plate to plate feedback - measurements

[hidnplayr]

Here is my new creation, a 807 push-pull amp, supposed to give 60W in AB2.

I added different plate to plate feedback resistors and measured the output of a 10kHz square wave, using a real speaker as the load of the amp.
I used a real speaker so I could demonstrate the effect of the dampening factor
Measurements could be repeated when I find the resistors to make a dummy load. (so i can calculate the real numbers..)

Another thing to consider is that when I use smaller feedback resistors, the voltage on the input pentodes rises. Perhaps i should have increased the current in this stage to get the same voltages on every measurement.

Here you will see the output of the 10k square, with no plate to plate feedback.
Clearly, the amp doesnt have a good dampening factor. (altough I find it sounds rather nice already)

[hidnplayr]

Now, with plate to plate feedback.
From left to right: 56k, 82k, 150k

[hidnplayr]

180k, 200k, 220k

[hidnplayr]

Can someone explain to me what happens when I use a too small feedback resistor?

[Michael Koster]

Originally Posted by hidnplayr

Can someone explain to me what happens when I use a too small feedback resistor?

Plate voltage goes > 250V? maybe 270V?

What's happening is as the voltage drop across the resistors becomes less and less (as you decrease the resistance) there is very little remaning available driver headroom (voltage and current) to swing positive. Remember that the 807 plate voltage is decreasing as the driver voltage is increasing. It's running out of drive current.

Try coupling your feedback with capacitors to block the DC current from the 807 plate. If that doesn't do it you may need more than 5-7 mA idle current in the driver instead.

Or maybe the performance is great with the larger Rfb and you're golden!

Cheers,

Michael

[DF96]

With a small feedback resistor you are simply putting a square wave through a limited bandwidth, so it gets rounded off. This is what you should see at 10kHz. With a higher resistor you are seeing OPT ringing at its HF resonance, as the anode impedance is too high to damp the resonance. This looks superficially better, but is actually less accurate.

[richwalters]

I've tried this method on many power amps without real benefit. From a distortion view, each and everycase using good quality o/p transformers it made the 10 kHz sine distortion slightly worse, although it cleaned up the square wave ringing with the effect of reducing the already placed UL snubber values. In such cases, the application of global nfb (20dB) has such profound effects on the whole performance of the amplifiers that I dispensed with the configuration. I can only assume that the technique works best with lower quality output transformers, but one will find by altering the quiescent output stage current at 10kHz working this also has a profound effect on THD. i.e ye higher the quiescent current, the better looking squarewave and lower sine THD. Goes hand in hand. By tackling the 1st stage Zobel and the global feedback step response in a proper way I found I gives optimum results..........i.e nom 20dB circuit global nfb, the amplifier if designed properly should cope with another 15dB on top without any signs of oscillation or instability....into a resistive dummy load. It can be a tough test if the layout is poor. Alot of output stages, in my view run too light at higher audio frequencies which has a big effect on sound quality definition.

richy

[Allen Wright]

I have experimented with anode/anode fb many times, but no matter the improvements in measured performance, the amps always sound "thin".

No measurable bass roll off, but sound thin.

Anyone else experienced this?

Regards, Allen

[DF96]

Change in damping factor affecting bass resonance?

[artosalo]

Another thing to consider is that when I use smaller feedback resistors, the voltage on the input pentodes rises.

This is obvious since DC-voltage at anode of 807 is essentially higher than at the anode of the driver. The smaller the feedback resistor, the bigger current thru it will flow and results increasing anode voltage at the anode of the driver. Simply add a series capacitor.

Can someone explain to me what happens when I use a too small feedback resistor?

With small feedback resistors the higest frequencies seem to attenuate.
This can be seen as rounded edges of the square wave.

I would suggest you to experiment with some amount of global feedback (...10 dB) from the secondary of the output transformer AND also this shunt feedback, but only a little (3dB...). You could find a great combination.

By the way: You can get up to 70 W from an AB1-class PP-pair of 807.
Makes the whole construction much easier...