blackface PI cap question

[Randall]

Is there a reason that the 0.1 cap on the lower grid of the PI in most blackface and some silverface amps is rated at 200v when all the other signal caps are 600v? All I have on hand are 600v, is there any reason not to use it?

[The Dude]

I'm not sure which "blackface" you are talking about, so this is a bit of a guess. My bet is that you are talking about an amp where the lower grid is fed from the cathode of the PI. There is far less than 600V potential there, so a large cap is not necessary as opposed to the upper grid that is fed from the plate of the PI where voltage is much higher. And, yes, you can use a 600V cap there. It won't matter. It's just not necessary.

[Leo_Gnardo]

What The Dude says ^^^. The signal is passed from 1st triode to 2nd here thru the cathodes. The first triode "shakes" the second triode's cathode because they are connected together, and the second triode's grid is AC-shorted to ground, or nearly so, via that 200V 0.1uF cap. Since it's a cap that's doing the shorting the grid can "float" to whatever voltage it will via the triode's DC current bias, but audio frequencies pass thru the cap to a near-ground point, reducing their amplitude to a small fraction. "Near" ground but not exactly there, as it's attached to the feedback node so some small portion of the amp's output is fed to that grid via that cap, and in this way negative feedback is accomplished. So the second triode is driven from the first triode mostly by the signal delivered thru its cathode, plus the sum feedback signal arriving via the cap wired to its grid, and whatever small fraction of the signal induced thru the tube but attenuated thru the same cap wired to a near-short to ground. Tricky, clever circuit, innit?

[Randall]

"Tricky, clever circuit, innit?"

It is. And I STILL don't REALLY get it. The PI circuit on tube amps is what finger picking is to guitar playing for me. Elusive.

[Enzo]

It all revolves around the fact that as far as the tube is concerned, the bias is the voltage difference between grid and cathode. So is the signal.

Look at a regular single triode gain stage. Your signal enters at the grid, and it controls current through the tube. Current through the tube also flows through the cathode resistor, varying the voltage drop across it. In a cathode follower, that is the whole deal. But in any case, the signal at the grid is also on the cathode. The cathode moves up and down with the signal.

We tend to think in terms of ground. We think of the signal as grid to ground and cathode as cathode to ground. But really it is the grid to cathode where the signal applies.

REmember the Peavey Heritage VTX or any of the Music Man amps, where the power tubes are fed signal by their cathodes? They hold the grid steady, and move the cathode. In the grand scheme of things, it doesn't much matter if we hold the cathode steady and move the grid a volt or if we hold the grid steady and move the cathode a volt. it is the same result.

Now look at the long tail pair phase inverter. You feed signal into the top side grid. The tube amplifies that and results in the signal showing up on the cathode as well. The cathodes of the two triodes ar wired together. SO when the top side cathode goes positive, so does the one on the lower side. Your big cap to ground from the low side grid essentially grounds that grid signal wise. So that low side triode has a steady grid with its cathode moving to the signal.

Consider phase now. A positive signal into the upper grid results in a positive on the cathode too. But on the lower side, a positive on the cathode but a steady grid is the same as a steady cathode with a negative going grid. The opposite of the input signal. Thus the lower triode inverts polarity

In the above when I say positive, I am not referring to DC voltage, I am talking about the direction the signal takes.

[Justin Thomas]

That's the clearest explanation of an LTPI. I've ever heard. Thanks!

Justin

[Chuck H]

Ok... For the purposes of education (not the least of which, mine) can we discuss the value of said cap? I've read a few places that it should be valued at 10X the LTP input cap. And, of course, we've all seen that criteria ignored on many designs. Most designs keeping to the .1uf value like it was a paradigm similar to that same value being used on many presence circuits (even when that value seems less than ideal for the NFB impedance!) I'm not savvy enough to figure the impedances at every node of the LTP circuit relative to NFB, etc. So a rule of thumb like 10X is appealing, though perhaps arbitrary. I've perused the Valve Wizard and Aiken articles about the LTP design. Enough of it goes beyond my skill set that I still don't have a bead on how to value that LTP bypass cap effectively.

I assume if the value is too small that LF would suffer due to the non inverting triode grid creating local NFB. Suffer being relative since I also notice that some designs that use both LTP grids as inputs omit the bypass cap altogether. Ok, so what of too large a value? Is there a time constant consequence? And how is the bypass cap value interacting with any global NFB? It looks to me that the bypass cap value would be somewhat relative to the global NFB circuit impedance, like the presence control cap value. Yet the BF type designs (relatively low impedance NFB loop) and the Bassman/Marshall type designs (relatively high impedance NFB loop) both use the .1uf LTP bypass cap value.?. Why is .1uf "simply the value that IS used" by the entire industry like a mindless congregation genuflecting?

[mikepukmel]

. . . Consider phase now. A positive signal into the upper grid results in a positive on the cathode too. But on the lower side, a positive on the cathode but a steady grid is the same as a steady cathode with a negative going grid. The opposite of the input signal. Thus the lower triode inverts polarity . . .

Enzo, thanks I think Im really close. OK maybe not. Anyway, re the part of your comment above: positive signal on the upper half of the LTP grid, results in positive signal on the cathode, which results in a positive signal on the lower trodes plate as well and so output of the lower half is positive (at least from this signal's contribution) wrt input to the grid of the top half of the LTP. If the feedback signal is positive in this example, then it gets inverted when it goes through the lower half of the LPT (grid to plate), so comes out as negative on the lower triodes plate. So, for negative feedback, we'd want the signal from the OT to be tapped on the side of the transformer that is in phase with the signal going into the top half of the LTP. Is that correct? (I have to get a scope and SEE some of these signals!!)

[Dave H]

It looks to me that the bypass cap value would be somewhat relative to the global NFB circuit impedance, like the presence control cap value. Yet the BF type designs (relatively low impedance NFB loop) and the Bassman/Marshall type designs (relatively high impedance NFB loop) both use the .1uf LTP bypass cap value.?. Why is .1uf "simply the value that IS used" by the entire industry like a mindless congregation genuflecting?

The presence cap is working with the 5k presence pot. I think the relevant impedance for the bypass cap (and input cap) is the impedance 'looking into' the LTP grids which can be several Meg for both BF and Marshall so the bypass and input caps can be smaller. 10n would be enough. 100n is more than enough.

[Leo_Gnardo]

Why is .1uf "simply the value that IS used" by the entire industry like a mindless congregation genuflecting?

Ach! HERESY ! ! !

Just kidding...

I guess because it covers a very wide range of frequencies including down into infrasonic territory. Why does everybody? Just like 22 or 25 or 100 or 250 uF bypass caps are used in preamp cathode circuits. Agreed, we don't need to have that kind of sub bass bandwidth. So exercise your creativity, use smaller caps and shape your tone with them, I have no problem with that.

Genuflect .... now you have me thinking of an old Tom Lehrer song.

[Randall]

That's the clearest explanation of an LTPI. I've ever heard. Thanks!

Justin

Amen to that!

[Randall]

Good stuff Enzo! I'm a lot closer now, but still one thing. How does the phase relationship occur? I think I don't get why this:

"Consider phase now. A positive signal into the upper grid results in a positive on the cathode too. But on the lower side, a positive on the cathode but a steady grid is the same as a steady cathode with a negative going grid. The opposite of the input signal. Thus the lower triode inverts polarity"

[Enzo]

Assuming that last part:

Hold the grid steady, and signal raises the cathode. That means grid-cathode voltage goes down. That would be the same as the grid voltage going down on a common gain stage. So grid going down means plate voltage going up - a positive change. Remember we started with a positive going signal on the upper grid, which results in its plate going down - a negative change. SO ther we have it, a positive grid signal on the input grid results in a negative signal on upper plate and a positive signal on the lower plate.

[tubeswell]

FWIW, if you like nice big bandwidth clean sounds, you can make the amp sound cool by upping the value of the non-inverting stage decoupling cap to 220nF or even 470nF (although the higher you go the more distortion issues you raise, but on the flip side its always cool to push the design parameters). I've done this in a few on my own BF clones, personally prefer them this way. YMMV

[Randall]

Thanks for the insight Tubeswell! Always good to hear from an informed Kiwi!