Coupling caps, mix resistors single channel

[EFK]

I'm building a little single channel/1 input amp like this: input --> 1/2 triode --> t/b tone stack --> volume --> 1/2 triode --> PI --> power. What I'm wondering is:

(1) since there are no channels to isolate, do I need a "mixer" or grid stopper going into the PI? I just stuck a 220K (Fender value) there now, but does it really need to be there? What are the effects of removing it altogether, or raising it to 'Marshall' values i.e 270K or 470K?

(2) is the 'coupling cap' off the second stage plate essentially the PI input cap as well? It seems kind of silly to set up, say, a .022 coupling cap into a .01 PI input cap - wouldn't one just use the .01 cap right off the 2nd stage triode plate and into the PI (or through the mixer/grid stopper into the PI)?

I'm sure these are remedial questions, but any info much appreciated!

[d95err]

No, you don't need mixing resistors, although sometimes a grid stopper can be useful. Experiment! Coupling caps are to prevent DC getting from one stage to the other, so you only need one.

I suggest taking a look at some schematics for amps similar to the one you're building, e.g. the Blackface Fender Princeton (AA964).

[Chuck H]

(1) since there are no channels to isolate, do I need a "mixer" or grid stopper going into the PI? I just stuck a 220K (Fender value) there now, but does it really need to be there? What are the effects of removing it altogether, or raising it to 'Marshall' values i.e 270K or 470K?

It doesn't need to be there. There will be a small increase in gain and you could get high end oscillations since a single resistor in series has a small low pass effect. You can reduce gain in other ways, a split plate load is my favorite, and use a smaller resistor right at the input tube grid to help squash oscillations. When you don't need mixing resistors the split plate load / small grid stopper arrangement sounds better to me.

If you just want to stick with a simulated mixing resistor then, obviously, the higher resistor value the less gain amd the greater low pass effect. But you have to get really big resistance to make the low pass audible mostly it affects ultra sonic stuff. But something to pay attention to is also whats happening on the end of the "other" mixing resistor that isn't in your circuit. For example, say the other mix resistor in the amp you copied the single channel from goes through the mixing resistor a .022 cap and then straight to the plate load with very little resistance in between. That's 220k-.022uf-100k-0 impedance. You would have a mild high/mid filter effect that would be missing from a single channel because the second channel circuitry isn't there.

(2) is the 'coupling cap' off the second stage plate essentially the PI input cap as well? It seems kind of silly to set up, say, a .022 coupling cap into a .01 PI input cap - wouldn't one just use the .01 cap right off the 2nd stage triode plate and into the PI (or through the mixer/grid stopper into the PI)?

In a circuit like you describe you only need the .01 cap. You do need the cap right before the PI in the signal path with no ground loads in between. If your circuit allows you to do that with one cap then one is all you need.

Chuck

[Bruce / Mission Amps]

Chuck's reply seemed worded a little odd to me in the first part.
Correct me if I am wrong but I think Chuck meant that not using the series resistor "could" result in some kind of high freq oscillation.
The amount of gain loss is really really low though and I think they sound slightly better with a 100K to 220K resistor before the DC blocking cap... also, it doesn't hurt to throw a 220K to 470K resistor to ground before the cap too.

[EFK]

Thanks everyone - this has cleared it up for me to some degree, although I had no clue what a 'split plate load' was until I started reading up on it a little. That's quite an interesting concept. I was wondering about the ability of the resistor to prevent oscillation and it seems like it ought to be a concern. Bruce, by adding a resistor to ground as you mention, wouldn't I basically be creating a voltage divider? Can you elaborate on why you think this would be a good idea?

[tubeswell]

by adding a resistor to ground as you mention, wouldn't I basically be creating a voltage divider? Can you elaborate on why you think this would be a good idea?

Depends on how much gain you need tame once you have it up and running. I've discovered, in my amp experiments, that its better to build the board with extra eyelets etc so I can add a few parts easily here and there once its all assembled.

[EFK]

"...its better to build the board with extra eyelets etc so I can add a few parts easily here and there..."

Me too! Always comes in handy.

So you're saying the purpose of the added resistor to ground/divider - if utilized - is to *reduce* gain? I don't think I'll need that, given that the tone stack and volume between triode stages (only two, both plate driven) are probably going to suck up a lot of signal?

[tubeswell]

Well of course, whether you want to reduce gain or not depends on all the other bits of the circuit. But if you are experimenting, then why not either make that divider variable with a 250k-1M pot. Then you can muck around with the gain that best suits yourself. Either way you will need some form of grid leak.

[EFK]

I'm having a hard time understanding why you say I NEED a grid leak off the grid resistor (I jumped down to a 150K there now, not really a mixer as there's only one channel) as the amp seems to be working just fine as-is. ???

[tubeswell]

I'm having a hard time understanding why you say I NEED a grid leak off the grid resistor (I jumped down to a 150K there now, not really a mixer as there's only one channel) as the amp seems to be working just fine as-is. ???

Are you saying you already have a 150k grid leak? (We might be talking at cross-purposes here). Are you able to post a schematic?