Beginner needs advices for 807 PP project

[oshifis]

I am rebuilding a 807 PP amplifier that I built some years ago, this time more focused on audio quality. I don't want to modify the PSU, so the following voltages are fixed:

425 V plate voltage
-50 V bias voltage

I would like to use the 807s triode connected, run in AB1 and drive them with a tube phase inverter, not interstage transformer. I don't want overall NFB, and 5 to 10 W @ 3% would be sufficient. Since I have not much experience in various circuit arrangements, I have some questions:

- Which is better, fixed bias or autobias? In the latter case do I need to bypass the cathode resistor, or is the local NFB an advantage? What is the max. value of the grid resistor in either case?

- Can I connect g2 to the plate? The data sheet dictates Ug2 < 300V.

- Can I run the tubes with 60 mA idle current without degrading their lifetime?

- Which phase splitter sounds best? Initially I used two 6DJ8s SRPP connected, with a CSS in the lower cathodes, a kind of long-tail pair configuration. Now I am considering a simple LTP with anode resistors and CSS, followed by a DC coupled cathode follower.

- I have a 25:1 output transformer. For optimum match to 8 ohms load do I need to go up to 27.5:1 or even 30:1?

- Is cross-feedback of the center-tapped OPT secondaries into the cathodes á la Quad II of any advantage (linearizing effect)? I tried it but the AC voltage on the secondary is too low compared to the grid AC voltage. Perhaps a 27.5 : 5 : 1 transformer (the 27.5 and 5 coils center-tapped) is better, the 5 coil going to the cathodes in opposite phase?

Any advices regarding the sonically best options welcome.

Laszlo

[Tweeker]

807s arent going to give very much power in triode A1 or AB1 at all, and 425V is pushing them, but may be ok. Datasheet gives triode connected limit as 400V. G2 connects to plate via a 100 ohm resistor for triode. If PSU is full wave center tap, you can drop voltage by connecting a zener diode to CT.

Grid circuit resistance limit is 100k ohms for fixed bias, 500k for cathode bias.

60ma idle @425V is too high for class AB1, is fine for class A.

With a 5K ohm OPT, if cathode biased the resistor should be bypassed.

[oshifis]

Tweeker,

Thanks for your input, this reassured that I am not on the wrong track. What about my other questions?

Laszlo

[diegot]

oshifis,
Using the 807 triode connected in class AB2 will give you much more power. Try it.

[Tweeker]

Cathode feedback can be good, the Quad II implemented using a very fancy OPT with a bifilar tertiary winding though. Feedback requires more gain. PP cathode feedback using the secondary requires said secondary to have a center tap and be symmetrical for optimal performance, many (if not most) OPTs are not

Id suggest fixed bias AB2 with cathode follower drive, idled at ~40ma. Maybe 6SL7 or 5751 LTP for more gain than 6DJ8s, should be just fine as driver with CF buffer. If you want class A, you probably should have a higher ratio OPT wound.

Consider tetrode connection using gas tube regulated screens with cathode feedback and/or global feedback. 807s perform very well this way, though its not required for 10 watts.

[gerryc]

I'm sure there are many others but this one comes to mind. Classic Williamson circuit with triode-connected 807s.

http://www.bonavolta.ch/hobby/en/audio/807.htm

[oshifis]

Thanks for all your comments. Yesterday I modified the circuit like this:

Input: E88CC long-tail pair with 56 kohms in the anodes, 8.2 mA CCS (1N5312 + 1N5313 current regulator diodes parallel) in the common cathode to -50 V.

Buffer: E88CC cathode follower DC coupled to LTP, 56 kohm in the cathodes to GND.

Output: 2 x 807 (russian!) with 560 ohms autobias shunted with 2 x 10 uF 160 V WIMA MKS-4. OPT cross-coupled to bottom of cathode resistors. Grid resistors 47 k to GND. G2 resistors 47 ohm to anode.

100 ohm grid stoppers at every tubes' grids. No overall NFB.

I measured the amplifier with a Tektronix SG505 oscillator + AA501 distortion analyzer. It performed 4 V output @ 0.135 % THD with 1 V input, this is 2 Watts @ 8 ohms. At 2 V input (8 Watts output) the THD was 0.63%. The distortion goes up rapidly beyond this level, it reaches 3 % at 10 V output (12.5 Watts). Frequency response is 15 Hz to 70 kHz.

I am very satisfied with the result, and I leave it like this for now. Listening tests will come...

[Bandersnatch]

hey-Hey!!!,
I have found good results with 807's running 40% U-L on a 10k a-a load. Input stage built with two video pentodes. 12HL7, 6CL6, EL84, SV83, 12BY7...
cheers,
Douglas