Totally silent switching circuit - FINALLY!!

[tempus]

Hey all;

Here it is:


After much experimentation, I finally found a circuit that switches completely silently (on my breadboard anyway - hopefully there aren't any other problems to creep into my now perfect universe). I plan to control this from a uController. The 2nd MOSFET shunts the signal to ground, and will always be on during any switching, after which it is immediately turned off (or kept on to bypass). 2 of these switches will be required for each effect.

Thoughts?

Thanks

[JKowalski]

Congrats!

Wouldn't it be a lot nicer to have one control input, with low and high? It seems awkward to have two inputs needing separate switching. Is there a reason for this?

[cpm]

Wont body diodes produce clipping?
my guess is that if audio path is biased around Gnd, when the "switch" is on it would clip to ground on positive peaks, and when its off, it would leak negative peaks. is it so?

Also, if used as a building block for a spdt or a/b sort of switch, being the output put to ground, you cant just tie two outputs together from complementary switching blocks, as it would result in the output being always grounded

[snap]

crosslink:
http://www.diystompboxes.com/smfforum/index.php?topic=76209.0

[tempus]

Wouldn't it be a lot nicer to have one control input, with low and high? It seems awkward to have two inputs needing separate switching. Is there a reason for this?

It would, but for this to be silent, the timing has to be right. The 2nd MOSFET shunts the signal to ground, which is why it switches silently. Is there a way to do this using one control input?

Wont body diodes produce clipping?

I thought they would too, but no matter how hard I pound my guitar, there is no distotion of the signal.

Also, if used as a building block for a spdt or a/b sort of switch, being the output put to ground, you cant just tie two outputs together from complementary switching blocks, as it would result in the output being always grounded

I'll look into this further, but I think it will work as posted. There will be one of these circuits on the input of the effect, and one on the output of the effect. When I want to bypass the effect, the input will be selected, which means that the signal will be passed on to the next effect in the chain. Essentially a pair of these work as a SPDT switch selecting between the input and output of each effect.

[igor12]

I notice the BS170 has an Rds(on) of 1.2 ohms.  That's pretty sweet! Do you hear any differences in bypass sound?

[tempus]

The low on resistance is one of the reasons I tried this particular device, although something with a higher on resistance probably wouldn't make much difference. That being said, there is a limit to how much you want to put in the signal path. I'm going need 2 of these for each effect, which for me would be a total of 8 BS170s in series, plus the on resistances of the JFETs in my effects pedals. So it does make sense to try to design for the best possible case, which 1.2 ohms of on resistance definitely helps with.

There is no difference at all (to my ears) between bypassed and non-bypassed sound.

[R O Tiree]

How about having one "logic high" and using a NOT gate (or ganging the 2 inputs of a NAND) to provide the opposite sense?

Even better would be a dual flip-flop - that would provide the 4 logic states you need to drive the 2 of these switches you'd need for each effect - 2 high and 2 low, swapping every time you click the foot-switch. One dual flip-flop chip, a footswitch, 6 resistors, 2 caps, 4 BS170s. Cheap, simple, not much PCB footprint. You'd need to connect the /Q output from one of the flip-flops to the other one's input as well as to its respective BS170, so that one flips whenever the other flops.

[MetalGuy]

After much experimentation, I finally found a circuit that switches completely silently (on my breadboard anyway - hopefully there aren't any other problems to creep into my now perfect universe). I plan to control this from a uController. The 2nd MOSFET shunts the signal to ground, and will always be on during any switching, after which it is immediately turned off (or kept on to bypass). 2 of these switches will be required for each effect.

It's been done long time ago but most of the companies don't like MCUs although in the last years they're getting over it.
I started using MCUs for switching 5 years ago and it works very well. In my designs I'm using J174 and J175. You can also use LDRs or other transistors as well.

[tempus]

How about having one "logic high" and using a NOT gate (or ganging the 2 inputs of a NAND) to provide the opposite sense?

Even better would be a dual flip-flop - that would provide the 4 logic states you need to drive the 2 of these switches you'd need for each effect - 2 high and 2 low, swapping every time you click the foot-switch. One dual flip-flop chip, a footswitch, 6 resistors, 2 caps, 4 BS170s. Cheap, simple, not much PCB footprint.

I'll be using this circuit to switch in and out different combinations of 4 different effects (plus amp functions) in a switching matrix, so a uController would be a far simpler way to implement this particular setup, and would take up much less board space. In this case, a toggle switch (like a flipflop) wouldn't be of use.

Thanks

[R O Tiree]

Aah... I hadn't appreciated that what you are trying to switch in and out was far more complex than 1 or 2 effects, in which case you need a cleverer logic system, fer sure.

Might still be worth a look-see for a simpler system...

[tempus]

Simpler than a uProcessor? I think that's about as simple as it can get isn't it? I mean, all I need is 2 output pins on the micro for each effect, which will connect directly to the circuit I've shown here. My current setup is more discrete, and it's got latches, inverters, relay drivers, and more diodes than Mouser in it. The micro will simplify all that and be easily reprogrammable.