max. output for two 5881 tubes?

[tonskulus]

We had nice testing session here.
Amplifier:
-2x5881, pushpull. Hammond 1650H 6.6k/4-8-16ohms output transformer rated 40W hifi, 70-80Wmax.
-phase splitter: ECC83, cathode connection.
Anode voltage: 330V. Screengrid voltage: 330V
Grid bias: adjustable 0...-70V
Class AB
Load: adjustable highpower 100W resistor, 0...20ohm
test equipments: signal generator, oscilloscope, multimeter.
We used both 100 and 1000Hz frequencies and measured voltage output. Oscilloscope was used to indicate wave shape at output.
Max. voltage for 6ohm load was 12volts:
Max. output power was 24watts, 8ohm taps connected to 6ohm load. Any possible load configurations were tested, no any better results. Bias voltage somewhere beetween 20..30volts.

First run gave only 10watts output, using 260V g2 voltage (there was 7kohm resistor).

So do we really need more plate voltage than 330V to get more than 24watts output power?

[SY]

So do we really need more plate voltage than 330V to get more than 24watts output power?

Yes.

The 5881 is nearly identical to the 6L6 series. Use those operating conditions as a guide. In pentode AB2, you should be able to squeeze out 50 watts. Power diminishes rapidly with most other connections. I haven't checked the datasheet, but offhand, 24 watts sounds about right for that B+.

edit: The datasheet shows that for pentode, fixed bias, 6600 ohm plate to plate, 360V B+, you should get 26.5W. So your numbers are certainly sensible.

[tonskulus]

So now I need some good advices to get that 50W output. Any ideas?

[SY]

With a 6k transformer and 480V on the plates, 6L6GC should get you the 50W you seek, at least at midband. From your description of the output transformer, it sounds marginal.

Keep in mind, 50W does not sound "twice as loud" as 25W.

[Dave Cigna]

In my own experiments with 5881/6L6 tubes (years ago) I was able to get 45W with 450V on the plates and screens. IIRC that was with a load in the neighborhood of 3.5k using an ECC83 LTP driver. I found that biasing the tubes cold to get closer to class B did not help increase the power, but I might have been limited by the driver in this case.

- Drivers capable of swinging more volts *might* give more power with colder bias.

- Lower screen voltage (~400V or so) might buy a few extra watts as this will allow the plates to swing a few extra volts before running into the knee of the curve.

- If you want to run AB2 (drive the grids positive) then you will need to use a lower load to avoid hitting the knee in the plate curves. 3.5k might be OK depending on how positive you drive the grids. And, you will definitely need a beefier driver; ECC83 won't do it even as a cathode follower.

Incidentally, I know the 5881 has a max plate/screen rating of 400V, but I also know they can easily take 450 as long as you do not exceed their dissipation rating.

-- Dave

[anatech]

Hi,
Correct me if I'm wrong, but 5881's are rated at 25 W for plate dissipation. A 6L6GC is rated at 30 W.

-Chris

[Dave Cigna]

I forgot about that. I also forgot that I never let it bother me as the 5881's I played with seemed able to handle any 6L6GC application I tried. To be fair, the 5881's that I tried were not real 5881's, but Russian (Sovtek branded) 'equivalents', so who knows...

-- Dave

[anatech]

Hi Dave,
There have been some Fenders out there that happily ate some 5881's. I just wanted to clear the air and stop the myth. Especially with the NOS nonsense floating around.

-Chris