Optimal location in circuit to implement channel switching?

[Zippy]

I'm contemplating the build of a two channel amp - 1) EF86 and 2) 12AX7 - similar to a DC30 but with a two-knob tone stack on the EF86 rather than just switching the coupling cap. I'm thinking that switching later in the circuit would add less noise than switching at the inputs. If you have the opportunity to do a scratch build, what is the optimal place to switch between or combine the channels? Just before the phase inverter?

Thanks,

g

[Tdale]

I might be mistaken, but how would this work?

I assume you want just one input, but two channels. If you switch between the preamps right before the PI, how would that work? If the two preamps are in series, like the Dumble, it's no problem, but when you have two separate preamps, the signal would have to go through both of them at the same time if you want to switch right before the PI...

Or am I missing something?

Tommy

[David Root]

Assuming parallel channels, if you use two separate SPST switches you can dump either signal to ground quietly or have both on together, pretty much wherever you like. You have to passively isolate the two channels immediately before the summing point, eg with 220K/470K resistors, to eliminate back signals when both are on together. See any Fender BF reverb circuit. Or you can use two pots instead of switches, which is a bit more flexible. Or you could use a tweed (anode) mixer as the last preamp stage for both channels, which very effectively sums and isolates the signals.
If you're adding the stages in series, (which I doubt is the intent here) the Dumble arrangement is proven and straightforward.

[Zippy]

Thanks, David. Yes, I am talking parallel channels. I've done the job with the passive summing resistors and separate volume controls before. It is versatile but tends to beg having both channels on and mixed rather than switching between them. I thought that if I could implement switching at would allow A/B/A+B, that it would make the amp more versatile still.

Since the upcoming build is from scratch, I thought I'd open myself up to thinking outside my ol' box.

[Bob-I]

I did this on the 3 channel build. I tried the switch first at the input, then at the output of the channels. At the input not only did I have more noise, but the channels interacted, my guess is that the output impedance was changed by the 2 output stages being connected.

I then paralleled the inputs and switched using a DPDT relay, grounding the unused channel. The result is far better. Quieter and almost no switching noise using a 44M resistor across the contacts.

The only issue I have relates to the overdrive section. I have the second channel connected to a switchable OD section ala Dumble. If the OD is engaged and I switch to channel #1, some of the OD bleeds through. Checking carefully for this bleed through I believe that it's through power supply rails.

[David Root]

Following your logic, then you need to find out exactly where it's getting into the DC (could be more than a single source, depending on your circuit), or where it is feeding from DC into the channel it's showing up on, and then set up a bleed circuit or cap to ground that removes this AC signal.

In the unlikely scenario that it is coming from two out of phase origins, eg a plate and cathode, you could theoretically null the two at a summing node, although this might be difficult or impossible in practice.

[Bob-I]

Following your logic, then you need to find out exactly where it's getting into the DC (could be more than a single source, depending on your circuit), or where it is feeding from DC into the channel it's showing up on, and then set up a bleed circuit or cap to ground that removes this AC signal.

I believe it's getting into the DC right from the plate of V3, the OD tube.

In the unlikely scenario that it is coming from two out of phase origins, eg a plate and cathode, you could theoretically null the two at a summing node, although this might be difficult or impossible in practice.

Actually my scope disagrees. There appears to be AC coming from both OD stages, only it's stronger on the second. There appears to be some cancelation but the wave still seems to get through.

The solution was simple, when I switch to channel 1, turn the OD off. It was a simple change in the relay ckt.

[mlp-mx6]

The only issue I have relates to the overdrive section. I have the second channel connected to a switchable OD section ala Dumble. If the OD is engaged and I switch to channel #1, some of the OD bleeds through. Checking carefully for this bleed through I believe that it's through power supply rails.

Bob, I did a 3-channel build too, though my own "design" and layout. I have the EXACT same issue - If I'm on the clean-only channel, and the clean/OD switchable channel is on OD, I get some noise in there. If I turn off the "inactive" OD, the dedicated clean channel sounds correct.

I tried everything I could think of, to no avail. For now I just live with it.