DIYstompboxes.com

DIY Stompboxes => Building your own stompbox => Topic started by: Dimitree on December 14, 2016, 07:10:26 AM

Title: photorelay for bypass
Post by: Dimitree on December 14, 2016, 07:10:26 AM
hello
is this photorelay:
https://toshiba.semicon-storage.com/ap-en/product/opto/photocoupler/detail.TLP175A.html (https://toshiba.semicon-storage.com/ap-en/product/opto/photocoupler/detail.TLP175A.html)
suitable for stompboxes bypass?
they describe it as a mechanical relay replacement,
but I'd like to be sure that its specs are ok for audio (and if yes, what are the specs that tell us so)

thank you
Dimitri
Title: Re: photorelay for bypass
Post by: R.G. on December 14, 2016, 10:14:01 AM
That is a solid "maybe". Depends on the exact photo relay.

MOSFETs are a better variable resistor than JFETs, and so they're very good for switching analog signals, except for some special considerations.

The special considerations all relate to the substrate/body diode in the MOSFET.  A single MOSFET is a variable resistor and an even better off/on switch for analog signals, excepting that there is a built-in diode from the source to drain. This diode is not removable, because it's part of the way the silicon MOSFET is made. So signals over about a tenth of a volt get progressively more distorted as this diode turns on in one polarity.

MOSFET relays get around the problem by using two MOSFETs connected source to source, so their body diodes are opposed. What turns on one body diode turns the other off. That means that the MOSFETs do all the controlling, but there have to be two of them in series.

So the photo relay ought to function as an audio switch. But there may be oddities about how they function very near the zero crossing point - or may not. Depends on the MOSFETs and how they're driven.

The best advice I have is to try it with the photo relays you can get. You might like the "Clinton Bypass" from the Technology of Bypassing.

It's worth noting that the CMOS switch chips are in fact MOSFET devices arranged specifically to switch audio signals. They have the advantage that they can force the MOSFET substrates to be at the power supplies, not just the sources of the MOSFETs, and that they can use both N-channel and P-channel MOSFETs, so they can do a better job of "covering up" any oddities of the individual MOSFETs by using an N-channel and a P-channel MOSFET in parallel, not two Ns or Ps in series as in the photo relay.

I've always thought that a photo relay would make a good speaker relay, able to open up the speaker connection FAST!! if there was fault detected, but never had the time to test it out very well.
Title: Re: photorelay for bypass
Post by: midwayfair on December 14, 2016, 10:59:59 AM
It's similar to the TLP222G, which some people have used, but the datasheet says 5mS, which is slower. Slow enough to maybe notice when you switch. I still prefer the H11F1 and I don't think I'd use the linked chip over the 222G if I couldn't use it.
Title: Re: photorelay for bypass
Post by: Dimitree on December 14, 2016, 11:57:37 AM
thank you both
both TLP222G and H11F1 seems about 2x expensive then TLP175A. Is that just because of the 5mS timing or do they have overall better specs?
Title: Re: photorelay for bypass
Post by: Dimitree on December 15, 2016, 01:38:48 PM
in the meanwhile, I searched about TLP222G here on the forum and looks like everybody is using it to mute the audio while switching with a relay.
Why not directly use it instead of a relay?
Title: Re: photorelay for bypass
Post by: Transmogrifox on December 17, 2016, 11:16:21 AM
Quote from: Dimitree on December 15, 2016, 01:38:48 PM
Why not directly use it instead of a relay?

Try it and report back.  The company I work for uses these in certain products as an alarm output contact.  I always associate them with switching power, or at least 24/48V DC to contact inputs so the things that are not specified in the datasheets are the things that may be issues for audio signal levels.  You don't get to find that out until you try it for an audio application.

One such example that comes to mind is switch signal feed-through.  Probably for this the best bet would be something with a really high isolation voltage rating because by nature of the design the device would have a smaller capacitance between the input side and the contact outputs...but then if that was the problem you could switch the input more slowly so you don't apply a fast rate-of-change to feed through parasitic capacitance (common mode).