Matchless Hotbox Layout Corrections, Need More Gain

[PRR]

> 100k increase to 120k results in increase output from 20V to 24V.
> Now, I assume this is a simplistic example

Exactly true IF the input is a 0.1mA signal.

We almost never control current directly. We wiggle grid voltage, which influences current. But in triodes, plate voltage also influences current. And in the opposite direction.

The increase of gain is always less than the increase of plate resistor.

A large plate resistor with the same supply voltage and reasonable plate voltage implies a lower current. At lower current, a change of grid voltage causes a smaller change of current. You are starving and weakening the tube.

Reasoning from First Principles is good for the brain, but not a good way to design.

As a general rule, the tube plate resistance should be much smaller than the load impedance. Even if there is no other load, the input capacitance of the next tube is around 100pFd, which is 100K impedance at 17KHz. For guitar-amp design, you should generally expect the load to be near 250K. Unless it is less: fix-bias power tubes should have 100K grid resistors.

The plate resistor should split the difference between tube resistance and load impedance. "Geometric mean" is better, but it isn't critical. But if plate resistor is larger than load impedance, gain and swing will be poor.

We have a choice of Mu and, at reasonable currents and loads, tube resistance is proportional to Mu. You want a high Mu for high gain in high impedance loads, you favor a lower Mu for good swing in lower impedance loads.

Since next-grid and stray capaciticance does not vary a whole lot, and we lean to the highest Mu we can get away with, we can write these general guides:

Rp= 470K, 5KHz response
Rp= 220K, 10KHz response
Rp= 100K, 20KHz response

Several 470K stages may have more gain at 100Hz and 1KHz, but less gain at 10KHz, than several 100K stages. Use 470K in AM radios, in mellow amps, and in very-simple amps. Use 100K in many-stage amps with extended treble response. The rules are not exact (and added capacitance will reduce response more), but give you a guide.

[skyboltone]

SB - Do you know if the commonly reprinted RCA book (RC-30, 1973) has R-C tables in the back for the EL-34/6CA7 and 6L6? Earlier RCA versions available on-line have data sheets for those, but not R-C tables (do for the 12AX7).

That's the one I have and it has no curves for the 6CA7. It has an abbreviated set for the 6L6GC. There are a few pentodes in the RC tables but usually only those that also have a triode inside and serve as stand alone audio sections of radio and TV receivers, or drive a pair of push/pull tubes from a line level input. Like the 7199.

If you want to squint the curves go here: http://frank.pocnet.net/index.html

Dan H

[PRR]

> R-C tables ..... for the EL-34/6CA7 and 6L6?

Ummmmm... "R-C Tables" are for Voltage Amplifier work. Generally taken to mean high voltage gain and reasonable audio voltage swing into "high" impedances.

That's not what the Big Bottles were made for! They were made for POWER. Voltage gain is a minor detail. R-C coupling would be stupid, because the best power efficiency of an R-C amp is 4%, while for an inductance-coupled amp it is 50% in theory and nearly that good in practice.