Channel switching in tube amp with solidstate relays

[dobrota]

Did anybody try it ?
E.g. Vishay LH15xx ? They seem to be almost ideal - linear operation, maximal signal amplitude 200V and more, isolation voltage 4500V, bounce free switching, low consumption...

[Steve Conner]

I've never tried it, but there are schematics over at Mark Huss's site.
http://mhuss.com/18watt/schematics/MosfetSwitchTrex.gif

It's also been discussed on this forum, try the search function.

I think the only drawback with them is the high shunt capacitance in the OFF state. It might require a series-shunt pair of switches to stop signal bleeding through.

[dobrota]

Yes, I also noticed relatively big capacitance, maybe it's major problem..

[J M Fahey]

I may be wrong, but I doubt very much these can replace a metallic contact relay as *audio* switches.
I see them as power switches, turning lamps, entire preamps or small motors on, rather than feeble audio signals.
I had a slight hope when I saw some of them labelled "Telecom switches" thinking they might be part of complex Telecom audio matrixes, but not, they handle on/off hook (switching from 24 to 48V DC) and 48V Pk dialing pulsed.
In short, they are not 300V capable CD4016/66/53 etc.
I see the 18W.com drawing closer to a brainstorming idea than an actual working schematic.
Of course, it might work as we hope.
In short, I see them closer to Triacs and such than to audio components.
If anybody bench-tested them, welcome.
I'll see if I can grab a pair.
EDIT: makers post no transfer curves at all !!!!
Besides, being Led triggered, they "eat" a lot of power for what they do.

[Don Moose]

There are the IR PVT series

Haven't tried them, and the capacitance issue is still there, but it does look suitable for signal switching.

Hope this helps!

[J M Fahey]

Maybe. Hope so.
At least they show a voltage referred to current graph, it seems to be linear and it seems to cross "0"
That's what worried me before, because I *have* worked with solid-state relays before, and they were meant to mimic an electromechanical leaf-type relay turning motors on and off and stuff like that, but they were useless for audio, they had a voltage threshold that distorted audio signals worse than crossover distortion, were not linear, etc.
If you turn-on, say, a garage door opener motor, whether it's 120VAC or even 12VDC, "losing" 1 volt is nothing, but on audio it's terrible crossover.
These *seem* to be linear.
What I'd love to have is whatever Telecom companies use on matrix switchboards.

[Steve Conner]

JMF: You get two different types of solid-state relays, MOSFET and triac. You may have had triac ones. The MOSFET type are what we're discussing here, and they are theoretically linear ("ohmic"), just like JFET and analog switches, but I can't vouch for them in practice as I've never tried them.

I've used the H11F1 JFET optocoupler in amp circuits before, as a switch and a voltage-controlled soft clipper, but it's a bit of a wuss.

The telecom stuff is all digital now, but they used to use relays and uniselectors.

[raiken]

I used some Panasonic AQW210E dual SSRs in an SPDT configuration to switch between two preamp volumes and they didn't work worth a damn.

On paper, they looked great, high voltage, low resistance, etc.

In practice, however, they didn't turn off properly with a high-impedance source and load (driven from a 1Meg pot into a 1Meg grid load). I had to lower the load resistance way down before the off channel wouldn't come through at any appreciable level.

I don't know if it was capacitive bleed-through on the off switch, or just something with the way the signal modulates the MOSFET channel, or what, but I was in a hurry to get a prototype done and didn't get a chance to further investigate it, I just ripped them out and put in some real relays.

I plan to look into them further when I get a chance. I suspect they would be great for shunt switches, but I don't know about series switches - you may have to use an L configuration with a series and shunt to ground.

The high capacitance of some of them may also adversely affect frequency response or inadvertently pass through very high frequencies in the off state (which would add a "buzz" into your clean channel if you were switching between clean and dirty).

Randall Aiken

[Steve Conner]

Thanks for the reality check, Randall!

I had similar weird problems when using the H11F1 as a clipper. If it was biased partway on, the signal would couple capacitively into the floating gate, and modulate it on and off, making nasty, funky sounding distortion. I never got it sounding quite right, and went back to a Marshall-style diode bridge clipper. But H11F1s worked fine for me in other applications, like changing gain by switching cathode resistors.

Maybe the MOSFET SSRs also suffer from this kind of "Miller effect" - if that's the right name for it, when you use them at audio frequencies.

Someday I want to try back-to-back power MOSFETs as a speaker protection "relay" in a hi-fi power amp. I wonder what their THD measures like.

[raiken]

Someday I want to try back-to-back power MOSFETs as a speaker protection "relay" in a hi-fi power amp. I wonder what their THD measures like.

I'm going to be looking into some of those myself, soon. If I come up with anything interesting, I'll let you know.

RA

[J M Fahey]

Hi Steve.
Yes, you are right, the SS relays I used repairing industrial machinery *must* have been Triac based, yet I never got one open, though.
I know these are Mosfet based, I've seen the datasheets, and precisely there lies my doubt: Mosfets are great for switching, *can* be used for audio, but unfortunately they are enhancement devices (that by itself is a small nuisance but not a fault) *BUT* they are *very* nonlinear for the first 2.5 to 3.5 volts, just look at the curves.
I guess that must produce something that sound very close to crossover distortion, which we all know is horrible .
What you and others add about crosstalk and feedthrough make me want to stay away for now.

[Merlinb]

In general it is always better to use FET switches as shunt switches. If the source resistance is large, then sometimes this is all you need.

They can be used as series switches too, but you ideally want a low source and load impedance. Unfortunately they have higher resistance at low signal levels (tens of ohms), so you get cross over distortion. But if you load them with a high resistance then you fix the crossover distortion, but capacitative bleed-through and poor offness become a problem, so you pretty much always need a 2nd one to act as a shunt and improve the offness. Using more than one shunt in parallel improves the offness of course (6dB improvement for every doubling of the number of FETs you use).

[J M Fahey]

Yes, that's true.
Fact is, we have many options: Fets, CMos gates, even bipolars shunting signal to ground, if we work with low voltages, say , 15Vpp and a few handle 30Vpp, but for higher levels as usual in tube amps, we are stuck with either LDRs or relays.
Anyway we should try everything new, someday ......

[Michael Allen]

Just finished a 5E3 with an added stage switchable by SSR. Used two avago chips, switching between volume controls after the first stage and then directing the second stage either to the PI or into the third stage for overdrive. I think it works well, no issues to speak of. I don't think there are any audio degradation issues or incomplete shut off. I can't hear any overdrive coming through on the 5E3 channel. Perhaps it depends on where in the circuit you want to use them, but unlike Randall's post above, they worked well for my switching preamp volume controls...

[J M Fahey]

*VERY* interesting, please post what you implemented.
Thanks.