Effect of output transformer primary impedance?

[fuzz_addict]

Hi folks, first post here. Long time reader. If I've posted this in the wrong place please let me know.

Just as a foreword, I have a background in electronics and have somewhat recently gotten into designing/building valve amps.
I'm feeling fairly confident in the knowledge I've gained so far, but there is one concept that I just can't wrap my head around - the relationship between output transformer primary impedance and current draw in a push pull class AB amp.

My understanding is that a lower primary impedance will raise the current draw, but then the conflict here is could one not just bias the tubes at an appropriate level and negate the effect of a lower primary impedance? Doesn't seem right, would love if someone could explain or point me in the right direction.

A real life example would be using a 6k6 primary impedance OT instead of say 8K for a pair of 6v6's. Comparable to swapping in a deluxe reverb output transfomer into a princeton reverb. Wouldn't this lead to a higher current demand and overwork the power transformer? This is assuming a fairly "typical" PT like a 330-0-330 100mA affair.

Thanks in advance, would dearly appreciate some guidance for a newcomer.

[Stevem]

There's a lot more to it then that and it would take a ton of typing to post !

I would suggest you go to the Aiken amps page, then go to the tech listing , then go to white papers and then skim down to the output transformer listing.

[martin manning]

My understanding is that a lower primary impedance will raise the current draw, but then the conflict here is could one not just bias the tubes at an appropriate level and negate the effect of a lower primary impedance? Doesn't seem right, would love if someone could explain or point me in the right direction.

This is a question about load lines. I suggest reading here: http://www.valvewizard.co.uk/pp.html
The short answer is, yes lower primary impedance will increase the current draw and plate dissipation when signal is applied. The bias point is an idle (zero signal) condition, which changes the average plate dissipation by raising or lowering the dissipation on the Class A potion of the load line. It does not affect the high-power Class B portion.

A real life example would be using a 6k6 primary impedance OT instead of say 8K for a pair of 6v6's. Comparable to swapping in a deluxe reverb output transfomer into a princeton reverb. Wouldn't this lead to a higher current demand and overwork the power transformer? This is assuming a fairly "typical" PT like a 330-0-330 100mA affair.Thanks in advance, would dearly appreciate some guidance for a newcomer.

There you have a case where two 2x 6V6 amps are running different load lines, and the Deluxe Reverb is definitely demanding more current from the power transformer and working the output tubes harder (for the same plate voltage).

[R.G.]

This was a big issue back during the Golden Age of tubes. In the 50s and early 60s, it was common to see charts of combined output power, second harmonic and third harmonic distortion versus the plate to plate impedance. The idea was to pick a plate loading impedance that gave you the best combination of output power and low distortion. The general result was that the curves showed a gently sloped hill of power and two different gentle valleys for the distortions.
These charts presented the tradeoffs clearly: the peak power, minimum second harmonic, and minimum third harmonic happened at different loading points. Taking the 6L6 as an example (because that's the one I remember most clearly) the lowest combination of distortion happened at about 6000-6600 ohms plate to plate, the power maximum was about 4000-44000. This accounts for most music/hifi amps using 6.6K and most guitar amps using the 4k loading. The guitarists wanted LOUD.
It is a matter of load lines, all right. A tube can only deliver the voltages and currents that its plate characteristic shows. These are foggier than the collector curves for a bipolar or drain curves of a MOSFET, but they're just as binding. There's a maximum current that the tube can deliver at 0V grid to plate (and the fogginess of positive grid drive can deliver a little more) and a maximum voltage that the tube can withstand at grid cutoff. The plate loading amounts to a down-and-right slope within this range of possible voltages and currents. Picking a plate to plate impedance amounts to picking a slope of this loading curve. Picking a bias current amounts to picking a starting point for the curve. The maximum power hyperbola (that is, the locus of the points where plate to cathode voltage and plate current equals the tube's maximum dissipation) puts limits on where you can have the load line go as well. Your pick of the slope of the load line and where it starts and ends by picking bias and plate voltage determines the output power and distortion numbers.

There was a semi-infinite amount of textbook and article information on this kind of thing in the Golden Age.