Got some money, going to buy OPTs for PP GU-50

[nerdorama]

Erhard Audio sells autobias boards which would probably work very well with a toroid OPT. They are made in the Czech Republic I believe.
Auto Bias

I have one of the ABS-Q+ boards for a stereo PP amp I am getting ready to build. Not using a toroid OPT but thought I would give it a try.

John

[gerrittube]

You can buy these autobias boards directly from audioamp.eu. Mine worked quite well, but with oscillations in the audio range. I switched to a Dutch autobias alternative. So far so good!

Regards, Gerrit

[rongon]

Do these auto-bias boards enforce Class A operation? Or are they voltage regulators that stick the cathodes at a particular voltage, no matter what else happens?

Have you all tried using a TL783 as a constant current sink in the cathode of each output tube? I was thinking of that, to enforce constant current/Class A operation. I'd lose power, but that would enforce balance in the OPT core.

Yes, a TL783 voltage regulator IC has worsening performance as frequency goes up, regulation declining with increasing frequency above 5kHz or so. I don't think this is anything to worry about in this particular application.
--

[nerdorama]

Separate current regulators, one in each cathode, would require a bypass capacitor on each or you wouldn't get any audio. The auto bias board has a 10 ohm resistor for each cathode and the circuit has low pass filtering so it only responds at or near DC. Then it varies the grid bias voltage to maintain the DC bias currents separately for each tube.

You can also use a current regulator for each output channel pair which will enforce class A but maintaining balance is more difficult. I have done this and it sounds vey good.
John

[rongon]

Absolutely, yes, each current regulator would have to be bypassed. Otherwise it would be like having a >100k resistor there as far as AC signal is concerned.

(The auto-bias board) varies the grid bias voltage to maintain the DC bias currents separately for each tube.

So that is also a current regulator (cathode current is kept constant, cathode voltage is allowed to vary).
--

I'm still left wondering what OPTs I really need for a weird amp that draws 90mA of plate current per tube but is only expected to output 10W or so of (AC) audio signal power to the load.

Is the maximum plate current (Ip) the OPT can handle purely a matter of winding resistance/heat dissipation?

If a PP OPT has 100R per half primary, and there is 90mA quiescent plate (anode) current (Ip) from the tube it's loading, then that's 100R*0.09A = 9W of heat being dissipated by that primary half-winding. That seems excessive to me.

But a push-pull OPT with reasonably well balanced output tubes will have cancelling currents on its primary windings, because the two tubes will always be conducting in opposite phase. Does this mean the primary winding doesn't really 'see' the full Ip of the connected tube?

I read the chapters on power amplifiers in the RDH4 last night. It didn't shed any light on this, unfortunately. Or maybe I missed the part that addresses this. That's quite possible, being that a lot of the math in the RDH4 is well beyond my understanding. I did get that high frequency performance from the OPT is worsened by leakage inductance and inter-winding capacitance.

Questions:

Is it true that -- all other things being equal:
1) Will a larger OPT have worse leakage inductance than a smaller one?
2) Will a larger OPT have worse inter-winding capacitance than a smaller one?
3) Will a larger OPT have a worse insertion loss than a smaller one, due to iron losses in the bigger core?

I've read time and again that smaller OPTs 'sound better' (look at comments on McIntosh MC-240 vs. MC-260, or Dyna ST-35 vs. Dyna MkIII for instance). If the 50W OPTs subjectively 'sound as good' as the 25W ones, I would simply buy the bigger/higher power OPT and get on with it. But I'm hoping to make this thing 'sound good' too. I don't think I want to wrestle with a big, dopey sounding OPT that weighs 9 lbs and rolls off at 15kHz. Am I being silly?
--

[nerdorama]

All of the Hammond outputs are rated at 100mA, or more, "per side" which to me means each primary half. Some years ago the 5k 100w Hammond got good reviews for sound quality from I believe Douglas Piccard.

[tikiroo]

Originally Posted by rongon

If a PP OPT has 100R per half primary, and there is 90mA quiescent plate (anode) current (Ip) from the tube it's loading, then that's 100R*0.09A = 9W of heat being dissipated by that primary half-winding. That seems excessive to me.

--

It's not as bad as that, P=I^2R so only 0.8W per half.

[rongon]

Originally Posted by tikiroo

It's not as bad as that, P=I^2R so only 0.8W per half.

Ah, I had the math wrong. Big surprise, eh? Thanks for straightening me out there.

I found a comment on Edcor CXPP25 OPT quality from a reliable source, namely Pete Millett of 'Engineer's Amp' fame. On his Engineer's DCPP Amp page, Pete M wrote:

Note that to a large extent, the frequency response and THD vs. frequency performance are dictated by the output transformers. I used Edcor transformers here (CXPP25-MS-8K), which are affordable and measure pretty well. I also tested using some expensive "imported" iron; the THD at high frequencies was considerably better (nearly flat to above 20kHz) and the HF response extended slightly.

--

[Wavebourn]

You may checking this thread, I am getting rid of some heavy stuff while relocating.

Relocating, selling stuff.

[6A3sUMMER]

The frequency response of an output transformer depends on many things.

Here is a pair of factors:

1. The Plate resistance, rp, that drives the primary.
The low frequency inductive reactance, and the high frequency capacitance versus the rp are one factor.

2. The presence of negative feedback that samples the feedback signal from either the primary, or the secondary. That negative feedback lowers the effective rp of the output tube(s) (refer to # 1 above).

Those factors are true for both low frequencies, and for high frequencies.

I will not list the many other factors that affect the frequency response of an output transformer that is in-circuit.