Switching an output gain stage with an SPDT... this won't work, will it? (schem)

[aion]

Trying to think if there's any way to switch an output gain stage using a SPDT switch. Here's the schematic:



Input is coming from F2, and then it either goes through the gain stage (F3) or takes a shortcut to bypass it (F1). Trouble is, I have my doubts that the FET circuit will stay out of the way when bypassed since the transistor is still biased up and doing its thing regardless of whether the signal is going through it. Is there any way to accomplish this with a SPDT switch, or would I need to use a DPDT to fully bypass the section?

[Digital Larry]

I'd do a few things.

Put the coupling caps on the output between the FET drain and top of the volume pot.  Generally you do not want a DC voltage across an audio pot.
The bottom of the pot can go the GND.  Not sure what "VR" is here but ground is just as good.

Then you could put the pole (common) of the SPDT to the top of the volume pot.
One side of the SPDT then goes to the caps.
Other side of the SPDT could connect to the FET gate.
Might want to add a 1M from the first side of the SPDT to ground.

This would send both pre and post gain through the volume pot.  Maybe that's not what you wanted?

Also Rx5 at 1M hardly has any effect in parallel with a 12k (assume the cap is low impedance at audio frequencies).  What are you trying to do there?  If the point is to increase the gain at audio frequencies then Rx5 needs to be around the same value as RX4 or smaller.

[aion]

Shoot, RX5 should indeed be 12k - I think I just duplicated RX2 in Eagle and forgot to change the value. Good catch.

And thanks for the tip on the order of the caps and volume pot, I'll get those switched around.

VR is Vref (1/2V) - this is a mod to the end of a Bluesbreaker and nearly all the variants I looked at had the volume draining to Vref instead of Gnd. However, I did actually include a second pad near lug 1 to connect the pot to ground instead of Vref just in case.

Is there any rule of thumb on when the volume should be connected to ground vs. 1/2V? The Tubescreamer volume pot also drains to Vref, but I've noticed some variants (like the Eternity) go to ground instead.

[Digital Larry]

I'll give you my theoretical perspective which may differ from those with more practical experience  

The ground is supposedly the voltage "reference".   That is to say, anything meaningful about the signal is generated and/or measured relative to that reference.  In the case of op-amp circuits powered by a single battery, these usually split that voltage to an internal Vref because op-amps want to have a split supply based on their design.

If you're combining single-ended and split-supply sections in a circuit, ideally they would all reference the same "Vref".  But sometimes they don't.  What's the risk?  It depends on the quality of your power supply and the impedance of the voltage reference and other currents in the circuit.

Sometimes Vref is developed by putting a couple 10k or 4.7k resistors in series from 0 to +9V of a battery.  If you return 1 mA to that point, it will cause that reference voltage to shift!  If that current is actually an audio signal, all of a sudden your so-called reference is bouncing around at audio frequencies (= bad).  [Edit: this could be reduced somewhat by a filtering cap across one of the resistors.] Other times you'll see an op-amp voltage follower hanging off the resistor divider which creates a MUCH lower impedance for the rest of the circuitry which is not so bad.

If you use a battery for power, you don't have to worry about ripple.  But if you use a wall adapter, even with regulation and filtering it's going to have some (hopefully small) ripple.  If you use +V for some references (e.g. positive ground PNP fuzz circuits) and +V/2 for some others (op-amp EQ stage) and 0V for yet others (for whatever reason) then it might work out OK IF your power supplies are very clean and very low impedance.  Otherwise it is best to use the same signal as the reference point for everything in the circuit.  The risk is that the ripple will get amplified because all the circuit "knows" is that such and such a voltage appeared between these two points.

I'm working on a circuit that uses + ground for PNP fuzz as well as +5 for some digital stuff.  I went ahead and added circuitry to generate -9V for the fuzz so that everything could be referenced to the same point.

[Digital Larry]

dual post