Calculate desired OT primary - how?

[EFK]

When looking through data sheets for a 6L6GC, I see there are a range of OT impedances listed depending upon bias type (cathode, fixed) and various plate/screen voltages. Seems to be a pretty wide range, I see anywhere from 3.8K up through 5.6K. If one was spec'ing a OT, is there some kind of formula one would use based upon various voltages etc. in a particular amp to determine what would be the best impedance to extract maximum power? Curious about how these impedances are determined. Anyone?

[MWJB]

Impedances that I am aware of are from 2.5K up to 7K, brutal truth is that some budget manufacturers used what was available cheap. Fender started with higher impedances (7K-ish) with some of the tweeds, dropping to 4K, or just under into the tolex amps. There is no direct correlation to plate voltage.

Load lines can be determined, but these rarely tie in with what popular manufacturers used.

[Wilder Amplification]

The only way to determine what OT primary impedance will yield maximum power transfer at a given plate/screen voltage is to draw a load line. A load line allows you to determine what load will intersect the Vg1 = 0 line just below the knee of the curve at a given B+/screen voltage for maximum voltage/current swing.

[Steve Conner]

Yes, or for a single-ended amp you can subtract 50 from your B+, then subtract your cathode voltage, and finally divide by your idle current.

(You're assuming the stage is biased for symmetrical clipping, and the "knee" is at a tube drop of about 50V, then "drawing" a load line.)

[MWJB]

OK John, out of interst, I want 485vdc on the plates, 483vdc on the screens, 57W output from a pair of 6L6, what's the OT primary I need?

[Wilder Amplification]

OK John, out of interst, I want 485vdc on the plates, 483vdc on the screens, 57W output from a pair of 6L6, what's the OT primary I need?

Assuming the plate voltage can drop down to about 50 volts at peak swing I'd say you'd need a 4.5K plate-plate impedance for that output. Of course this assumes a 100% regulated supply, which doesn't exist in a guitar amp. Accounting for the B+ drop under load I'd say a 4K plate-plate impedance would be a more suitable load for this output power.

[MWJB]

Thanks Jon. That's what I've got.

[jpfamps]

The best way I've found to work out the impedance for maximum power is to actually measure the output power into various loads using a large power resistor.

This will thus accommodate any sag etc in the power supply.

One aspect I've found that does have a significant effect on the loading is sag in the screen supply, either as a result of using a resistor in place of a choke, or by use of larger screen grid resistors.

Much of the data provided in the data sheets specifies a regulated screen supply.

[EFK]

That's what I just did - using the 4.2K bassman transformer I currently have mounted, I shifted the reflected impedance up and down. Moving it down raised the AC voltage on the tube grids slightly but decreased AC output at the speaker jack, and moving it up decreased both measurements, so I'm going to leave it right where it is which seems to make max power output. The only unfortunate effect here is that doing it into a dummy load doesn't permit you to actually hear it, and for all I know, shifting it one way or another may reduce power slightly but create a more pleasing tonal effect.

[pdf64]

'I shifted the reflected impedance up and down. Moving it down raised the AC voltage on the tube grids slightly but decreased AC output at the speaker jack'

If the load resistance on the secondary is reduced, the Vac across it might reduce but the actual power going to the load could increase.

'moving it up decreased both measurements'

If the secondary load is increased in value, surely the Vac across it increases also?

[MWJB]

EFK wrote: "and for all I know, shifting it one way or another may reduce power slightly but create a more pleasing tonal effect." Well, that's the key, power within reason may be "enough" but you're not going to settle for a less than ideal sound because you get a few more W. Trouble being that it's very hard to compare like with like, unless you have the same speakers in more than one ohm rating (I used to have a healthy collection of Eminence 102 in both 8 & 16 ohms that were useful for AB testing loads). It's always worth just trying a few variations, you may well end up back at 4.2K...then again maybe not.

Peter, the VAC may increase due to the larger load, but power might not necessarily change with it, with the higher loads the speakers see less current than with the lower loads, so it can be swings & roundabouts.

[octal]

...on the 'trying different load impedances' topic- I happened to notice that Weber has a speaker matching autoformer. I don't know that I'd want to use one permanently, but it might be a good way of experimenting with the sonic effects of different reflected loads. It's got some taps (5.3 and 2.6 ohms) which would allow for tweak mismatches beyond the typical factor of two differences based on using mismatched taps on the main OT.

Nathan

[MWJB]

Indeed, but you can also use three speakers in an array, in parallel to try the "in between" impedances. This can also give you a workable compromise when switching between 6V6 & 6L6 for example.

[octal]

Naturally. I was just suggesting that a $30 X-mer might be an alternate, less expensive way of trying different impedances than a three pack of boo-teek speakers.