Switching devices at tube amp voltage levels

[Kona311]

Hi, interesting forum. I found this site while trying to google an answer to a problem I have. I am trying to use some type of transistor (bjt? Jfet? Mosfet?) to switch between a number of interstage coupling capacitor options in a tube preamp. Essentially I plan to build some kind of logic to turn on or off different combinations of these coupling caps or filters. My plan to deal with the high voltage in tubes is to first use a high voltage, high capacitance cap to block DC from the previous stage. Since grid voltages are usually low, once the high voltage from the plate is blocked, I could get away with low voltage components.

The issue I am unsure of is which device is best suited to this purpose? jfets seem best but I am concerned about the high voltage swings of tube amp signals, 100 to 200 volts. Assuming the transistor voltage ratings would have to be rated for the swing of the signal, no? If that's the case then jfets are all but out. Mosfets would be next, but I hope I don't need a power mosfet that costs over a dollar each, as I was hoping to use a few dozens of them. Which leaves me with bjts.

If a schematic of what I am trying to do would help explain it better, I can upload one. I am not married to any components, and it is safe to assume I will have access to at least a 12V DC source in addition to the high voltage source for the tubes. Ideally the situation would be achievable using discrete through-hole components. Keeping an analog audio path.

Is what I am trying to do feasible without breaking the budget? Is there an effective way to use solid state devices to route signals that are hundreds of volts in amplitude?

Thanks for any help.

 

[MikeMl]

I'd say High Voltage NFET or PFET, but watch out for the body diode.

 

[Kona311]

I'd say High Voltage NFET or PFET, but watch out for the body diode.

Do you have a specific device in mind that might fit the bill?

Body diode... I feel like I should know why you are warning me about that but I'm drawing a blank as to its significance. I'll google it.

 

[crutschow]

If you are switching interstage signals in a tube circuit then I doubt that the signals are more than a few volts AC. Please post a schematic.

 

[Kona311]

If you are switching interstage signals in a tube circuit then I doubt that the signals are more than a few volts AC. Please post a schematic.

I better see if I can take some readings on an oscilloscope then. I'll do that before anything else.

 

[rumpfy]

I'll back crutschow.
Audio preamps never had to deliver more than about 1 volt peak to an audio power amp to drive it to full power with valve equipment.
There is a device, an HEF 4052 which is a dual 4 channel multiplexer used commonly in CD/tuner/tape/aux selection in commercial audio sets. I think it uses a bcd code to select the channel.
hope this helps.

 

[Kona311]

Maybe I should have been more specific. I'm talking about a guitar preamp designed for taking an input signal and distorting the heck out of it.

 

[crutschow]

Maybe I should have been more specific. I'm talking about a guitar preamp designed for taking an input signal and distorting the heck out of it.

If it has a low level output then all the internal signals would also be low level.

 

[Nigel Goodwin]

To be 100% safe and sure (and provide the best possible quality - pointless though that is with valves) use relays.

Anything else is just a poor imitation of a relay, and doesn't work as well.

 

[WTP Pepper]

I too would choose relays. They do have the tag of being old fashioned devices, but my plasma TV has them in as does some of our state of the art test equipment at work. My Hi-Fi amp uses them as anti-thump switches.
You can also more or less rule out your signal path through MOSFET switches adding to any distortion.

You also say you are going to use a HV cap to remove plate voltages, then end talking about signals hundreds of volts in amplitude. Just how many watts of output power are we talking about?
e.g 100VRMS across 4 ohms is a serious amount of output power.

 

[Nigel Goodwin]

You also say you are going to use a HV cap to remove plate voltages, then end talking about signals hundreds of volts in amplitude. Just how many watts of output power are we talking about?
e.g 100VRMS across 4 ohms is a serious amount of output power.

You didn't read the original post

It's for an effects pedal, so you overdrive the valves and then attenuate the output back down to a sensible line level.

So 'no watts' of output.

 

[dr pepper]

My 'scope has loads of relays inside.

An mpsa42/92 can handle fairly high voltages, have you considered a diode switch?

 

[Kona311]

Thanks for the replies. Yes, this is just a preamp. A few gain stages, some basic passive tone controls, then lower the output back down if need be to levels suitable to be handled by a power amp.

I will look into that mpsa42/92 mentioned. Relays are possible but are going to draw much more current than the transistor would, no? That's not a major concern since I am building this to sit in a rack, but it could impact transformer choice. As I said, I was hoping to switch dozens of these using a micro controller programmed to turn on and off certain combinations to give different results. I also thought relays would give more of a pop than solid state devices would.

I measured the output of my preamp prototype with all knobs set to max. There is some losses due to a passive fender-style tone stack and I planned to use a few more stages. The voltage coming out was in excess of what my oscilloscope could read when using my guitar as the input. So greater than about 40 volts peak to peak at least for the initial impulse of the plucked string, plus some time to diminish a bit. For a swept 20Hz-20kHz signal, I was measuring a voltage gain of around 115 depending upon frequency, but not drastically different.

I guess the million dollar question: for a B+ of 315Volts before filtering and typical preamp biasing (100k plate resistor, 1.5k or so cathode resistor bypassed by a large value cap, 1M grid resistor to ground "self biasing" I believe that configuration is called), what is the maximum voltage swing one could expect before hard clipping occurs? I should be sure anything that I use can handle that peak to peak voltage level, plus some margin to be on the safe side.

Thanks for all the replies.

 

[WTP Pepper]

You didn't read the original post

It's for an effects pedal, so you overdrive the valves and then attenuate the output back down to a sensible line level.

So 'no watts' of output.

Sorry I must of missed the reference to an effects pedal. Especially if you re-read the last sentence in the OP

 

[MikeMl]

The trick is not switching things like capacitors and resistors in/out of the circuit. The trick is to do it without introducing a huge thump (transient) into the output...

 

[Nigel Goodwin]

Sorry I must of missed the reference to an effects pedal. Especially if you re-read the last sentence in the OP

Perhaps I extrapolated 'effects pedal' , he actually said "switch between a number of interstage coupling capacitor options in a tube preamp."

Which to me reads as an effects pedal

 

[Kona311]

The trick is not switching things like capacitors and resistors in/out of the circuit. The trick is to do it without introducing a huge thump (transient) into the output...

I could always time delay the function if I'm using a microprocessor to mute the circuit temporarily while the switching is being performed. The price is right on that MPSA42

 

[dr pepper]

Said 'scope contains reed relays, these burn more power than a transistor but still not that much.
If the preceeding side of the relay contact/transistor is decoupled with a cap there shouldnt be a pop, dc is the worst offender for causing that.
I'm not a valve expert however grid leak biasing isnt great, you'll get less voltage swing, cathode biasing would be a better approach in my opinion, voltage swing at the o/p depends on lots of other factors, and I spose the schematic is classified.
Sounds like an interesting project.

 

[Kona311]

Said 'scope contains reed relays, these burn more power than a transistor but still not that much.
If the preceeding side of the relay contact/transistor is decoupled with a cap there shouldnt be a pop, dc is the worst offender for causing that.
I'm not a valve expert however grid leak biasing isnt great, you'll get less voltage swing, cathode biasing would be a better approach in my opinion, voltage swing at the o/p depends on lots of other factors, and I spose the schematic is classified.
Sounds like an interesting project.

Schematic is basically any high gain amp... Google Bogner Fish, Bogner XTC, Mesa Dual Rectifier, etc etc etc. It's been done so many times it's hardly worth drawing up. Cookie cutter. The only thing that changes are the R's and C's.

 

[dr pepper]

Mesa boogie make some good sounding bass amps, galleon kruger in my opinion are excellent.
What valve were you thinking of, an ecc83?