PI Driver Gain Differences

[plexi50]

Here are two LTPI pics. One is a Traynor and the other a Fender. Which one has more drive (gain) and why?

Here are more detailed pics and voltages.

[HotBluePlates]

Need to see all of the schematic, including tube type, plate loads and supply voltages.

Also the Traynor schem looks curious, as there appears to be two caps running from the grids to ground. In the absence of other info, that would appear to shunt the signal, meaning the Fender circuit will actually work (and so has higher gain).

[DummyLoad]

traynor is a YSR-1; PI B+  is around 325V to PI - tube CAD says AV is about 21 - Vpk is about ~65V

http://www.lynx.bc.ca/~jc/moddedYSR1.gif - looks like a factory mod? then B+ is 338V to PI. so then Vpk would be slightly higher.

fender is AB763 SR; PI B+ is around 430V to PI - AV is around 25 - Vpk is ~100V

fender - higher B+, higher gm tube. traynor should have used a 12AT to drive 100k load?

--DL

[jojokeo]

It is the circuit with smaller resistance values on the tail

with everything being the same then you have a point but then everything is not the same in the two given schems

[plexi50]

Not trying to be difficult, but I thought you were asking which PI circuit has more gain, It is the circuit with smaller resistance values on the tail

Not at all drgonzonm. I have had the week from H*ll. Get a wod of money and BINGO the van needs calipers and a master cylinder. So i have been busy and away from the computer most of this week. I was wondering why some amps used 470K while most use 1m PI grids. And why the tail resistor varies from 6.8K to 22K and what the gain factors were. So in general do the lower value resistors produce more gain?

Ok jojokeo i will keep that in mind. I know there are many variables and i want to understand them

[sluckey]

I bet the 12AX7 beats the 12AT7.

[PRR]

> do the lower value resistors produce more gain?

No.

More max drive when you pluck harder.

[frus]

Ok jojokeo i will keep that in mind. I know there are many variables and i want to understand them

there's a pretty good calculator here:

http://www.ampbooks.com/home/amplifier-calculators/long-tailed-pair/

for the amps in question, it gives the gain around 22 for the Traynor and 25-26 for Fender (Fender is actually poorer balanced)

Traynor has 100k grid leaks on power tubes which are loading down poor 12ax7's more than "usual" 220k. 12at7's have lower plate impedance so those grid leaks are less of a factor

[printer2]

The gain of the circuit is determined (given the same tube) by the ratio of the cathode resistor and the plate resistor. On some designs the plate resistor is a different value (or split load) to make the two side's gain more balance. The tail resistor is there to force the two sides to behave, the bigger the tail resistor and the more equal the two sides are which allows you to use the same resistor value on the plates.

Now as nice as it would be to have a big tail resistor for better balance, it comes at a cost. The bigger the tail the greater voltage drop across it and less voltage for the tubes to amplify the signal with. You will see amps with EL84's with larger tail resistors that you would with 6V6's or 6L6's as those tubes need more signal than the EL84 to produce full output. The signal gobbling tubes are usually paired with a PI that has more gain (but without the greater voltage supply the extra gain makes the amp more sensitive but it still is short of signal to drive the tubes to full output) so you will see lower cathode resistors down to 470 ohms. An EL84 amp may have a 820 ohm resistor and a 56k tail as compared to a low value in a big Fender as  a 6.8k. 

Because the EL84 is more sensitive you do not want to hammer it too hard so the PI's output voltage capability is made smaller than a 6L6 amp. It keeps the distortion more pleasant. The grid leaks are also adjusted to complement the overload characteristics of the output tubes. Lower ones cause the PI to work harder and distort earlier. Different amps have different points where the output tubes start to distort and the PI starting to distort. This is one of the things that makes one amp sound different than another.

[HotBluePlates]

... The tail resistor is there to force the two sides to behave, the bigger the tail resistor and the more equal the two sides are which allows you to use the same resistor value on the plates.

Now as nice as it would be to have a big tail resistor for better balance, it comes at a cost. The bigger the tail the greater voltage drop across it and less voltage for the tubes to amplify the signal with. ...

If you divert to oscilloscope circuits instead of guitar amps, then you're likely to have a negative supply voltage available (on the order of -150vdc). The plate supply for such a circuit would likely be around 300vdc, or a total of 450v. Rather than "long-tail" the circuit type is a differential amplifier, but the bones of each are the same.

An infinite tail resistor would give perfect balance between halves, but requires infinite voltage supply. A constant current source acts as a very high impedance, and a pentode might look like a semi-constant current source due to its high internal plate resistance.

So some o'scopes use a pentode instead of a tail resistor, because they have ample supply of various voltages (usually regulated). The 450v of total voltage mentioned earlier then allows the big voltage drop needed across the pentode. Balance between the two sections becomes near-enough perfect.

You're not likely to see this in a guitar amp. Scopes have numerous supply voltages, and each serves multiple functions. Good scopes have regulation for almost all the supply voltages. They also need very accurate reproduction of the input signal. Guitar amps don't have or don't need most of this. Unless you need the most pristine clean or absolute max output from the amp, perfect balance just isn't that necessary.

[jeff]

Am I using this right? It seems that the 12AT7 in the stock values has (very slightly) more gain than the 12AX7.
And where do you enter the supply voltage?

[frus]

Am I using this right? It seems that the 12AT7 in the stock values has (very slightly) more gain than the 12AX7.
And where do you enter the supply voltage?

1) yes it has, because it has lower plate resistance and can drive those power tube grid resistors better. try a 12AX7 with 220k grid resistors (more usual value)
2) you don't, that's why I said "pretty good". supply voltage would influence the maximal signal swing, this only calculates the gain (also probably just takes plate resistance from datasheet, although PR does depend on supply voltage and bias point). but it's a good approximation

[plexi50]

So maybe i should change out those 100K grid leak resistors for 220K. Im running a 12AX7 in the PI. I tried a 12AT7 but like the 12AX7 better

[DummyLoad]

So maybe i should change out those 100K grid leak resistors for 220K. Im running a 12AX7 in the PI. I tried a 12AT7 but like the 12AX7 better

if you like the 12AX7 with the 100K GL then leave them be? perhaps what you're liking about the tone is the 12AX7 is worked harder.

--DL

[PRR]

> maybe i should change out those 100K grid leak resistors for 220K.

The spec-sheet limit for 6L6 and several other tubes, in fixed-bias, is 100K per grid.

If the grid return path (including bias-trim, trem-depth, and bias supply impedance) is more than that, a marginal-spec tube with high grid leakage will drift to higher and higher (more positive) grid voltage, higher current, until it melts.

Yes, Fender did violate the spec-sheet on many amps. Some of them are typically biased well on the cool side, so a little drift does not make trouble. Fender also knew that the leaky tubes were rare, especially in the Golden Era of good tube production. Finally, if a tube melts, you replace it... nobody dies, this isn't rocket technology.

[plexi50]

Very interesting. The PI 12AX7 working harder with 100k gl does sound good to me. I think i will leave it alone. Im at a point that i think im trying to find something wrong with the amp just to keep fiddling with it.

[labb]

c 33 (47pf) cap across the plates was added by Fender to reduce some high frequency oscillation.  First time I saw it was on the 5B3 layout ( not on the schematic)