Problem with switch noise on simple input stage

[MrShhh]

Hi there,

I've set up a simple balanced input stage with 12 position switchable gain trim (range -6dB to -11.5dB)

The gain is controlled by swiching R3 & R4 with a dual gang rotary MBB switch. R3 & R4 are carefully matched to preserve the CMRR.

I'm having trouble with the switch crackling and popping randomly, sometimes it's silent and sometimes it produces loud clicks. This only happens when the stage is passing audio.

Is this a problem with DC building up on C1, C2 or C3? Or something else?

I don't fully understand what might be happening here, it's the first time I've encountered this problem. I'd be grateful if someone could point me in the right direction. Thank you! 

 

[RuudNL]

Did you use a 'make-before-break' switch?

 

[MrShhh]

Cheers Ruud, I did use a make-before-break switch.

Btw: Removing C3 doesn't help, neither does tying C1 & C2 to ground through 1M resistors.

The clicking is worse on the gang that switches R3.

Also, I rebuilt the switch 3 times just to be sure it wasn't a soldering issue.

 

[RuudNL]

Did you try a different OpAmp?
If the OpAmp has an offset, the offset voltage is amplified with the selected gain.
Changing the gain could then mean DC-jumps, causing audible 'pops'.
I would try a 'good-old' TL0xx OpAmp, IMHO there is not much performance gain in choosing 'exotic' OpAmps in line level stages.
(OpAmps with bipolar transistor inputs usually have a bigger offset than FET OpAmps.)

An other possibility could be that the circuit is (on the edge of) oscillating. Did you check this with an oscilloscope?

 

[MrShhh]

I started out with NE5534 but the problem is the same. TL071 makes no difference either.

It doesn't appear to be oscillating but I might be missing something on the scope (I have the probe set to x1 and all looks fine). The 33pF cap rolls off the response at 482kHz at gain of -6dB. This increases to 900kHz at gain of -11.5dB. Is that less than adequate?

 

[Harpo]

The gain is controlled by swiching R3 & R4 with a dual gang rotary MBB switch.

Wild guessing, due to mechanical/tolerance reasons between both switch gangs they won't switch simultaniously.

 

[MrShhh]

Wild guessing, due to mechanical/tolerance reasons between both switch gangs they won't switch simultaniously.

e.g. this would mean that R3 Might be 10K while for the briefest moment R4 might be 9K53 (or vice versa) but the clicks are a loud discharge, sounds a bit like relay popping when there's no flyback diode. I don't know how the 0.5K or so momentary discrepancy could cause such a loud pop. 

I will try a different kind of switch to rule this out. I will first try increasing the compensation cap to see what happens.

 

[MrShhh]

I changed the compensation cap to 100pF but still no joy....

 

[JohnRoberts]

Can you leave the fixed largest value gain resistors always at R3 and R4?  then switch lower values in parallel to effect lower gains.

There should be a similar value cap across R4 as R3 for HF balance.

Noise can be caused by intermittent open circuits during switching. 

JR

 

[MrShhh]

Thanks John,

I tried exactly as you say thinking it might be momentary unterminated connections messing with the feedback, but it didn't help.

This shouldn't be an issue with a MBB switch but I wanted to be sure it wasn't to blame.

 

[JohnRoberts]

If that doesn't fix it then issue is probably DC offset in the opamp. The changing gain causes a step impulse when the DC error is multiplied by the different gain.

Fix for that is to cap couple so DC gain remains constant and only AC gain changes.  Again DC couple the max resistor values and cap couple the switched in shunt values.

Then any clicks remaining will be from signal present during switching.

JR

 

[MrShhh]

Aha, now why didn't I think of that....!!! Thanks very much  ;D I'll give that a go tomorrow and get back with the results.