Mixing dry/wet signals

[antonis]

2N4125 has a minimum h_FE of 25 so a 5k6 Collector resistor should be much more sensible..

For higher h_FE (within reason), NFB will settle things up..

Anyway, if you insist on BJT mixer, I'd propose someting like below:



P.S.
C4 is oversized due to next effect unknown input impedance..

[thunderaxe]

thank you for this!
but i certainly don't insist on using a BJT, i just want to keep it as close to the original phase 90 circuit as possible, but more importantly understand why they did it the way they did -- why did they choose a PNP BJT rather than an op amp to do that job? it certainly seems like another op amp stage would be the obvious choice there, which raises the question for me of why MXR didn't?

[m4268588]

Do you want to analyze the MXR's circuit?
Just in case, the mixing pot you're trying to do can't be realized with this circuit.

• Vref = Q1_base + Vbe

Sounds right?

       Vc-(Vref-Vbe)   Vbe   Vbe
Vc = ( ------------- + --- + --- ) x hFE x R4
           R7          R16    R8

R7 = R8 = R16, so...

• Vc = (Vref-3*Vbe)*hFE*R4/(hFE*R4-R7)

[PRR]

> why did they choose a PNP BJT rather than an op amp to do that job?

Is a mystery to me too. Just being clever? Or using-up excess parts? Commercial decisions do not happen in a vacuum. Not even a logical vacuum. If Mike Matthews tells you the 2SC1234 is THE high-mojo transistor to use, you say "Yes sir!" (you idiot). No, Mike doesn't do that, but other bosses get wild ideas.

Remember that until very recently, a transistor was cheaper than an opamp. Especially since if one type-number goes out of stock, there's surely another type you can use. (Opamps aim for interchangeability but don't quite get there.)

[jorg777]

I'd trade active for passive parts in this situation.  Assuming non-correlated signals (more or less true for phaser or chorus) it would provide minimal loss and constant loudness.

[thunderaxe]

I'd trade active for passive parts in this situation.  Assuming non-correlated signals (more or less true for phaser or chorus) it would provide minimal loss and constant loudness.

but as mark pointed out, this configuration won't let you fully mix out a signal for true 100% dry and 100% wet at the pot's far ends

[Matthew Sanford]

If you want it 100% (ish) wet or dry, couldn't you just take a smallish resistor (say 5k on that 50k pot) from each side to ground, that way if the pot is zero resistance to wet then no issue for dry but wet gets a 50k/5k divider, so then 91% or 9%? Maybe up the pot to 100k and use 1ks to ground? Or...not good enough?