Best way to bypass a circuit without pop

[daniel1010]

I have a simple circuit here and would like to make a true bypass without pops.
I use a non shorting button (SPUJ193700) but when I switch I get a short pop.
I assume this is because the unbalance audio connection is short disconnected (non shorting / break before make).
Any recommendation? Thank you very much for the newbie question.

 

[Bo Deadly]

Your op amp is missing a bias resistor. You cannot just have an op amp input floating open like that. There is a small bias current which, in this particular case, will cause the capacitor to charge up (or down). When the switch closes, that current will discharge rapidly through it which is equal to a "pop".

You also need to make sure that the voltage on either side of the switch doesn't drift for other reasons. Your circuit is too simple to understand where and how that might occur. You are leaving out a lot of parts. Provide a more complete schematic.

 

[ruffrecords]

The problem is most likely due to retained charge on the 10nF capacitor (or anny other coupling caps at the input or output). You need to make sure that when the switch is in the bypass position the cap can discharge to 0V so that when you switch back there is no dc difference between the two. Just make sure there is a cap discharge path at both the input and output of any active electroncis.

Cheers

ian

 

[daniel1010]

Ian. Thank you for the support

I can follow your statement, but when I remove the capacitor from the circuit I have the identical effect.
In the middle of the button, when it is not yet directly switched through (DPTP / non shorting / break before make button) I hear a hum. It sounds like the signal is hanging in the air.

I think that would explain the pops. But how can I get around that?
Daniel

 

[daniel1010]

Your op amp is missing a bias resistor. You cannot just have an op amp input floating open like that. There is a small bias current which, in this particular case, will cause the capacitor to charge up (or down). When the switch closes, that current will discharge rapidly through it which is equal to a "pop".

You also need to make sure that the voltage on either side of the switch doesn't drift for other reasons. Your circuit is too simple to understand where and how that might occur. You are leaving out a lot of parts. Provide a more complete schematic.

Bo you will laugh but that's exactly how I just built it on a prototype board .
Goal, an active lowpass with LME49720 (Audio Level -10 dBV) and a simple bypass without pops to achieve.

 

[daniel1010]

The problem has been fixed. Firstly, I made a coupling for input and output. On the other hand, the LME OpAmp were not cleanly suppressed via the necessary capacitors. The rest now fits and runs great.

 

[Matt Syson]

a truly 'pop free' changeover is actually remarkably difficult to achieve. the fact the wiper of your switch goes 'open during a part of it's travel is significant, as well as the slight DC offsets you might have. many years ago the 'test' for pop/click free switching was that it must not cause a PPM to register a 'peak' disturbance of more than 4dB above the residual noise level. So yopu would amplify the noise 9hopefully just hiss (white noise) to get a 0dBu reading on the PPM then operate the switch..
One possible 'dodge is to have a largeish resistor that links the wiper to one of the two 'input' sources (one with the highest source impedance probably, such that when the wiper was part way you would still receive some signal until the wiper contacts the lower impedance source signal which would swamp the previous source via a high resistance.
A cheat but quite effective but you need the right conditions available to do it.

 

[daniel1010]

Thanks Matt for your great comments. I will look into this in detail.

As mentioned the pop has now become much quieter, but when switching still a spike to hear.
You have also explained well why.

What surprises me only, is my use case so strange. How do others solve this?

 

[ruffrecords]

What surprises me only, is my use case so strange. How do others solve this?

Back in the day at Neve we use 4M7 resistors between switch contacts.

Cheers

ian

 

[daniel1010]

Back in the day at Neve we use 4M7 resistors between switch contacts.

Cheers

ian

Ian, thanks also for this important comment.
I have not so large resistors there and will probably have to bundle them to a very large cluster

I keep you up to date

 

[Matt Syson]

Don't worry about a specific value. In fact I think Ian is remembering some spoecific instances where indeed a 4M7 resistor is used to ensure the non polarised capacitors on switched frequency bands of the EQ section use the resistors to bleed off/ensure the DC potential is the same or actually something near the instantaneous signal voltage at the moment it is switched. If there is an electrolytic cap involved then values down to ever 100K might be appropriate. Again EXACT circuit details matter.

 

[daniel1010]

Matt and Ian, I thank you very much for going in detail. I'm excited.

However, I ask for your understanding that I can only provide the schematic in simulated overview form without the wiper change.

The solution the wiper to one of the two 'input' sources is an excellent idea which I am currently testing.

 

[Matt Syson]

The 100uF 'output capacitor should be in the output of the op amp and then have the 10K 'bleed' resistor after it. You must also ensure there is NO DC on the 'bypass route either so whatever is feeding the first wiper of the switch MUST be DC free. Of course you could actually feed the filter all the time which would eliminate the first switch contacts altogether. by switching anything when not necessary is designing problems into the circuit. dual and multipole switches do not make and break at the same 'time' between different poles. This 'subtlety can be the make or break of a State variable filter circuit where it is possible for it to become an oscillator momentarily during switching.

 

[Matt Syson]

The bias current for op amp inputs flows out of the input terminals. FET devices have minimal bias current of course. There are some around that extol the 'virtues' of having capacitor free linking between amplifier stages BUT ensuring there is NO DC offset in any permutation and over a wide temperature range can produce gear that plops and bangs during switching (or scratching when pots are rotated). Adding a 'DC servo to hopefully minimise DC offsets may not actually help much because the extra op amp for the servo function can also drift with temperature and have it's own inaccuracies.

 

[daniel1010]

I'm still trying to work out a solution which is not so trivial
I think so far the most sustainable idea is just to switch the filter capacitor 10nF away from ground. Certainly a not so good idea, probably.
I suspect the problem is the AC coupling because the faster I switch the quieter the pop.

 

[Matt Syson]

If you switch the 10nF capacitor away from ground it will still click when you operate the switch on some occasions. Putting a large value resistor across the switch contacts would help but if the resistor is too low a value it will modify the overall frequency response a bit.

 

[ruffrecords]

I'm still trying to work out a solution which is not so trivial
I think so far the most sustainable idea is just to switch the filter capacitor 10nF away from ground. Certainly a not so good idea, probably.
I suspect the problem is the AC coupling because the faster I switch the quieter the pop.

Your solution needs to ensure that in all switch positions, all capacitors have a dc discharge path from one side of the cap to the other.

Cheers

ian

 

[daniel1010]

You are both great and I am still learning very well. I appreciate that too much.

A defined capacitor DC discharge prevents the "pop". From 100k ohms, the switching is transparent to the filter and would commit me to 1M ohms.


A "silent" bypass to achieve will probably still take some time but you have delivered good approaches for this. Thank you!

 

[JohnRoberts]

"Pop" is not a technical term but suggests full range including LF content (thump includes even more LF). This usually is the result of switching between two sources with different DC levels.

Even with levels matched you can get "clicks" (more HF content) when switching musical signal in the middle of a waveform.

Ironic perhaps audio purists hate extra capacitors added to their audio paths, but customer really hate ticks, and pops, and thumps, so design engineers do what they can to keep the paying customers happy.

JR

 

[Whoops]

Some good and enjoyable short reads on the subject:

http://www.muzique.com/lab/pop.htm
http://www.muzique.com/news/pulldown-resistors/
http://www.geofex.com/Article_Folders/box_pop.htm