1W Power Amp with 6922

Hi,

Very simple design for this 0.5/1W power amp. It is connected to 8-ohm small speakers, directly fed by a line source (about -10dBV) such as portable cd players, dat, mini-disk, iPod, etc.

Input xfm is a step-up and phase-splitter to directly drive the triodes. The tube is a 6922 (one for each channel).
Each triode: I have chosen a 5K loadline, quiescent point 100V-10.5mA, grid bias -2V.

Are my calculations correct? Am I missing something? Will I get1W of power to drive my speakers?

Any suggestion welcome.



Val

The 600Ω input winding may not be happy fed with a 10K pot. At -6dB, the wiper impedance is about 2,500Ω. Some "600Ω" windings will be fine, others will lose bass. Many "line level" outputs will drive an unloaded 600Ω transformer, but some really are not happy with less than 10K or 2K.

> quiescent point 100V-10.5mA, grid bias -2V (and 155V B+)

Remember that the "10K" of the transformer is for audio frequencies; at DC it will be much less. Ideally zero, in practice about 1/10th of the rated AC impdedance. I penciled 500Ω per side. The plate voltage will be nearly B+. To stay near a 10mA idle current, the grid bias has to be raised to almost 4V. So the plate-cathode voltage is more like 147V than 100V.

The common cathode resistor should be more like 200Ω, and I don't think it needs to be anywhere near as large as 4W. Heck, the whole amp only draws 3.3 Watts, and the cathode resistor a very small part of that.

Ignoring the fact that ECC88 is only rated 130V: yes, it does look like about 1 Watt at clipping at 4Ω loading. I suspect distortion will be fairly high (several percent), damping factor about 2, tube life will be more than an hour but not a decade because of the overvoltage. At 8Ω loading it may be more like 0.7 Watts, a little lower THD, a little better damping, same tube life. Reducing the B+ to stay inside the plate rating will cit power in half, roughly.

MANY years ago when TV tuner tubes were cheap, I did a similar thing with 250V-rated tubes running at 330V. I got about 50 hours of use before the sound degraded. When tubes were $0.50 this was not excessively expensive. With twice the voltage on a slightly higher load impedance with lesser tubes like 6BQ7 it was about 0.3W-0.5W depending on the tube and how close to clipping I considered "maximum". FWIW, that was with a 12AX7 driver/splitter.

The grid swing with 140V-150V on the plate will be about 3.7V peak per grid, 7.4V across the total secondary, 7.4/6.5= 1.1V peak or 0.8V RMS at the primary.

I think it is a killer low-Z headphone driver, and may be a very fun speaker driver with the right speakers and listener.

[quote author="PRR"]The 600? input winding may not be happy fed with a 10K pot. At -6dB, the wiper impedance is about 2,500?. Some "600?" windings will be fine, others will lose bass. Many "line level" outputs will drive an unloaded 600? transformer, but some really are not happy with less than 10K or 2K. [/quote]

Betterwith a 2K transformer and a 100K pot?

> quiescent point 100V-10.5mA, grid bias -2V (and 155V B+)

Remember that the "10K" of the transformer is for audio frequencies; at DC it will be much less. Ideally zero, in practice about 1/10th of the rated AC impdedance. I penciled 500? per side. The plate voltage will be nearly B+. To stay near a 10mA idle current, the grid bias has to be raised to almost 4V. So the plate-cathode voltage is more like 147V than 100V.

The common cathode resistor should be more like 200?, and I don't think it needs to be anywhere near as large as 4W. Heck, the whole amp only draws 3.3 Watts, and the cathode resistor a very small part of that.

Ignoring the fact that ECC88 is only rated 130V: yes, it does look like about 1 Watt at clipping at 4? loading. I suspect distortion will be fairly high (several percent), damping factor about 2, tube life will be more than an hour but not a decade because of the overvoltage. At 8? loading it may be more like 0.7 Watts, a little lower THD, a little better damping, same tube life. Reducing the B+ to stay inside the plate rating will cit power in half, roughly.

Click to expand...

Will re-design for a 6N1P, which is similar, still easy to find, and 300V rated.

The grid swing with 140V-150V on the plate will be about 3.7V peak per grid, 7.4V across the total secondary, 7.4/6.5= 1.1V peak or 0.8V RMS at the primary.

I think it is a killer low-Z headphone driver, and may be a very fun speaker driver with the right speakers and listener.

Click to expand...

:roll:

Why not use a (stereo) pot AFTER the transformer?

Hi,

Mods are very few:
Re-designed for a 6N1P tube, B+ is now 180 volts, output transformer primary is 25K, so static loadline should be 1,250 ohms (right?), quiescent point is 170 volts at 6.5mA, grid bias still -2V, so common cathode resistor is now 150 ohms.

Will I still get about 1W of power?
Predictions about distortion/damping?

[quote author="gyraf"]Why not use a (stereo) pot AFTER the transformer?[/quote]

Input trafo mod.: 2K5 prim., so I think 10KB pot. will be fine now.

Any suggestion/comment welcome.

Peace,
Val

The driver transformer can't be both 2K5:25K and 1:6.5. 2K5:25K is 1:3.13 turns ratio, or a 1:6.5 ratio with 2K5 input is 105K secondary.

We could calculate our butts off, or just find available iron and build it. I think it will be a happy amp, but not big power (obviously). The input sensitivity should be OK with home hi-fi, a little insensitive to get the most out of 3V-power devices like portables (however, they could stand a 100Ω:10K 1:10 transformer, if you can find one).

[quote author="PRR"]The input sensitivity should be OK with home hi-fi, a little insensitive to get the most out of 3V-power devices like portables (however, they could stand a 100?:10K 1:10 transformer, if you can find one).[/quote]

This should be a reasonable compromise: Input transformer has prim. 2K5, and sec. 160K, turns ratio is now 1:8.

Other than that, any suggestion/comment more than welcome.

Peace,

Val

I still wonder if the input pot is gonna affect freq response. Much better to do like gyraf sez and put a stereo pot on the hi-z side.

[quote author="Larrchild"]I still wonder if the input pot is gonna affect freq response. Much better to do like gyraf sez and put a stereo pot on the hi-z side.[/quote]

Unsure if I will leave the pot, input transformer is 1:8 5K:320K.

Respect,

Val

:wink: