Chaining CMOS switches

tysonlt · January 07, 2012, 09:36:19 PM

Hi all,

I'm designing an effect switching unit using cmos 4053, ala the geofx article. I would use two of the channels for guitar signal and ground, and the third channel to drive a red or green led. Overall I would have about ten loops.

I would like to ask a few questions:

1) in this setup, even with all loops off, the signal goes through ten chips. Is this going to be detrimental to the signal? Specifically thinking of the cumulative treble loss that occurs when chaining boss pedals together.

2) R.G's design has a resistor and cap on each CMOS channel. If I'm chaining them together, do I only need these on the first chip?

3) I plan to use a separate 9v regulator for the cmos chips because I heard this increases the headroom. I'm not quite sure if running the cmos at higher voltage will still let me drive the control pin from the pic running at 5v. I read the datasheet and I think that I will need at least 6v to flick the switch. Is this a job for a transistor? Would the emitter/collector simply be wired between the 9v rail and cmos control pin?

Thank you everyone. I already feel so much more confident with electronics thanks to this forum.

alparent · January 08, 2012, 08:19:38 PM

I'm not that smart so I can only help on one question

Yes .... run the CMOS with 9v and use a transistor triggered by your 5v trigger the CMOS gate.

I've already asked lots of questions on this is these threads
http://www.diystompboxes.com/smfforum/index.php?topic=89555.0
http://www.diystompboxes.com/smfforum/index.php?topic=89918.msg762786#msg762786

Might help?

tysonlt · January 09, 2012, 01:09:21 AM

Great, thanks. Some good homework there!

Have you successfully built something using cmos switching? Have any trouble with noise or pops?

potul · January 09, 2012, 03:17:14 PM

Regarding your question number 2...

These caps and resistors are there to remove the DC component and bias the signal to the midpoint. In each loop you will be connecting different devices (pedals and so on), and you never know what these will do to the signal. So I think you should keep them all.

Mat

tysonlt · January 10, 2012, 12:01:47 AM

Thanks mat.

If the effect is being bypassed and going straight from one cmos chip to another, does that path need to be biased as well?

potul · January 15, 2012, 10:37:45 AM

then it would be ok, because you've alraedy biased the signal once.

tysonlt · January 15, 2012, 08:19:24 PM

I think I need to understand the basics. What is actually being biased? I thought guitar signal was only about 0.005 volts... Do we want to increase that to 4.5 volts?

Could some kind soul explain what is happening to the electrons when they come in from the input jack? I need to understand things on a physical level. I am pretty confused by the series capacitor, because I thought they stopped the electron flow once they were 'full'.

And... I still don't understand the concept of a pullup! Feel the shame

DavenPaget · January 15, 2012, 09:51:27 PM

Quote from: tysonlt on January 15, 2012, 08:19:24 PM
I think I need to understand the basics. What is actually being biased? I thought guitar signal was only about 0.005 volts... Do we want to increase that to 4.5 volts?

Could some kind soul explain what is happening to the electrons when they come in from the input jack? I need to understand things on a physical level. I am pretty confused by the series capacitor, because I thought they stopped the electron flow once they were 'full'.

And... I still don't understand the concept of a pullup! Feel the shame

Somewhere in GEOFEX on a article about Opamps did he mention it .
It actually gives the input signal a voltage reference . "Ground"

News:

Chaining CMOS switches