Opto Channel Switching Question

[JoeK]

I know, Another One , sorry if this has been elsewhere in a different thread. I looked but didn't see what I was after. I have a simple two channel amp and I'm using it as a test bed for a few ideas and I'd like to have an opto setup to switch between channels. I would like to have one input jack feeding two totally separate preamps that feed back into each side of a LTP Phase Inverter. It's an 18watt power amp with no NFB and the amp is up and running fine but with separate input jacks for each channel. I've got the power supply and switching schematics but I was after some advice about where to place up to say 4 optos in this very rough schematic to achieve the best result. Also wondering about how to run the input jack into the separate triodes at the start. I know a lot of amps share the first input triode but I'm hoping to keep what I've got.

[Enzo]

I'm not the design freak, but in general I'd avoid like the plague doing a bunch of stuff to the the incoming guitar signal before it goes through a tube. The whole point of a common input stage is to boost the weak guitar signal up away from the noise floor so the amp can then deal with it through the rest of the circuits. And splitting it and running it through a bunch of optos or anything else qualifies as doing stuff in my book.

[JoeK]

Thanks for the response Enzo, I know it's a good idea to get the input signal straight up off the noise floor but wasn't sure if any amps did split the signal between two grids with dire consequence. I can only think of the patch between channels on a 5E3 or a Marshall that actually split an incoming guitar signal? I could live with running the signal for both channels through the 220K triode and then diverting with optos.

[Enzo]

Oh there are amps done that way, and dire isn;t the word I'd use. It just isn't the best practice, and that is the reason it is not often seen. No reason you can't try it that way.

[cbarrow7625]

Joe, I have an amp with multiple individual inputs. At first I just had the signal go to all the stage 1 grids in parallel and used a common 1M grid reference resistor. The tone really suffered (I have 5 different channels in this amp - my test bed amp). I solved the problem by putting a large resistor in series with each stage 1 grid. For the active channel, I shunt around the resistor; all the other channels are grounded right at the grid. The grounded grids also ground the resistors in series with them, thus providing the grid reference resistor. Since I never really truly break the guitar signal this scheme works well and is fairly quiet. I do it with relays but I don't see why optos couldn't work as well.

For your 2 channel amp, just put a 1M resistor in series with each grid. For the active channel, shunt around the 1M and ground the grid of the other channel. If you want to make sure it is extra quiet, put a 2.2M resistor in series with each grid & a permanent 2.2M resistor to ground. That way when you switch channels you never lose the ground reference (probably not an issue with optos but it is with relays)You need the equivalent of a DPDT relay to accomplish this. I don't know much about what are available as standard optos so this might not work for you. Just using a single DPDT relay is really pretty quiet; maybe you could consider that.

[J M Fahey]

From a practical, no-mojo point of view, I'd first and foremost boost the input signal a moderate amount, say 4X or 5X (slightly over 12dB) with a TL072 or NE5532 and only *then* split that clean, noiseless , louder, low impedance signal at will.
Will this "SS cr*p" kill my sound/tone/sex life?
Think again.
You´ll be showing much more respect for it than trying to have that poor weak pickup, 15 or 30 feet away, drive the combined input capacitance of a few 12AX7´s.

Yea, I know, to the Golden Question: "did Jimi Hendrix do that" I must truthfully answer: "no, He didn´t"
In fact he even used 1000 pF per foot coiled cord.
Oh well.

[imaradiostar]

In fact he even used 1000 pF per foot coiled cord.
Oh well.

Hah! So that's why my plexi is always too bright!

I like Enzo's thinking- plus it buys you an extra triode for additional gain or other fun, should you decide you need it later.

jamie

[Wombaticus]

I know, Another One , sorry if this has been elsewhere in a different thread. I looked but didn't see what I was after. I have a simple two channel amp and I'm using it as a test bed for a few ideas and I'd like to have an opto setup to switch between channels. I would like to have one input jack feeding two totally separate preamps that feed back into each side of a LTP Phase Inverter. It's an 18watt power amp with no NFB and the amp is up and running fine but with separate input jacks for each channel. I've got the power supply and switching schematics but I was after some advice about where to place up to say 4 optos in this very rough schematic to achieve the best result. Also wondering about how to run the input jack into the separate triodes at the start. I know a lot of amps share the first input triode but I'm hoping to keep what I've got.
[ATTACH=CONFIG]15706[/ATTACH]

Looks to me like you have a spare triode. If it were me I'd at least consider using that triode as a virtual earth mixer to blend/switch the signals instead of mixing at the LTP. Then you can use just two vactrols to accomplish the switching. Just use the vactrols in series with series resistance resistors for each channel at the mixer. The 5-50 Meg off resistance of a vactrol is enough to pretty much completely kill the gain in the mixer for the unused channel.

Edit: looks like I miscounted, no spare triode after all...ah well

[diagrammatiks]

why not just use 2 relays?

one to switch the input and another to switch the output.

[wizard333]

At first I just had the signal go to all the stage 1 grids in parallel and used a common 1M grid reference resistor. The tone really suffered

Can you be more descriptive on the tonal affects? I have a couple two channel amps that run exactly that way and I don't hear an issue, but I'm curious what you heard.

A lot of multi channel designs tend to share a common input stage, I think for the reasons Enzo mentioned. That is something of a design problem though, since what sounds good as the 1st stage on a clean channel isn't necessarily what sounds good on the 1st stage of a crunch or high gain channel, and vice versa. In fact, I can't think of any amp that is done on where I thought "gee, the clean and gain channels on that amp BOTH sound great". Some amps that use something similar to what the OP was talking about are the Mesa DC series, which have an LDR right up front that runs the input grids in parallel when the gain channel is on. Some others I can think of run the grids of the clean and gain channel(s) in parallel all the time and use relays or LDRs later to ground later stage input grids or select the clean or gain channel(s). On those, all gain channels share a common input stage that is separate from the clean channel input stage however; there aren't 5 parallel grids.

I haven't noticed issues there, perhaps it is because only 2 grids are in parallel vs. 5, but maybe whatever is happening, if present, is adjusted for later. In my couple of amps that do that, I didn't notice a negative tonal issue when I went to that configuration vs. shared 1st stage.

[cbarrow7625]

I haven't noticed issues there, perhaps it is because only 2 grids are in parallel vs. 5

I think that is likely the issue. I too have seen plenty of amps with 2 grids in parallel for the reasons you stated; it's the reason I chose that approach. All 5 1 st stages use different rk/ck combinations. The negative tonal affect I objected to was that it just sounded like the guitar signal was being loaded down too much. It was dull & lifeless and lacked any presence. One day I decided to experiment and disconnected the inputs to the 4 channels that were not in use and, voila, the guitar sound came back to life.

I do not tend to run any effects before my amp so any combined parallel effects of the first grids act directly on the high impedance output of the guitar pickups. I assume it is mostly input capacitance of the tubes. I suspect that if you run any effect, or non-true bypass in the off setting, that the effect pedal would provide the buffering necessary to mostly alleviate the problem.

I'm sure there are some minor tonal effects running two tubes in parallel at the first stage. It's easy enough to check next time you have one in front of you. Just disconnect the signal going to the unused channel & see if you hear a difference.

[wizard333]

Were you using a separate grid stopper on each input stage, or just one at the input like Fender does?

I wonder if there is a way to isolate that loading w/out putting in huge resistors up front. The big series resistors will lead to big resistor noise before the first gain stage, the last place you want it.

[cbarrow7625]

Actually, I do not have any grid stoppers in this amp. It is a test bench/home amp so I can get away with it. Adding grid stoppers on each input wouldn't change anything though. For the active channel, you would still shunt around the "huge" resistor (so it is out of the circuit on the active channel), leaving the grid stop resistor in series with the signal and then ground the grid of the "off" channels. If I did have grid stoppers, I would use individual ones. One common grid stopper at the front would forma voltage divide with the grid leak resistor; it is a negligible voltage loss but I just try to avoid it on principle. Why lose any more input signal voltage than absolutely necessary?


Maybe you misunderstood the application of the "huge" resistor? For the active channel that resistor is shunted out of the circuit by a relay (so it would not be contributing any noise). The large resistors in series with the other grid(s) would all be pulled to ground when the grid is grounded. The total of these other resistors in parallel should be sized so that they equal your standard target grid leak resistor value (1M, usually). Does that help clarify it at all?

[wizard333]

Yeah it's making more sense; hard to visualize from a description w/out a schem or layout. Don't you get RF like Nigel at the Air Force Base without an input grid stopper?

Still wondering if there would be a good way to isolate that up front without adding a lot of noise.

[cbarrow7625]

cb Input Switching Scheme.pdf

Take a look at the attached (terribly drawn) schematic. I hope I uploaded it correctly, never done it before. Hopefully that will help it make more sense. I used 2.2M resistors because it was a 2-channel amp drawn. If it was a 3 channel amp I would have used 3.3M, etc.

To get Nigel-like RF, it has to be present in the first place. In my workshop I have no such worries, so I get away with it. I imagine I would have issues if I ever decided to take the amp out and about. In that case I would definitely put in some RF suppressor resistors. I drew the schematic with them in place so you can see how it would work.