do I need bias voltage on the signal with electronic switch?? (4016)

[maxime]

Hello,
I am currently elaborating a nice sequenced object.
I am planning the use of the electronic switch 4016 chip. and I can't find a clear reply on the datasheets of the chip. I want to use it to treat an audio signal. simple ... here's my question:
Let's say supply voltage is 5 volts. does the switch want to see half that value at the input or will it be acting like a real physical switch, not caring wether the values of the signal are above or below ground potential?
Is a 2.5 volts bias needed at each input?

thanks

Maxime

[Processaurus]

I'm curious about this as well, actually, two excellent pedal specific articles on cmos switching I've read have differing views.  It certainly would be easier not to have to bias everybody...  One thing for sure is the signals being switched must have exactly the same DC bias, or it will pop just like a mechanical switch.

http://geofex.com/Article_Folders/cd4053/cd4053.htm

http://www.geocities.com/thetonegod/switches/switches.html

[SISKO]

I dont know if would pass signal below ground, i think it would. But you definetely  need a bias voltage at the input and output so you wont have a pop

[maxime]

nice to have you around!

it seems like a bias is needed..

thanks

maxime

[R.G.]

Let's say supply voltage is 5 volts. does the switch want to see half that value at the input or will it be acting like a real physical switch, not caring wether the values of the signal are above or below ground potential?
Is a 2.5 volts bias needed at each input?

I'm curious about this as well, actually, two excellent pedal specific articles on cmos switching I've read have differing views.  It certainly would be easier not to have to bias everybody...  One thing for sure is the signals being switched must have exactly the same DC bias, or it will pop just like a mechanical switch.

All CMOS switches (4016, 4066, 405x, and others) use pairs of N-channel and P-channel devices to do the switching. The pairs are hooked up drain to drain, source to source. The substrate diode which prevents using discrete MOSFET devices as analog switches in most cases is arranges so that if the signals are within the power supply limits, the diodes are not activated, so signals within the power supply are switched normally.

The N-channel switch works GREAT for signals from about ground to 2/3 of the power supply voltage, where it starts having successively higher resistance. Th P-channel device works GREAT from about the power supply down to 1/3 of the power supply. With the two in parallel, you have good switching and low switch resistance from rail to rail.

However, all switch elements have some capacitance from the control signal to the signal path. This can cause pops directly. In JFETs we slow this signal down so the pops are sub-audio and not heard. You can't do that with CMOS, as the control signals at the switch devices are all internal. What you can do is take note of the fact that the N-channel device injects a little charge into the audio path (and "charge injection" is what you see on the datasheet) when its control line moves. The P-channel device SUCKS a bit of charge out of the audio path at the same time.

The amount of charge injected and removed depends on where the signal voltage is at the time the control signal moves. If it happens to be nearly at the middle of the power supply so that both N and P devices get about the same amount of voltage difference to the path, the charges tend to cancel, and the overall pop from control signal charge injection is so close to zero, you can't hear it.

This is why I put biasing to half supply in the article on CMOS switching at GEOFEX.

Some newer CMOS switches that are not of the CMOS 4xxx family have de-popping stuff inside. That's really nice if you are using them. But they still will not switch signals outside their power supply range well, if at all. The CD405x family has a "Vee" terminal. You can make Vdd be as much as +7.5V and Vee as much as -7.5V and switch signals with a DC voltage of 0Vdc while using a 0Vdc to +7.5Vdc control signal. That sounds like switching signals without biasing them, but actually the thing works best with signals having a dc voltage of 0V - again, right in the middle of the power supply.

What matters in technology is the facts, and the physical reality underneath the facts. Things like gravity and thermodynamics don't care what your opinion of them is, and don't change their operation depending on how many people believe in them; in that way, they are kind of the exact opposite of William Jefferson Clinton, whose idea of leadership was to take a poll and then go do what the poll said. Go find the facts. Go dig out how the things work, and don't rely on internet views.