Two-channel switching

[pedaltothemetal]

I need a solution to allow digitally-controlled switching between two different audio signals and I'm trying to stay away from mechanical relays.

I will only have 9V power so solutions requiring bipolar power won't work for me.

My current solution is MOSFET-based SSRs in series/shunt configuration... The problem is the isolation between the two channels sucks. I can still hear the other channel quite well when it's supposed to be off due to leakage current.

I will have from 5V-9V available to switch the relay. Switching frequency is low as this is for a human-operated control.

Anyone have any suggestions? If worse comes to worse, I'll go for a mechanical relay. Anyone have any favorites?

[dschwartz]

fets?

[composition4]

I like optoFETs and optoresistors.  Or photoFETs and photoresistors, however you want to call them

H11F1 optoFET works well for series/shunt switching, despite what a lot of people claim (maybe if they tried it they'd hear how well it works)

An even better series/shunt uses the Silonex NSL-32SR3, NSL-32SR3S, or even NSL-32SR2 in some cases (http://www.silonex.com/audiohm/index.html).  Whichever way you go, you can set it up like the following diagrams:

http://www.silonex.com/audiohm/softswitch.html

Figure 4 might be what you're after

Let me know if you have any q's

Jonathan

[potul]

What about CMOS switching? CD4066 style.

[sean k]

Excuse my ignorance but could you elaborate on what digital switching actually means. Even digital sources would, I suppose, use some conversion system to control analog style devices like fets or cmos switches or even relays.

For control, and switching of Pulsed DC I was told 4066's would work after trying plain old transistors and having the base voltages effect the Pulsed DC signal I was trying to switch on and off. So in my ignorance I'd like to know what the digital control signal is.

[composition4]

I understand his question as wanting to have the control voltage to be from a logic source (i.e. high/low, 0/1)

[pedaltothemetal]

I like optoFETs and optoresistors.  Or photoFETs and photoresistors, however you want to call them

H11F1 optoFET works well for series/shunt switching, despite what a lot of people claim (maybe if they tried it they'd hear how well it works)

An even better series/shunt uses the Silonex NSL-32SR3, NSL-32SR3S, or even NSL-32SR2 in some cases (http://www.silonex.com/audiohm/index.html).  Whichever way you go, you can set it up like the following diagrams:

http://www.silonex.com/audiohm/softswitch.html

Figure 4 might be what you're after

Let me know if you have any q's

Jonathan

Hey dude, thanks for the suggestions. I'm aware of the LDRs and stuff and they'd work great, but they're a bit spendy. The problem with figure 4 (series configuration) would be the leakage. Series-shunt works with no leakage for one channel, but you can't combine two channels like a SPDT with that configuration as one will ground the other, muting them both.

The H11F1 looks interesting but on resistance is a bit high.

[pedaltothemetal]

I understand his question as wanting to have the control voltage to be from a logic source (i.e. high/low, 0/1)

Yes, that's correct.

[pedaltothemetal]

What about CMOS switching? CD4066 style.

You'd need a bipolar supply (or a suitably biased input) for audio and I won't have the necessary power for that.

[composition4]

I like optoFETs and optoresistors.  Or photoFETs and photoresistors, however you want to call them

H11F1 optoFET works well for series/shunt switching, despite what a lot of people claim (maybe if they tried it they'd hear how well it works)

An even better series/shunt uses the Silonex NSL-32SR3, NSL-32SR3S, or even NSL-32SR2 in some cases (http://www.silonex.com/audiohm/index.html).  Whichever way you go, you can set it up like the following diagrams:

http://www.silonex.com/audiohm/softswitch.html

Figure 4 might be what you're after

Let me know if you have any q's

Jonathan

Hey dude, thanks for the suggestions. I'm aware of the LDRs and stuff and they'd work great, but they're a bit spendy. The problem with figure 4 (series configuration) would be the leakage. Series-shunt works with no leakage for one channel, but you can't combine two channels like a SPDT with that configuration as one will ground the other, muting them both.

The H11F1 looks interesting but on resistance is a bit high.

Not sure how you mean... in Figure 4, if OP1 is ON and OP2 is OFF, then INPUT2 won't "ground the other", as it is separated by OP2's off resistance ( which is up to 25M).  Leakage will be inaudible with the impedances you're dealing with...

The on resistance of the H11F1 is under 100 ohms from memory... again, a series resistance of 100 ohms most likely won't make a difference with impedances you're dealing with.

I don't disagree with you about the prices though!  I picked up a bunch of H11F1s for a design of mine and didn't pay much (i.e. less than a dollar each), but seeing the prices after a quick search now, it looks like I got lucky

Jonathan

[potul]

What about CMOS switching? CD4066 style.

You'd need a bipolar supply (or a suitably biased input) for audio and I won't have the necessary power for that.

It's  simple question of biasing the input signal, you don't need more power for that. Can't you do that?

[pedaltothemetal]

What about CMOS switching? CD4066 style.

You'd need a bipolar supply (or a suitably biased input) for audio and I won't have the necessary power for that.

It's  simple question of biasing the input signal, you don't need more power for that. Can't you do that?

I have a signal that will potentially swing more than 1/2 the power rail positive or negative.

[pedaltothemetal]

Not sure how you mean... in Figure 4, if OP1 is ON and OP2 is OFF, then INPUT2 won't "ground the other", as it is separated by OP2's off resistance ( which is up to 25M).  Leakage will be inaudible with the impedances you're dealing with...

The on resistance of the H11F1 is under 100 ohms from memory... again, a series resistance of 100 ohms most likely won't make a difference with impedances you're dealing with.

I don't disagree with you about the prices though!  I picked up a bunch of H11F1s for a design of mine and didn't pay much (i.e. less than a dollar each), but seeing the prices after a quick search now, it looks like I got lucky

Jonathan

Figure 4 isn't a series-shunt config, it's purely series and it is incapable of completely silencing the signal.

Series-shunt configuration is capable of silencing the signal, but cannot be used to switch between two channels like a SPDT simply because the shunt portion will ground both sides.

On resistance doesn't make a difference for a series switch but makes a big difference for the shunt switch.

[potul]

I have a signal that will potentially swing more than 1/2 the power rail positive or negative.

Oh, I see. If your signal can be higher than 5v peak to peak CMOS will not work. In this case I would go for relays.

Potul