What Tech 21 has to say and show about true bypass

[greigoroth]

Thanks for the ideas guys

[mojotron]

Andrew (TTG)  has an excellent set of ideas here too. Besides some simple debouncing in some cases, I really don't see the down side to CMOS switching. Here is a link to Andrew's "Wicked Switches" write up:

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

And RG has another excellent write up of this as well in his "Bypassing and Switching with the CD4053 CMOS Analog MUX" write up:

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

Thoughts?

[pjwhite]

4066-type switches are OK for switching into a high impedance circuit, but you should be aware that the switch ON-resistance is not a constant.  It varies with the level of input signal, temperature and the power supply voltage.  With a poor circuit design, this can lead to significant distortion and signal loss.

Here are some excerpts from the TI data sheet for the 4066.  This first shows typical switch ON-resistance over a range of input voltages at different temperatures with a 5 volt power supply.


Same again for a 10 volt power supply:


And for 15 volts.


You can see that the variation in ON-resistance is lowest with the 15 volt supply.

This shows, together in one graph, the range of ON-resistance for the three different power supply voltages.


The result of these variations in resistance can be minimized by:
1. Keeping the input signal low (near 0 volts).
2. Keeping the impedance of the circuit that that the switch feeds much higher than the expected switch ON-resistance.
3. Use a high power supply voltage.

[Mark Hammer]

VERY nice summary, and useful information.  Many thanks.

[Ben N]

That Tech21 article seems to me to be mostly self-serving. I immediately had the same reaction as Jack about the swiitches, and my feeling is where someone tries to snow me with one point, the rest loses credibility too.