more on electronic switching, coupling caps, internal restance

[JN]

I'm debating whether to try RG's method and or Andrews.  I feel RG's might be better for switching values WITHIN an effect, because it seems to require coupling caps to isolate from the chip.
 
See RG's use of the CD4053 here,
http://www.geofex.com/Article_Folders/cd4053/cd4053.htm

Andrews use of the CD4066, does not need coupling caps, so you can get a truer bypass, as it were.

Right so far?

So for the CD4066, as seen in the data sheet here,

http://pdf1.alldatasheet.co.kr/datasheet-pdf/view/26882/TI/CD4066.html

Am I to asume that "125ohms typical on-state resistance for 15V supply"  means that when the switch is closed there is 125 ohms of resitance in the signal path?  Because that stinks.

Please help.

Thank you.

[R.G.]

Here's a novel idea: why not try both?

The CD4053 and 4066 have very similar internals. What is different is just the way the switches are controlled. Both will switch signals at or near ground or the power supply, or anywhere in between.

With the CD4053 setup I was solving a different problem. CMOS switches pop least if the DC level of what they are switching happens to be dead in the middle of the switch's power supply. I was choosing to force that with DC biasing. But the 4053 will switch without those DC levels and caps, too.

You can also do this another way with the CD4053. The 4053 lets you put a negative voltage on the Vee pin. Then you can switch signals at ground level with no popping too. A charge pump would make for an easy way to do this. Or you can just go ahead and switch signals at ground without the biasing and capacitor isolation.

As for something being a truer bypass; it may use less parts, but resistors and caps are pretty "true".  The ugliness in the eye of the true bypass fanatic should be in the fact that the signal touched a silicon device in the active switch, not that it had to be coupled through caps. Maybe what you meant was that it's closer to the "no other parts needed" model of a hard switch.

And that's the subject of your next question - the on resistance of the semiconductor switches. You read the data sheet correctly. When the switch is on, it's typically 125 ohms. Whether that stinks or not depends on (a) how much signal loss it gives you, and (b) whether it distorts the signal.

On (a), we know that amps and many effects have input impedances of 1M or more. What's the signal loss if we have 125 ohms in series with 1M input impedance?

Well, it's 1M/(1M+125 ohms) = 0.999875    or a loss of 0.001085668 db. Since humans perceive a change of 6db as a just perceptible change in loudness, it's likely that minus one thousandth of a db is not perceptible.

There exist analog switches with much lower on resistance, down in the 1 -20 ohm range. They're not as cheap, though.

On (b), the on resistance of a CMOS switch is not constant. It varies with the nearness of the signal to either V+ or V-. And CMOS switches are flattest right in the center.

Better CMOS switches will specify flatness. Then again, it depends on that you are driving. Driving a 1M load with 125 ohm resistor, it doesn't matter much, because if you can't hear the whole 125 ohm resistor, you're unlikely to hear it wavering around a fraction of the 125 ohm value.

So try them. You may like one of them.

[JN]

RG, thanks for your explanations.   They are encouraging.

I understand your points, and think I should simply try both.

But, two quick (I think) quesitons:
1.  can you reduce the the closed switch resistance in half by stacking the chips without any will effects?
2.  Do you know of any lower resistance electronic switches?

Maybe I am falling prey to purist hype, but I think there might be circumstances-- granted bedroom ones-- where you might hear the effects of an extra 125 ohms.  Probably not, but mabye.

Thanks,
Jake

PS thanks for the grat detail on the cd4053 in the previous post.

[Pushtone]

I'm debating whether to try RG's method and or Andrews. 

I've reading up on this too.

But what is "Andrews" method. Is there something in the Gallery?
Please post a link to a schem / layout  if possible.




Very helpful reply R.G., thanks.

[puretube]

T.T.G. "wicked switches..."

[Pushtone]

T.T.G. "wicked switches..."

oh, that Andrew.

Thanks

[The Tone God]

That article is starting to show it's age. I should update it. When I wrote it I had not gotten around to trying the 4053 for audio switching as I usually used the 4066 because the whole IC made the DPDT that I needed and the LED was controlled by a spare gate. I have since played with the 405x series so I should update for that as well as address a number of other things that I keep getting questions about.

I never made very clear that the way I was discussing how to use the switches was for general purposes not specificlly for audio. The switches can be used to switch other types of signals like DC control signals (i.e. switch between LFO and manual sweep control) or digital signals for more complex arrangements. This why I choose to leave decoupling/bias networks off. If someone needed one they could add it themselves with little trouble.

With the use of the switches in audio situations I took some liberties and made the assumption that the user would use the same supply as the effect. It is likely the circuit has been designed to not exceed the power rails thus the signal would be within the range of the switching IC because it would be sharing the same supply. Also I make the assumption that if the switch is being used to bypass an effect it would have decoupling caps and pull down resistors to keep the signal at or near center.

Just a few thoughts.

Andrew

[Pushtone]

I never made very clear that the way I was discussing how to use the switches was for general purposes not specificlly for audio. The switches can be used to switch other types of signals like DC control signals (i.e. switch between LFO and manual sweep control) or digital signals for more complex arrangements.

Andrew

Funny you should say that. I've been struggling with switching controls sets (pots) in my dual Small Clone Chorus build.
I figure it would take 5 SPDT switches to switch the RATE and DEPTH pots. Which is why I'm so keen on this thread.

Thanks for pointing out the "barebones" aspect of your design. Helps my understanding.

So if I wanted to do the pot switching in the Small Clone Chorus I could go "barebones" and get the job done with the minimum parts?

On the other hand, for an audio switch application, the decoupling caps and pull down resistors are cheap insurance.

Have I got that right?

[The Tone God]

Funny you should say that. I've been struggling with switching controls sets (pots) in my dual Small Clone Chorus build.
I figure it would take 5 SPDT switches to switch the RATE and DEPTH pots. Which is why I'm so keen on this thread.

Thanks for pointing out the "barebones" aspect of your design. Helps my understanding.

So if I wanted to do the pot switching in the Small Clone Chorus I could go "barebones" and get the job done with the minimum parts?

Well I can't give a definitive answer as I have not evaluated the signals you wish to switch but I have done similar things with them so it may be possible. The decoupling caps suggested for audio use would block the DC levels that some of those controls may need hence why I left my design "barebones".

On the other hand, for an audio switch application, the decoupling caps and pull down resistors are cheap insurance.

Pretty much. As mentioned earlier I took some liberties with my designs. I can get away with that as I design all my stuff from scratch and making sure my audio signals are conditioned before they hit the switches. I have not used them with other people's designs. The decoupling then rebiasing is a good "catch all" to make sure the signal(s) are conditioned to a known state. You can try to see if you can get away without it but you are flipping a coin at that point if you have something dedicated to that design like a PCB.

Good luck.

Andrew