Testing power pentodes

[martyh]

I have a few sets of power pentodes I would like to match for transconductance and idle current at the published operating points. I don’t have enough HV regulated supplies to supply the screens separately. Can I just use some simple isolation for the screens as shown below or do I really need to build another supply?
Thanks,
Marty

[trobbins]

Do you have an amplifier that uses the pentodes in PP mode? If so, then you could use it as the test jig.

[martyh]

Originally Posted by trobbins

Do you have an amplifier that uses the pentodes in PP mode? If so, then you could use it as the test jig.

I have a few pp amps but the basing is different on a few of the types I want to check. I have test sockets already wired up to terminal strips for prototyping so it seemed like a better idea to just lash a jig up.

[dcgillespie]

If you are only doing low level transconductance and quiescent current testing, then a simple isolation network for the screen should suffice. Most of the published characteristics for typical power pentodes use identical (or nearly so) plate and screen voltages anyway, so your isolation network will need to be a low impedance design.

Dave

[trobbins]

Yes, it starts to get difficult if the range of tube types to be tested becomes too dissimilar, and the amp is not set up for modifying or as a test platform. A preferred 'test-jig' amp would be fixed bias, and some valves may need higher bias levels than supported, and need lower ht or screen levels, or higher signal drive levels.

The advantage of an amp as a test jig is that large signal performance can be compared between sets of tubes (as per fit for purpose use if this is for hi-fi).

[martyh]

Thanks for the replies
Okay, I learned the hard way. Case in point two El34s that tested close but not exact in a calibrated TV-7 at roughly 5300 micromohs next I tried them with the screens derived from a regulated 265V plate supply and they measured close again and closer to the published data at roughly 8500 micromohs (9000 published with series screen resistor). I noted that the screen voltage was slightly different from one tube to the next which considering the dropping resistor I attributed to slightly different screen currents. So I dug out and repaired another regulated supply. I retested the tubes with all of the dc voltages regulated. This time they were an exact match at exactly the published gm. (11000 with no series screen resistor)
I should also note that I quickly abandoned the idea of measuring AC current as my multimeter didn’t work well in that setup. I ended up shifting the grid ½ volt each way from the published bias point and measuring IP at both points.
Based on what I saw I would say the only way to effectively match pentodes is with all of the dc voltages regulated and even a calibrated transconductance tube tester isn’t effective unless all of the supplies are regulated (is there such a thing?) Thinking about the circuits I’ve seen using this tube in push pull pentode, the screens are supplied from a common dropping resistor which considering the variation I saw makes sense. I wonder if the series resistance in the typical UL transformer are close and low enough to not matter. Based on what I measured even a small percentage difference has an effect on balance.