Using Jfets for switching in tube amps - high voltage concerns

[Gaz]

The 1M/100nF combo has a 100msec time constant which means that after it shorts the gate (LED in your case) to ground it should release shortly after that.
With disconnected SSR measure the voltage at the 1M/100nF junction. Then try a regular LED with resistor to 12V and cahtode connected as I suggested earlier and let us know.

Thanks, I did the LED test and it was still behaving in the same way (LED getting stuck on for several seconds). I checked the connections to the MAC97 again, and all was good so I replaced it with another I had just in case I had overheated it.

The new one behaves differently now, but still doesn't work correctly. Now the LED lights up when the relays are turned OFF, but stays on. When relays are switched back on the LED does not light at all. The effect with SSR connected is it mutes one channel and not the other (and the pop comes through on the unmuted channel as well). Hope that makes sense.

[Gregg]

Well, at least now we know that the problem is somewhere on the driving circuit side and the SSR is not the problem.

[Gaz]

Yes, true. I also hooked up J175 again, and circuit works as expected, so I think perhaps there need to be some modification for it work with an SSR, but it's beyond my ken to reason why it works with the p-channel fet and not the LED. I attached a drawing of exactly how I have it hooked up. Maybe I'm missing something.

[nickb]

Yes, true. I also hooked up J175 again, and circuit works as expected, so I think perhaps there need to be some modification for it work with an SSR, but it's beyond my ken to reason why it works with the p-channel fet and not the LED. I attached a drawing of exactly how I have it hooked up. Maybe I'm missing something.

[ATTACH=CONFIG]57215[/ATTACH]

With a 2.2K resistor the SSR LED current is about 5mA. The holding current for a MAC97 triac is typically 1mA and may be much less than that. Therefore the triac turns on and stays on. Enzo mentioned this above. A small triac just isn't suitable to use with an SSR. I suggest you use a transistor instead. The time constant with 0.1uf and 2.2k is going to be too short. You'd need more like 50uF but the current would need to be limited to avoid damaging the transistor, perhaps a 120 ohm in series with the transistor. I think it would be better to move the timing to the base circuit rather than the collector. Try 10K instead of 47 ohms and 4.7uf instead of 0.1uF. You can omit C2 in your schematic. Just about any small switching NPN transistor will do.

[Enzo]

Well sorta. The reason they used a triac in the first place - in my opinion - was because it will fire in the quadrants. In other words a trigger pulse can go either direction and trigger the triac. A transistor would only be triggered by a positive pulse (or negative depends upon NPN or PNP)

[Gaz]

With a 2.2K resistor the SSR LED current is about 5mA. The holding current for a MAC97 triac is typically 1mA and may be much less than that. Therefore the triac turns on and stays on. Enzo mentioned this above. A small triac just isn't suitable to use with an SSR. I suggest you use a transistor instead. The time constant with 0.1uf and 2.2k is going to be too short. You'd need more like 50uF but the current would need to be limited to avoid damaging the transistor, perhaps a 120 ohm in series with the transistor. I think it would be better to move the timing to the base circuit rather than the collector. Try 10K instead of 47 ohms and 4.7uf instead of 0.1uF. You can omit C2 in your schematic. Just about any small switching NPN transistor will do.

Thanks, Nick! I actually missed Enzo's reply (which I think I would have missed the point of with the triac anyway). Before I try to hack this up, I redrew the schematic. Is this what you meant?

[Gaz]

Well sorta. The reason they used a triac in the first place - in my opinion - was because it will fire in the quadrants. In other words a trigger pulse can go either direction and trigger the triac. A transistor would only be triggered by a positive pulse (or negative depends upon NPN or PNP)

Oops, missed this before my last post. I think you're correct because the relay clicks with high gain can be heard turning the relays on or off. It's seems there's more than one reason they don't use SSRs. I understand the tried and true method is cheap and simple as you said, but it would be nice to have an alternative muting device that could handle higher signal voltages, especially when trying to use relay switching in vintage style designs where there's not much opportunity to bring the signal down to safe levels.

[nickb]

Well sorta. The reason they used a triac in the first place - in my opinion - was because it will fire in the quadrants. In other words a trigger pulse can go either direction and trigger the triac. A transistor would only be triggered by a positive pulse (or negative depends upon NPN or PNP)

Yes indeed, thx. I missed that point. I had wondered why they chose a triac and what you suggests make complete sense. So, no, a transistor will solve the hold issue but won't get the dual edge switching unless you add quite a bit more complication.

The original Peavey design is starting to look close to optimal.

[nickb]

Thanks, Nick! I actually missed Enzo's reply (which I think I would have missed the point of with the triac anyway). Before I try to hack this up, I redrew the schematic. Is this what you meant?

[ATTACH=CONFIG]57224[/ATTACH]

Well, not so fast with the praise. Yes you understood

But in order to do both edges you need yet another transistor. [Schematic deleted as it was wrong]

[Gaz]

Nick, it works! However, when the relays are switched ON the mute is shorter. I can hear the click come through and visibly see the LED flash for a shorter period when I have it hooked up in place of the SSR.

Any reason you can think of for this? Right now (if I understand correctly), the mute should be about 47ms right with the 4.7uF/10K? I think it's right when the relays shut off, but not on.

[nickb]

Nick, it works! However, when the relays are switched ON the mute is shorter. I can hear the click come through and visibly see the LED flash for a shorter period when I have it hooked up in place of the SSR.

Any reason you can think of for this? Right now (if I understand correctly), the mute should be about 47ms right with the 4.7uF/10K? I think it's right when the relays shut off, but not on.

Sorry. It was late last night. With fresh eyes this pnp/npn thing will never work right. The led will never get fully turned off.

Honestly, I cannot think of a simple way to make this work with an SSR. The best I've got is to go a variant of the original JFET design but instead of the JFET doing the muting, it drives the SSR. Something like this:



This begs the question, can't you just use the JFET on it's own?

[Gregg]

There's another way to pick up the relay spike in both directions using an opto element with back to back diodes like the one on the schematic below.
The other ways is to use Mesa's mute circuit but in order for it to work properly you need at least one relay switching on and at least one off.

[Gaz]

Sorry. It was late last night. With fresh eyes this pnp/npn thing will never work right. The led will never get fully turned off.

Honestly, I cannot think of a simple way to make this work with an SSR. The best I've got is to go a variant of the original JFET design but instead of the JFET doing the muting, it drives the SSR. Something like this:


[ATTACH=CONFIG]57231[/ATTACH]
This begs the question, can't you just use the JFET on it's own?

Hey, yeah, I did notice (not at first) that the signal was reduced even with the LED "off."

That's cool, another idea to try, and really the original circuit is fine, but using an SSR means I could stick the mute anywhere in the circuit even with high signal levels. I am out of town for a few weeks so can't try anything until I'm home :/

[Gaz]

There's another way to pick up the relay spike in both directions using an opto element with back to back diodes like the one on the schematic below.
The other ways is to use Mesa's mute circuit but in order for it to work properly you need at least one relay switching on and at least one off.

[ATTACH=CONFIG]57232[/ATTACH]

Thanks Gregg, this is a cool schem too. Did you come up with this? What are the resistors for? I will have to try that when I get home in a few weeks too.

[Gregg]

It's not my idea. It comes from the link below (Pops, clicks section but there's other interesting info there as well) and it's all explained. I just added the SSR. It's tested and works.

http://guitaramplifiers.prv.pl/misc.htm#Pops