT siblings R530 measures 76 Ohm and R536 77 Ohm. When the R536 was out I measured R530 again and then it was showing 100 Ohm. With new R536 it went back to 76 Ohm. I understand that measuring resistance in circuit is not the right way to do it but shouldn't measurements between channels be similar?

 

[mkusan]

After removal of C530 there is no more jump across R530, same for R529. Just a beautiful straight line

I checked and even replaced R536, but the old one checked OK out of circuit.
Here is what I don't understand with my limited knowledge, why R535 in circuit measures 324 Ohm, R529 99 Ohm and their RIGHT siblings R530 measures 76 Ohm and R536 77 Ohm. When the R536 was out I measured R530 again and then it was showing 100 Ohm. With new R536 it went back to 76 Ohm. I understand that measuring resistance in circuit is not the right way to do it but shouldn't measurements between channels be similar?[/QUOTE


Let me rephrase my question with those values being of course Kohms, not Ohms.

 

[Mooly]

The scope readings are a puzzle...

So with C530 removed there is no jump or spike ?
And the only components that are isolated for that is C530 and C518.
Everything else connected ?

Leaving it like that (confirm again no jump across R530).
Now measure with scope to the point where C530 and R552 and VR502 join. Leave the scope ground where it is (point you used for measuring across R530).

Is there any spike at this point when switch operated ?
I can't see how there possibly could be unless there were some leakage but best to measure and discount it.

If no spike then it really is a puzzle.

---------------------------------------------------------------------------------------------------------

You are getting different readings in circuit because the DVM isn't just seeing the resistances around where you are measuring. The slightest residual voltage (from PSU caps charged up etc) cause tiny in circuit voltages which totally corrupt the DVM reading.
It also "sees" the semiconductor junctions and these vary from device to device.

It's absolutely normal what you are seeing. Also, although polarity of the meter leads has no effect on resistance measurement it does when reading in circuit because the meter applies a small test voltage (thats how it measures resistance) and that voltage is altered by components around it. Caps have different leakage values too that alter readings. So it's all absolutely normal and the reason why we never measure as a general rule in circuit.

 

[mkusan]

The scope readings are a puzzle...

So with C530 removed there is no jump or spike ?
And the only components that are isolated for that is C530 and C518.
Everything else connected ?

Leaving it like that (confirm again no jump across R530).
Now measure with scope to the point where C530 and R552 and VR502 join. Leave the scope ground where it is (point you used for measuring across R530).

Is there any spike at this point when switch operated ?
I can't see how there possibly could be unless there were some leakage but best to measure and discount it.

If no spike then it really is a puzzle.

With C530 removed there is nothing visible on the scope while operating low filter switch, my wife can testify to that because she is the one operating it since I have both hands full with the probes.

Only components that are isolated are C530 and C518.

However, I will double check those readings again tonight.

Only difference is I replaced R536 before I measured with the scope. But I repeat, old R536 showed declared resistance out of the circuit.

If you remember I wrote the popping decreases in intensity as the unit is getting warmer, can resistors be thermally unstable?

Maybe we solved the problem already, but we are just not aware of it?

 

[AndrewT]

as the unit is getting warmer, can resistors be thermally unstable?

Yes, most components will reveal some kind of flaw if temperature is swept across a wide range. Sometimes extreme cold sometimes extreme hot and sometimes just a little change in temperature.

 

[Mooly]

Maybe we solved the problem already, but we are just not aware of it?

Stranger things have happened

We musn't lose sight of the theory in all this and that is that the switch makes a noise because there must be a potential difference across the contacts. That difference causes a transfer of charge when the switch is operated and you hear it as a click. And because the preamp is connected directly to the power amp after the main volume control then the slightest voltage spike is heard clearly as a click because it gets the full amplification of the whole amp working on it.

For the effect to get less as the unit warms implies that the potential difference across the switch, however small it may be is reducing.

So another possible area to look at is that the DC conditions are "slow" to stabilise. We have isolated the stage around Q506 and it is silent, no jumping.

What if (I know I said I would be amazed if this were the problem but thinking about it...), what if the DC conditions around Q508 and Q504 are "slow" to settle to a final value. If we measure the DC voltage at the collector of each of these transistors it will drift as the unit initially warms. That rate of change of DC will be passed (however small the amount might be) into the stage around Q506.

It's very hard to prove exactly though as the effect is so small. Even a very tiny leakage current through any cap could cause a small shift in DC conditions that settle as the leakage reduces towards zero.

That could perhaps be proved by swapping parts with the good channel although thats not really a method I favour.


Yes,as Andrew says resistors can be thermally unstable although it's perhaps more an issue with old carbon composition types.

 

[Mooly]

Measureing the DC conditions could be something you could easily do.

Four measurements begining from cold measured with DVM. (Must be from cold initially)

Q503 collecter and Q504 collector.

Q507 collector and Q508 collector.

Write the results down for each and then repeat measurement every minute or so for say 10 minutes and see if the any one shows more change than the other channel twin.

 

[mkusan]

Measureing the DC conditions could be something you could easily do.

Four measurements begining from cold measured with DVM. (Must be from cold initially)

Q503 collecter and Q504 collector.

Q507 collector and Q508 collector.

Write the results down for each and then repeat measurement every minute or so for say 10 minutes and see if the any one shows more change than the other channel twin.

Will do those measurements, do you want C530 and C518 to stay isolated?
Any special ground point I should use for testing?
If I see nothing out of the ordinary there, can I then return C530 and measure with scope again across R530 to see if it is still silent while operating switch?

 

[Mooly]

Yes, leave them isolated for the measurements. Just use the same grounds as you have been doing. As long as you are consistent and keep using the same ground for each transistor measurement it will show the relative drift as it warms up.

Of course try it all again afterwards.

 

[mkusan]

Measureing the DC conditions could be something you could easily do.

Four measurements begining from cold measured with DVM. (Must be from cold initially)

Q503 collecter and Q504 collector.

Q507 collector and Q508 collector.

Write the results down for each and then repeat measurement every minute or so for say 10 minutes and see if the any one shows more change than the other channel twin.

I measured all of the above with C518 and C530 still disconnected, from cold every minute for 10 minutes.

Q508 and Q507 tracked each other perfectly.

Q508: 10.11-10.11-10.11...
Q507: 10.13-10.12-10.11...

Q504:13.36-13.46-13.53-13.60-13.66-13.70-13.75-13.78-13.83-13.88-13.92
Q503:12.74-12.82-12.89-12.95-13.01-13.03-13.08-13.12-13.14-13.18-13.25

I measured again with scope across R530, no spike.
Then I connected C530, again measured with scope, no spike.
At that moment I thought we nailed it, R536 was the culprit but how naive I was.
I reconnected C518 and my test speakers, low filter in-out, we are popping again.

I will be on a trip tomorrow, so I will not be able to post here or do late night soldering, but Friday is business as usual

Mooly, is this my next course of action?

"Now measure with scope to the point where C530 and R552 and VR502 join. Leave the scope ground where it is (point you used for measuring across R530)"

 

[Mooly]

Nothing out of the ordinary with those readings. It's normal to see a drift like this. If it drifted a few volts then there would be a problem but not as measured here.

Yes we now measure that junction with C530 isolated.

 

[mkusan]

Nothing out of the ordinary with those readings. It's normal to see a drift like this. If it drifted a few volts then there would be a problem but not as measured here.

Yes we now measure that junction with C530 isolated.

Hi Mooly,
and sorry for not being available yesterday.

Measuring across R530 with C530 in place and isolated I can see no difference between the two.

C530 isolated, I measured between ground of R530 and the junction where C530, R552 and VR502 join, no moving on the scope while pressing or depressing low filter switch.

I was contemplating about my problem yesterday, and that LEFT channel that is now silent is a puzzle to me. I remember vividly that LEFT channel was exhibiting the same behavior, maybe even more so than the RIGHT.
The only explanation I could think of is that it stopped popping and went unnoticed after one of my most recent changes, either:

a.) change of transistors in tone board
b.) change of C501 and C502 to film caps, ceramics 300pF were in place originally

Let me know what you think of it.

Thanks
Marko

 

[Mooly]

Hi Marko, no problem and thinking time is good

Not proving easy is it this one... however,

C301 and 302 can't really have an effect tbh. We isolated everything on the input side by removing C506. Those caps in any case are in a part of the circuit that is all referenced to ground and with the loudness switch OFF then they are disconnected anyway.

Hard to see how transistors could cause this effect too as the DC conditions should be stable (which we have just measured) with any decent small signal type fitted.

If nothing is showing on the scope even with the scope set to its most sensitive then have you actually listened to it again. Is the noise still there ?


Another idea that surely has to prove something is this, as we are struggling to even identify just where the problem is happening.

Look at the picture here.
We are looking at both channels now and splitting the circuit into the three distinct stages that make up the preamp.

I am thinking we could "cross couple" the stages in turn to see where the problem is which is easier than it sounds. As an example, that means isolating C529 and C530 (left hand "legs" on diagram) and then swapping the feed into the caps over. If the noise moves to the other channel it proves its being generated in the part of the circuit before the part in blue.

If the fault does swap over to the other channel we then put those caps back as they were and do a similar swap around with C517 and C518.

 

[mkusan]

Hi Marko, no problem and thinking time is good

Not proving easy is it this one... however,

C301 and 302 can't really have an effect tbh. We isolated everything on the input side by removing C506. Those caps in any case are in a part of the circuit that is all referenced to ground and with the loudness switch OFF then they are disconnected anyway.

Hard to see how transistors could cause this effect too as the DC conditions should be stable (which we have just measured) with any decent small signal type fitted.

If nothing is showing on the scope even with the scope set to its most sensitive then have you actually listened to it again. Is the noise still there ?


Another idea that surely has to prove something is this, as we are struggling to even identify just where the problem is happening.

Look at the picture here.
We are looking at both channels now and splitting the circuit into the three distinct stages that make up the preamp.

I am thinking we could "cross couple" the stages in turn to see where the problem is which is easier than it sounds. As an example, that means isolating C529 and C530 (left hand "legs" on diagram) and then swapping the feed into the caps over. If the noise moves to the other channel it proves its being generated in the part of the circuit before the part in blue.

If the fault does swap over to the other channel we then put those caps back as they were and do a similar swap around with C517 and C518.

Hi Mooly,
and thanks for your explanations.

Initially I did not try speaker test with C530 removed, but I did after your post and the noise is not there.

I believe that earlier when we were performing tests and when C518 was removed I tried on speakers and there was also no noise.

If I understand you correctly that puts origin of our problem into the red (coincidence?) area, between C506 and C518.

Do you want me to start from there with this new info?

 

[Mooly]

It gets confusing doesn't it
Lets just be sure first before continuing.

If C530 is fitted and C518 is not and the noise is not there then the problem seems to lie to the left of C518.

Removing C530 isolates most of the preamp on that channel.

We are working from right to left on this adding more stages to the chain and hoping to find which stage is reponsible.

So lets just be sure where we are up too and then carry on as outlined above.

 

[mkusan]

It gets confusing doesn't it
Lets just be sure first before continuing.

If C530 is fitted and C518 is not and the noise is not there then the problem seems to lie to the left of C518.

Removing C530 isolates most of the preamp on that channel.

We are working from right to left on this adding more stages to the chain and hoping to find which stage is reponsible.

So lets just be sure where we are up too and then carry on as outlined above.

It does get confusing, especially with new toy (scope) in the equation, and a bad habit of not taking notes while working.
This time I relied on my ears and soldering iron only.
And I know I said it multiple times trough this topic, but this time it is different, I think we nailed it

My work in stages, working from right to left:

C530 removed-no popping
C530 installed-popping

C518 removed-no popping
C518 installed-popping

C506 removed-popping, and not only popping while pressing and depressing low filter switch, but also a continuous loud buzz from speaker

And I know you said Mooly there is nothing wrong with measurement results of Q504 collector, but we did measure 13.92V where should be 15.6V.

Awaiting further instructions, and hopefully confirmation of nailing

 

[Mooly]

You've 'gotta keep notes

C530. That tells us what we already suspected that the problem is (and it has to be) to the left of C530. So with C530 refitted we continue.

C518. This is where we have to be very very sure.
C518 removed and no popping means that the problem seems to lie to the left of C518.
To be 100% sure C518 isn't at fault we have to substitute it by removing it completely and fitting either a new part or perhaps swapping with C517 which we know is on the good channel. Recheck the polarity too just to be sure. There must be no doubt whatsoever at this point.

Now it goes one of two ways... either it is still faulty or not.

Assuming it is still faulty we have to absolutely assure ourselves again that the problem is only present when C518 is fitted (which we know is good because we swapped it ).

So this is where is gets interesting...

The DC voltages on Q504 (and all the transistors) differ a little from those in the manual. That's absolutely normal. The supply voltages could be slightly different and there are component tolerances etc. As a technician I have seen so many manuals where the voltages quoted are just plain wrong so no problems there at all, however that doesn't rule out microscopic changes in DC conditions occuring rapidly due to a poor/faulty transistor.

Before we reach that point we can be sure... yes , that with C518 removed it's silent, and with C518 fitted the noise is present.
If we have reached this point it is possible it could be a transistor issue.

C506. With it removed the input the input to the preamp is "open" or "floating" and very sensitive to stray pickup. That's what the hum and noise is. That's normal.
So as we did before we must "short" the input to ground via a cap. This is where we reconnected C506 to ground (post #66)

http://www.diyaudio.com/forums/solid-state/202909-help-dc-volume-pot-4.html#post2838231

If it still pops now then we are looking at a problem around Q502 and 503... maybe the transistors, maybe not. Lets see what all this proves.