analog switches and thd

Joe Grisso proteus at
Wed Jun 25 20:07:37 CEST 1997

Martin Czech wrote:

> I want to avoide a lot of relays in an multiple input switching
> matrix for an 8-channel audio system with 2x8-track, 2x-track, 2xCD
> and 2xDAT
> interface with independent level and EQ control for up to eight
> spatial distributed
> loadspeakers (is this called octophony ?).
> It's a DIY birthday present for a friend.
> Up to now, he has to run arround to regulate the sound at EVERY
> active loadspeaker (Yamaha), and there's always some awful
> switching noise during the performance, since he has no fade out
> box.
> Real studio quality is not required concerning THD, since the
> loadspeakers
> will allready contribute some distortion, it's for live performances.
> But too much noise would be awfull.
> Does anybody know, how much distortion (thd) is introduced
> by using a cheap CD4066 CMOS-Switch instead of an expensive relay
> @ Vdd=7.5V, Vss=-7.5V, Vsignal 1Vpp, signal current 100uA ?
> And what about noise ?
> Crosstalk will be no issue, I could use different ICs in such a case.
> Any experience out there ?
> m.c.

Any 40xx CMOS Switch have NASTY feedthrough, which when used as an audio
switch, will induce a very loud "pop" into whatever it is driving. My
suggestion would be to use a J174 JFET switch fed into the inverting
input of an op-amp (5532 prefferably). This not only is a much quieter
switch, but has more bandwidth, and can be amplified and trimmed with
the op-amp.

Just my $.02 worth...


________________________/ Talon Audio \_
Joe Grisso
CEO/Head Engineer
proteus at
jgrisso at

More information about the Synth-diy mailing list