Loud pop on switches

[bluemonk]

Hello. I'm buiding an amp based on BF design. I put a couple of switches in the signal path:
the first select between different treble caps;
the second turns on-off a whole gain stage.
These switches pop loud (the latter very loud).
Is there some trick to tame this noise?

Thanks in advance

[Enzo]

I'll bet my lunch money you are leaving caps unterminated.

When your switch flips between caps, one cap is hanging free at one end. When the switch completes the circuit to it, that cap must suddenly charge up to whatever voltage. That pulse of charging current makes a pop.

You need to keep the cap charged. A simple way is with a high value resistor. Put a 1 meg or 470k resistor in series with the cap. Now use your switch to short across the resistor. The resistor keeps the cap chanrge, adn when the switch is open the large value of the resistor keeps the cap from having any effect. By shorting across it, the resistor gets out of the way and the cap now functions, but without having to charge up first.

[bluemonk]

Quite right! I left both ends of caps unterminated.
Now that you explained me the solution I am confused about the easiest way to do the work using a 6-PDT rotary switch and (possibly) flying wiring.
Any suggestion?

Thanks a lot

[Enzo]

Doesn't matter how many other poles there are. Look at each switching element. You have a cap going to a terminal of the switch. You have the other terminal of the switch that will complete the circuit. Add a 1meg or other high value resistor across those terminals. If it is a rotary switch with say three different caps, then add three resistors. One resistor to each cap, and then the three resistor remaining ends will all be wired together to the common pole of the switch. Now each cap has a resistor completing the circuit as far as DC is concerned, but the size of the resistor makes it mostly non-functional as a signal modifier. Switching the switch to a particular cap then winds up shorting across its resistor. The cap remains charged, but now the resistor is gone and the cap takes effect.

[bluemonk]

I did so and now the switch on treble caps works nice. Thanks!

Did the same (crossing 10M resistor) at another switch on signal path and it's still popping (altough less than before).
This switch enable/disable a whole gain stage and is placed just after the caps.
Is there a way to make it quiet?

[Enzo]

10 meg? Try a 1 meg or 470k.

Enable/disable how? Open series circuit? Short signal to ground?

[bluemonk]

I try to explain how I wired that switch (please be patient with my English!).
I wanted to build an hybrid aa964-aa1164 Princeton. Basic design is aa964. Since aa164 has one more gain stage I put a simple two-way switch (please look at the pict. It just show where the two-way switch is located).

1st option: the volume output goes to V2 (stock aa964). The bypassed stage goes to nothing.
2nd option: the volume output goes to the next gain stage (stock aa1164).
Is that uncorrect? Is there a better way?

I put a 10M resistor across the switched poles. It just tame something but still pops.

Thanks in advance

[Steve Conner]

You probably need to think about the DC conditions of the right-hand switch section some more. If you inserted the switch in a place where DC current normally passes, then the 10M resistor trick won't work. The switch should break AC only, and the resistor then mops up any capacitor leakage.

We can't tell any more from the section of schematic you posted.

[bluemonk]

Originally Posted by Steve Conner

You probably need to think about the DC conditions of the right-hand switch section some more.

Unfortunately my knowledge is poor. I can't say what else to do instead of leaving the right side floating... grounding?

Originally Posted by Steve Conner

We can't tell any more from the section of schematic you posted.

Well, the schematic I posted is from stock Princeton AA1164:
Prowess Amplifiers - Fender - Schematics - Princeton Reverb aa1164 - Schematic

I attached just a part to make the post not too heavy.

Thanks everybody for your attention.
I hope further suggestions... amp is almost done

[bluemonk]

Attached the schematic of my homebrew.
The switch placed after tone control pops. I put a 10M-R across the switch but with very little effect.
Is there a way to cancel the loud pop on switching?

Thanks in advance

[defaced]

Check out this thread. I've linked page two because that's where we solved the problem, but you might want to read the whole thing. Look at the schem I posted (post #44) and you'll see how to terminate the relay contacts and a way to switch this tube stage in and out without pops. This is the method used in an amp I am building and it works without pops.

12v transformer for relays

[Steve Conner]

The grids of both your tube stages are floating, and a single resistor can't cure that. You said "the switch", but there are two switches, so you need two resistors.

I recommend a 10M from each grid to ground.

[JoeM]

Instead of switching in the additional gain stage, why not just switch another resistor in parallel with the 3.3Meg/10pf? You dont have to be concerned with matching the dry and wet reverb signals.

[loudthud]

You can't leave the grid of a tube floating at any time, even between switch positions. A 10Meg is too high in that position, a negative voltage will deveolpe and you will still hear a pop. Install a 1 meg to ground at each grid. A 10meg will work when switching coupling or bright caps. If you are switching electrolytic caps in cathode circuits, 100K might be a good choice.

[bluemonk]

Originally Posted by Steve Conner

I recommend a 10M from each grid to ground.

Still pops but not really loud.
It seems to me that it's almost the same result than wiring a 10M across the grids (I did that first).
Thanks.

[bluemonk]

Originally Posted by JoeM

Instead of switching in the additional gain stage, why not just switch another resistor in parallel with the 3.3Meg/10pf?

Good idea... but since amp is finished I'm trying to debug my building before considering to change design. Thanks.

[bluemonk]

Originally Posted by loudthud

Install a 1 meg to ground at each grid.

WoW! That works!
Actually I used 820k assuming they were good substitutes for 1M... aren't they?

Thank you everybody

[loudthud]

I love it when a plan comes together 820K will be fine.