calculating plate load for non-traditional power tubes

[pvsage]

I'm looking at quite a few NOS tubes with the idea of using a preamp tube (actually part of a dual triode/pentode tube) as a low-wattage power tube. Is there a way to determine an appropriate OT primary impedance if none is specified in the data sheet? One tube I'm particularly interested in is the 6AN8, but I'm looking at other dual-preamp bottles as well.

Haven't been here in a while - I usually post at SEWatt, but Wattkins is down right now.

[R.G.]

In general, there is an impedance that produces the maximum power, and a different load that produces the minimum distortion, and these are not the same. This is one reason hifi OTs and guitar OTs tend to have different nominal impedances.

The simplest way to do this is to put any OT on it and vary a power resistor on the secondary. When you reach a power maximum on the resistor, calculate the plate loading that exists from that resistor. That's your power maximum plate load. Then go find a tranformer/load combination that makes this magic condition come true.

Remember - OTs do not have impedances. They have RATIOs.

[gmoon]

There's always the formula for output impedance of a tube:

Va = Plate (anode) voltage
Pa = Maximum plate (anode) dissipation

Zout = Va/(Pa/Va)

[pvsage]

gmoon: Are you sure about that formula? I just tried that with typical specs for an EL84, and I got about 13K for Zout - considerably higher than the 5.1K normally specified for this tube!

R. G.: So you're saying the only way to find the right plate load is experimentation? Also, I apologize for the ambiguous shorthand - I meant "primary impedance relative to speaker impedance" or something to that effect. I'm aware that the actual resistance of the windings is negligible.

So I measure the power on the resistor...how? Do I measure the current across the resistor and calculate power from that?

[R.G.]

R. G.: So you're saying the only way to find the right plate load is experimentation?

No, it's not the only way. But it is probably the fastest way.

I'm sure that some time spent poring over plate characteristics and plotting load lines, then calculating impedances would turn something up. But if you have a working power supply, OT and tube socket, messing about with resistors is quicker and more direct.

Also, I apologize for the ambiguous shorthand - I meant "primary impedance relative to speaker impedance" or something to that effect. I'm aware that the actual resistance of the windings is negligible.

OK; I just didn't know the level of your understanding. We get a lot of people who think a 6.6K to 8 ohm transformer is always 6.6K.

So I measure the power on the resistor...how? Do I measure the current across the resistor and calculate power from that?

In a resistor,

P = voltage squared divided by R; so measure the AC rms voltage, square it, and divide by the resistance.

and
P= current squared times R; so measure the AC rms current, square it and multiply by R.

The voltage measurement is easiest with a DMM set to AC volts.

If you measure peak volts with an oscilloscope, you have to convert to RMS by dividing peak by 1.414.

[pvsage]

I don't have the tube yet; I'd rather do the math/plotting to make sure I have a tube that'll work with a relatively "standard" 125_SE OT. The triode/pentode tubes I'm looking at have different pin-outs, so it seems to make sense (to me, anyway) to know what I have will work before I wire the sockets.

What I'm considering is a single tube variant of the Jones Octal One, using the pentode as a gain stage and the triode as the power stage. (A natural extension of this would be based on the Octal Fatness, using a standard 12AX7 for the cathode follower.)

I'll re-read the Valve Wizard's page on biasing power tubes and see if that gets me any closer to finding the right OT load...

[pvsage]

[

gmoon: Are you sure about that formula? I just tried that with typical specs for an EL84, and I got about 13K for Zout - considerably higher than the 5.1K normally specified for this tube!

OOPS!!! I misread your text - I was going with the actual dissipation at the given plate voltage, not maximum dissipation. The formula works fine. Now I have something I can work with!

[Old Tele man]

...rule-of-thumb: the OT push-pull Zoo value is typically somewhere between 1/6 and 1/10 of the rp-value of the output tubes being used; with 1/6 approximating minimum distortion and 1/10 approximating maximum power output.

[pvsage]

Hmmm...interesting. I *think* that agrees with the pentodes I've looked at. But I'm mostly looking at using a triode section as an output tube. Mostly. In the pentode/triode tubes I've looked at so far, the triodes have much lower Rp's, typically arount 4K.

Based on Zout=Va/(Pa/Va), I've found quite a few viable candidates if I wire the OT for a 10K primary...but again, this appears to be a pentode approximation.

[gmoon]

Based on Zout=Va/(Pa/Va), I've found quite a few viable candidates if I wire the OT for a 10K primary...but again, this appears to be a pentode approximation.

Gotta wonder if using a reverb transformer as an OT would work...certainly opens up a different set of options (~20K : 8)

[pvsage]

Thanks for the suggestion! But if I stick with the 6AN8-A and a Hammond 269AX PT, that gives me a "sweet spot" of 10K for the OT with 160V at the triode plate and just under one watt output - perfect for a practice amp in a small space. I'll only be able to use two secondary taps at 4 ohm and 8 ohm, but since I don't have a 16 ohm speaker, that's really not an issue.

Refresh my memory - why don't we typically use triodes in guitar amps? Is it because they produce primarily even harmonics? I'm sure it'll sound really warm if I can get the pentode to overdrive it!

Gotta go plot some dissipation curves...these tubes aren't typically used in output stages, so the data sheets don't include that little crutch...

[pvsage]

OK, I think I have a handle on this. Could someone check my math?

Based on the data sheet at NJ7P (http://www.nj7p.org/Tube4.php?tube=6AN8A), 160V at the triode plate and a max dissipation of 2.8W gives me about 10K for an optimal plate load.

If I bias the tube for about -3V at the grid, that gives me a dissipation of 2.56W. This is also very far from center bias - very warm. Won't it distort far below 2W? (I'm perfectly fine with this so long as it's a "good" distortion.)

[Old Tele man]

...with TRIODES, the "rule-of-thumb" is: OT Zoo is about 2-to-4 times the triode rp-value.

[gmoon]

OK, I think I have a handle on this. Could someone check my math?

Based on the data sheet at NJ7P (http://www.nj7p.org/Tube4.php?tube=6AN8A), 160V at the triode plate and a max dissipation of 2.8W gives me about 10K for an optimal plate load.

If I bias the tube for about -3V at the grid, that gives me a dissipation of 2.56W. This is also very far from center bias - very warm. Won't it distort far below 2W? (I'm perfectly fine with this so long as it's a "good" distortion.)

I think you've pretty close. For a 10K load resistance @ 160V, -3 to -4 volts bias is about as close to the "linear" as you're gonna get. It's a pretty steep load line so it's not gonna be linear, anyway (2nd order distortion?)

Will it sound good? I dunno, probably depends a lot on the output level of the preamp.

You could flatten that loadline a bit (14K load @ 200V), maybe connect multiple speakers to get 14K or 15K primary from your existing transformer.

(Are you avoiding the pentode of the 6AN8?)

[pvsage]

I think some 2nd order distortion would be fine! So long as we avoid blocking distortion. A typical 1K5-4K7 grid stopper should help prevent that.

I already have the 269AX and 125CSE and a small number of 8 ohm speakers on hand, hence my choice of 160V and 10K as operating points. The only really unusual thing I might consider for the power supply and output is maybe an old Jensen field coil speaker.

I'm looking at the pentode as a high-gain preamp, as I'd indicated earlier in this thread. If you consider the Octal One as a jumping-off point, using a pentode/triode tube in this manner really isn't too far outside the box.
http://www.freewebs.com/jonesamps/JonesOctalOne.pdf

I'll post a copy of my schematic once I determine the right values for using the pentode in the 6AN8 as a preamp.