Analog switch pop, click problems

[retrotubeguy]

  Hi. I am having problems with the Analog Devices ADG5434BRUZ Analog Switch. No matter what i do I can not eliminate popping in the audio. I have have tried all kinds of tricks but nothing works. All types of coupling capacitors, various values, rampping up the logic input voltage, all kinds of grounding schemes and nothing seems to work. Are these switches notorious for this kind of thing. Data sheet does say it can be used for audio. I am missing something fundamental but can't seem to put my finger on it. Getting so frustrated that I am considering using relays again. i prefer the analog switch due to its size, low mass and immunity to vibration. Any Ideas...

[Alex Eisenhut]

They make clickless switches for audio.

http://www.digikey.ca/product-detail/en/maxim-integrated/MAX4910ETE-/MAX4910ETE--ND/1779955

http://pdfserv.maximintegrated.com/en/an/AN4449.pdf

[danadak]

The AD switch is a break before make switch, so for a short period of
time the output becomes floating and if you are feeding a low Z load
that will discharge the input to the next signal block in your signal
chain. If the load is high Z then when switch open then it becomes
subject to stray signal pickup.

Its also speced for charge injection. You might examine the the load
effects of that.


Regards, Dana.

[retrotubeguy]

I understand that it is break before make. I tried to connect multiple switches in a make before brake  fashion controlled by a micro controller but same problem exists and there is also the problem that for some reason I get digital noise passing thru the dam thing. This could just be the bread board or jumper wires I guess. Made it up on strip board and still had this problem though. Maybe if I find an 12C or Spi IC and use a digital isolator. I should possibly try one of the maxim devices with this setup. I am reluctant to try the maxim click less switches as they are a claiming a reduced switch noise not total silence, I would like to avoid buying lots of experimental chips that will just sit in my junk bin. The analog devices stuff is not cheap but if it worked the way I had planed then I guess it is well priced. If anyone has any suggestions or experience with these devices used in audio I would be keen to see how it could or should be done. Signal relays or opto couplers are  a last resort. I have considered the vactrol opto isolators which are basicly an LED and an LDR, not sure how analog switches compare as far as distortion goes. I have not detected any audible or significant distortion with the ADG5434 providing it is set up correctly with adequate power rails etc.
Thanks for your valuable time. I do appreciate that it takes time to formulate answers to some of these questions.

[tggzzz]

What reason do you have for believing that optoisolators might help?

What is the voltage on the two inputs at the instant of switching? If one happens to be +2.453V and the other -3.6328V then there is a 6V transition which will be audible as a click.

What is the DC offset on any of the voltages? If any, then that may be audible as a click, depending on where the voltage is.

What is the charge "left in" a coupling capacitor when it was disconnected? That may cause a current to flow when next switched into the circuit.

[retrotubeguy]

I have virtually no DC offset. I have coupling caps with resistors to discharge them I cant seem to measure anything significant not sure of the voltage as I have pulled the circuit down and need to rebuild it to measure this. I understand what you are saying but even if connected directly to a guitar pickup then to an amplifier it still pops. Is it the IC itself possibly generating some DC. I have never used these switches before so I may not understand the device entirely but it seems simple enough. The opto isolator idea was considered if using SPI or I2c Just as a precaution, offcourse I have not tested the IC with or with out them. It is just an assumption at this point that there may be some benefit.

[tggzzz]

Have you ensured a good separation between the analogue and digital sections of the circuit? When voltages and currents change, they induce voltages and currents in the conductor and neighbouring conductors.

[Buriedcode]

Have you tested this on its own?  It could be charge injection when you're turning on the switch. 

A method I used a while back for a stompbox controller with analogue switches was to use an SPST analogue switch to switch the output (or 'input' to the system) to ground just before changing the analogue switch used for selection, then releasing the other. I didn't do this for pop/click suppression, it was for crosstalk (using cheap HC series analogue multiplexers) but I had no pops.

This app note might help:
https://www.fairchildsemi.com/application-notes/AN/AN-6044.pdf

[dmills]

Charge injection would be my guess, but too sudden switching will always have the potential to click simply due to the potential step change in output. 
Doug Self shows some ideas in his book on small signal audio using a couple of jfets per input and ramping the control voltages to provide a smooth transition.

Regards, Dan.

[nfmax]

First check - do you still get pops & clicks when switching between silent audio sources (driven, not floating)? If so, it's almost certainly charge injection. This is probably from the switching device itself, but it just might be capacitive coupling from the logic control line. One trick I have used in the past is to implement the switching at a virtual earth point, so that the voltages on the switch terminals do not change. You need two switches, one normally open and one normally closed. Arrange these so that the input signal is routed, via a resistor, either to the virtual earth input of an inverting OPAMP stage, or to real earth (the switches go at the 'earthy' end of the input resistor).

[retrotubeguy]

Some Stuff to mull over here. I have tested the Analog switch on its  own simply connecting a DC voltage to the logic pin with no audio present but connected , not floating. Still pops. Terminated all unused switches and logic inputs still pops. I think your right about the charge injection. Is there any way to overcome this. the data sheet is not very helpful. I don't think this is the best choice for this application. I think I may need to find another way. Even relays will pop unless switched at zero crossing, not as bad though at lower frequencies they still cause a fairly loud pop. I will need to investigate the suggestions that you have provided.

[dmills]

Self has a whole chapter on ways to do this, well worth the time.

Regards, Dan.

[retrotubeguy]

What / Who is self. Do you have a link or source. Or do you men yourself ?

[jahonen]

I think he means Douglas Self, an audio guy:

http://www.douglas-self.com/

Regards,
Janne

[dmills]

Yep, specifically this http://www.douglas-self.com/ampins/books/ssad2.htm book, chapter 21.

I had assumed that Doug Self was well enough known in the audio world to not need an intro.

Regards, Dan.

[nctnico]

Rule number one when switching analog audio: make sure the DC level on both sides of the switch is equal.

[retrotubeguy]

Thanks, I know the rule but it does not appear to be what is happening. I followed the link to the self site and I am going to be missing in action while I drool over the the thing for a while. All I can say is Ahwauuuuuwwwwwwwwww,drip,drip drip, pick up jaw.

http://www.thefreedictionary.com/drool

[Raj]

Hi. I am having problems with the Analog Devices ADG5434BRUZ Analog Switch. No matter what i do I can not eliminate popping in the audio. I have have tried all kinds of tricks but nothing works. All types of coupling capacitors, various values, rampping up the logic input voltage, all kinds of grounding schemes and nothing seems to work. Are these switches notorious for this kind of thing. Data sheet does say it can be used for audio. I am missing something fundamental but can't seem to put my finger on it. Getting so frustrated that I am considering using relays again. i prefer the analog switch due to its size, low mass and immunity to vibration. Any Ideas...

Try simple cmos swich triggered via micro-controller to delay and filter out the physical switch's pulse

[retrotubeguy]

Solved with LDR setup, less distortion , total isolation from the micro, absolutly silent transition between sources. minor insertion loss but was not buffered in test. Can still use the ADG 5434 or similar to operate the LDR circuit. a bit more circuitry but not a big deal. I havent given up on cmos yet but that charge injection seems difficult to overcome. More study time would be handy. Thanks to everyone here for the enormous amount of help.