SansAmp GT2 whistling problem

[BlackTooth61]

Hi everyone!

I've built a SansAmp GT2 clone (tonepad) and it works great. But I have a problem. I get a high frequency whistling sound when I turn up the level, high or drive knob after a certain point. I get most whistling in british amp mod. I think because of treble chacteristic of british amps.

Also when I add TS808 in front of the SansAmp, even in the worst scenario, there is no whistling at all.

Thanks in advance.

[Rob Strand]

It's probably related to your wiring.

The first place to look would be the input jack and the input side of the bypass switch;
(here I mean both the input jack and the input  to the PCB).
Keep those wires away from all the other wires especially those near the output, say all
parts from S2b through to the output jack.   Sometimes using shielded wires for the
input is good enough.

With so many wires there *plenty* of opportunity to get problems.

The wires around the input up to say IC2a should be kept way from the wires
around the output say from S2b through to the output.  That includes
the level pot and tone control pot wiring.

Short wires can also help but sometimes you need a bit more length to route them away from
the others.

[BlackTooth61]

Thanks for your answer. Here is my building. There's no wiring.

[Rob Strand]

Are you running the board out of the box, as shown in the pic?

If you have a metal enclosure, the problem might go away when the pot casings and the switch
casing are connected to ground via the mounting nuts.  You should also run one wire from the circuit 0V to the enclosure ground.  You can solder a wire from 0V to the back of one of the pots or use a ring lug.

If you are going to use a plastic enclosure then you might consider running a wire strip solder to the back
of all the pot casings then connect that to the circuit 0V.    This doesn't ground the selector switches or the footswitch, so you can still have problems.

[Rob Strand]

A mod to make the layout tougher against oscillation would be to put a cap across the 100k located at pins 1 and 2 of IC1a.   Something in the order of 15pF to 47pF.

[BlackTooth61]

Are you running the board out of the box, as shown in the pic?

If you have a metal enclosure, the problem might go away when the pot casings and the switch
casing are connected to ground via the mounting nuts.  You should also run one wire from the circuit 0V to the enclosure ground.  You can solder a wire from 0V to the back of one of the pots or use a ring lug.

If you are going to use a plastic enclosure then you might consider running a wire strip solder to the back
of all the pot casings then connect that to the circuit 0V.    This doesn't ground the selector switches or the footswitch, so you can still have problems.

I fitted the PCB into a 1590BB metal enclosure. I also did the grounding as you said. Here is the picture of the box.

A mod to make the layout tougher against oscillation would be to put a cap across the 100k located at pins 1 and 2 of IC1a.   Something in the order of 15pF to 47pF.

I will try tomorrow. Thank you.

[Rob Strand]

I will try tomorrow. Thank you.

If you still have trouble  I'd be checking all the resistor values just in case.
After that a sledge hammer fix would be to use a bigger cap, like 100pF.
The values I gave should have no effect on the tone but you might need
more than that (I don't know why).

EDIT:
Beyond that you might need to add a cap across pins 6 and 7 of IC4b.
Off hand maybe 100pF but I'd have to do some checks to come-up with a precise value.

Both cap mods will help various issues with oscillations.

[BlackTooth61]

I will try tomorrow. Thank you.

If you still have trouble  I'd be checking all the resistor values just in case.
After that a sledge hammer fix would be to use a bigger cap, like 100pF.
The values I gave should have no effect on the tone but you might need
more than that (I don't know why).

EDIT:
Beyond that you might need to add a cap across pins 6 and 7 of IC4b.
Off hand maybe 100pF but I'd have to do some checks to come-up with a precise value.

Both cap mods will help various issues with oscillations.

I tried to put 100pF cap across pins 1 and 2 of IC1 and pins 6 and 7 of IC4. Frequency of whistling is decreased. Less disturbing. But still there is.

[Rob Strand]

I tried to put 100pF cap across pins 1 and 2 of IC1 and pins 6 and 7 of IC4. Frequency of whistling is decreased. Less disturbing. But still there is.

So that says it is oscillating.   The other piece of information you gave is the whistle goes away when you precede the device with a TS808 pedal.   That suggests perhaps the input wires are contributing to the problem.   It could be the tracks around the jacks and the footswitch have input and output wires which are too close together or don't have sufficient ground tracks around or between them.

Another mod that might help that scenario is to add say 47pF to 100pF across the 1M resistor at the input to the circuit.    Ideally you want to use the smallest value possible.  If the fix works decrease the value until the problem disappears then use say 2x or at least 1.5x that value.   I will warn you that the unit can "whistle" at frequencies above your hearing.  You might think it's fixed but it is not.  This is where you need an oscilloscope.    Another position for the cap is pin 5 IC1b to ground - do one or the other but not both.

[BlackTooth61]

I tried to put 100pF cap across pins 1 and 2 of IC1 and pins 6 and 7 of IC4. Frequency of whistling is decreased. Less disturbing. But still there is.

So that says it is oscillating.   The other piece of information you gave is the whistle goes away when you precede the device with a TS808 pedal.   That suggests perhaps the input wires are contributing to the problem.   It could be the tracks around the jacks and the footswitch have input and output wires which are too close together or don't have sufficient ground tracks around or between them.

Another mod that might help that scenario is to add say 47pF to 100pF across the 1M resistor at the input to the circuit.    Ideally you want to use the smallest value possible.  If the fix works decrease the value until the problem disappears then use say 2x or at least 1.5x that value.   I will warn you that the unit can "whistle" at frequencies above your hearing.  You might think it's fixed but it is not.  This is where you need an oscilloscope.    Another position for the cap is pin 5 IC1b to ground - do one or the other but not both.

Again, it's oscillating at lower frequency. I will try with the smallest value I can get.

[Rob Strand]

Again, it's oscillating at lower frequency. I will try with the smallest value I can get.

I have a feeling something is actually wrong and we are solving the wrong problem.

The mods I suggested are really to get around minor issues, things where the build could
push it into the bad region and the mods push it back into the good region.   We have
pretty much thrown everything at it and clearly the problem is still there.

It would be worthwhile checking the follow:
- All resistor values, cap values etc.
- check all solder joints
- Make sure the switch contacts are in fact switching like the schematic implies.
  If the switch contacts were rotated by one position it would cause all sorts of weird behaviour.
  Another possible issue is the switch contacts are not closing due to a faulty contact.

It really needs a sanity check across the whole build.   It's only takes one crack and the whole thing collapses. Because it's a large circuit the best way to do it is to check each section is working.

If I had it here I'd be injecting a sine-wave at a few frequencies and verifying the measured frequency
response and gain at various points in the circuit agreed with that expected from the schematic.  I'd also check that the signal on one side of each switch contact agreed with the signal on the other side of the contact.

If I had to guess I'd say the problem is around the first stage.

[BlackTooth61]

Again, it's oscillating at lower frequency. I will try with the smallest value I can get.

I have a feeling something is actually wrong and we are solving the wrong problem.

The mods I suggested are really to get around minor issues, things where the build could
push it into the bad region and the mods push it back into the good region.   We have
pretty much thrown everything at it and clearly the problem is still there.

It would be worthwhile checking the follow:
- All resistor values, cap values etc.
- check all solder joints
- Make sure the switch contacts are in fact switching like the schematic implies.
  If the switch contacts were rotated by one position it would cause all sorts of weird behaviour.
  Another possible issue is the switch contacts are not closing due to a faulty contact.

It really needs a sanity check across the whole build.   It's only takes one crack and the whole thing collapses. Because it's a large circuit the best way to do it is to check each section is working.

If I had it here I'd be injecting a sine-wave at a few frequencies and verifying the measured frequency
response and gain at various points in the circuit agreed with that expected from the schematic.  I'd also check that the signal on one side of each switch contact agreed with the signal on the other side of the contact.

If I had to guess I'd say the problem is around the first stage.

You're right! I've checked the resistor values on my layout drawing. Somehow I had drawn the R3, R4, R5 and R6 as 2.2k instead of 1k. My silly mistake. I'm glad we find the issue. Thanks a lot again! I will report the result after changing the resistors.

[BlackTooth61]

Again, it's oscillating at lower frequency. I will try with the smallest value I can get.

I have a feeling something is actually wrong and we are solving the wrong problem.

The mods I suggested are really to get around minor issues, things where the build could
push it into the bad region and the mods push it back into the good region.   We have
pretty much thrown everything at it and clearly the problem is still there.

It would be worthwhile checking the follow:
- All resistor values, cap values etc.
- check all solder joints
- Make sure the switch contacts are in fact switching like the schematic implies.
  If the switch contacts were rotated by one position it would cause all sorts of weird behaviour.
  Another possible issue is the switch contacts are not closing due to a faulty contact.

It really needs a sanity check across the whole build.   It's only takes one crack and the whole thing collapses. Because it's a large circuit the best way to do it is to check each section is working.

If I had it here I'd be injecting a sine-wave at a few frequencies and verifying the measured frequency
response and gain at various points in the circuit agreed with that expected from the schematic.  I'd also check that the signal on one side of each switch contact agreed with the signal on the other side of the contact.

If I had to guess I'd say the problem is around the first stage.

You're right! I've checked the resistor values on my layout drawing. Somehow I had drawn the R3, R4, R5 and R6 as 2.2k instead of 1k. My silly mistake. I'm glad we find the issue. Thanks a lot again! I will report the result after changing the resistors.

I was going to change the resistors but when I opened the box, I've seen that all resistor values are correct. I can't remember what happened at that time. Probably I looked at unsaved layout file. Well, we are back to the beginning.

- Cap and resistor values are correct.
- I've checked solder joints. No problem.
- Switch contacts works well.

I wish that resistor values were incorrect.

[Rob Strand]

I wish that resistor values were incorrect.

Hmmm, that's a pain.

I'm still fairly sure something is *really* wrong.   It's a matter of what that is and how to find it.

Another "quick fix" attempt would be to wire 100nF caps across the opamp power rails.

A couple of other things that's easy to over look are:
- Electrolytic caps around the wrong way.
- unconnected pins on the IC's
- missing ground connections
- broken nets  (As in, you gave ground in one part of the circuit and ground in another, locally both parts
of the circuit look good.  However, there a missing track connecting those two parts together.)
- The jacks are wired wrong, in that one of your tracks goes to the connector pin that disconnects when you insert the jack.

Beyond that it's going to be a hard slog going through every part and every track on the layout verifying that all is what you think it should be.   I'd probably use a CRO to check the actual ckt.

Approaching the problem from the side of the built unit, you have to test each part of the circuit. That means lifting the 22nF cap on the input and 2u2 cap on the output then wiring each section to those points to see if they are working.   Another idea is to bypass sections of the circuit using a wire.    The aim is to see what has an effect and what does not.    You try to find a pattern or a problem area of the circuit.   Keep in mind bypassing a high-gain circuit inherently reduces the chances of oscillation problems.

These types of problems are the hardest to find and take a lot of leg work and persistence.  It takes long time with no guarantee of success - that's why you want to do as many easy checks as possible.   Checking DC voltages is a always a good thing to do.

I suspect the problem is around the first stage so maybe put more time into that area.

Unfortunately I'm getting to the end of list of easy-ish things to try.

[MrStab]

hi Blacktooth,

does the TS808 prevent whistling when the TS808 is in bypass mode? and did you definitely buy the correct type of slide switches?
also, what kind of guitar/pickups are you using?

it's definitely oscillation, just some ideas to help you guys narrow it down.

[Rob Strand]

it's definitely oscillation, just some ideas to help you guys narrow it down.

Your thread last year was a tough one.   At least that was a new design, you expect some trouble, but this one is a tried and true circuit.   What's new?  maybe the layout.  So maybe something subtle there causing issues.

[MrStab]

thanks again for all that help! that was last year? feels like last lifetime!!

could be worth contacting Tonepad about any revisions or similar reports, if all else fails.

[Rob Strand]

thanks again for all that help! that was last year? feels like last lifetime!!

No worries.   Pretty sure it was around Christmas Dec 2016/Jan 2017.  Seems like a long time ago for me too.

[rankot]

Thanks for your answer. Here is my building. There's no wiring.

Is this your PCB design, or you bought proven one from somewhere?

[BlackTooth61]

hi Blacktooth,

does the TS808 prevent whistling when the TS808 is in bypass mode? and did you definitely buy the correct type of slide switches?
also, what kind of guitar/pickups are you using?

it's definitely oscillation, just some ideas to help you guys narrow it down.

Whistling is going on when the TS808 is bypassed. I'm sure it's the correct type of switch. I tested the pins one by one. I'm using these switches (only 2 pole of course): https://www.taydaelectronics.com/rotary-switch-4-pole-3-position-alpha-sr2511.html 
Single or humbucker pickups, I tried both. Sometimes there is a whistling on 1,3 and 5 pickups positions (for SSS strat) and no whistling on 2 and 4 positions. But in the worst scenario, there is always whistling.

A couple of other things that's easy to over look are:
- Electrolytic caps around the wrong way.
- unconnected pins on the IC's
- missing ground connections
- broken nets  (As in, you gave ground in one part of the circuit and ground in another, locally both parts
of the circuit look good.  However, there a missing track connecting those two parts together.)
- The jacks are wired wrong, in that one of your tracks goes to the connector pin that disconnects when you insert the jack.

All grounds are common. I used the same jack connection structure several times without any problem. Here is the schematic drawing of my build. Layout of this drawing: https://gerber-viewer.easyeda.com/showcase/?#!id=4e556e22371311e899d9026a86b9cae7&type=top&layer_list=1-2-3-4-5



By the way, I must say that the circuit works well. I like the sound of it. Every mod, amp and mic settings work as expected. Only problem is in some settings, there's oscillation.

I have an idea, even though it may not be sensible. As I put TS808 in front of the SansAmp, whistling is gone. In order to change the input impedance, can it be helpful to change the input resistor of the SansAmp circuit?

Is this your PCB design, or you bought proven one from somewhere?

My design.

EDIT: Also, as I'm lowering the volume knob of the guitar, frequency of the whistling is going down gradually and whistling is gone around 7/10.