A neat trick?

[Salas]

Has anybody ever implemented a power pentode current source to cancel bias current in the opt of an SET like in here? If yes how is the performance? If no sonic penalty is involved, makes big SET building significantly more accessible.

[ErikdeBest]

I know that Menno van der Veen has some project online using this trick. Using the same driver, power tubes, PS and OT he compares different topologies.
http://www.mennovanderveen.nl/eng/project-frameset.html

the last project are with an EL34 set for current compensation in the output trafo. Not hard to do, it seems that almost any PP amp can be easily modified to accomplish it.
http://www.mennovanderveen.nl/eng/project25.html
http://www.mennovanderveen.nl/eng/project26.html

One day, I will try it too

Erik

[Salas]

Interesting, cuts costs, handy. Its a pity he did not take the full optimization path in his example so to tell us more. Its a bit demanding on supply ripple for start. Good concept idea and easy to check out since it's applicable on widely available, normal costing PP transformers and valves. I guess I will check it out next time I will be breadboarding something PP. Anybody here did it already? Some findings?

[Dave Cigna]

I haven't done it myself, but Jeffrey Jackson (Experience Music) has used the filament current of the output tube to cancel the 'idle' flux in an SE transformer. It takes relatively few turns of wire because the filament current is so high compared to the plate current.

It worked very well in his 50W 304TL amp and I believe he has been trying it in smaller amps. He's turned a few other people on to it too.

His filament supplies generally use a choke as the last element before the filament itself, which presents a high enough impedance with the large turns ratio. Of course a CCS could also be used... many people use a CCS to supply their filaments anyway.

-- Dave

[Wavebourn]

I had the Idea, but gave it up because of power loss that is already big enough in SE amp.

[tubelab.com]

There was a thread here a while back (maybe a year?) exploring this very subject. I reported on my experiments using a CCS IC to load the other side of a P-P transformer in SE use. It works, but I did not try it out with a high quality transformer so I can't compare it to a good SE transformer. I don't remember the details, but I belive some of the other writers may have tried pentode loading. I have a schematic (somewhere) of a comercial amplifier that uses a 211 in SE mode, and another (undriven, but equally biased) 211 on the other side of a P-P transformer. It would seem that this would be a good place for a current mirror, or some type of bias servo circuit to guarantee equal currents.

[Salas]

I would think it takes a very high rp of the source tube in order to avoid eating away efficiency. Did you experience hum? How was the sound with that modest pp you used? Had true SET character, even if compromised in bad iron?

[tubelab.com]

I did these experiments about 2 years ago, so It is hard to remember the sound. I have a bunch of P-P OPT's that were made for guitar amps. I have used them successfully in some amps, while they don't sound great in other applications. I am using them in my 300B push pull amp. It can be made into an instant SE amp by pulling one driver tube from each channel. I have tried this, but it doesn't sound as good as my 300B SE amp which uses custom Electra Print transformers, which is no surprise.

Most of my amp experiments (including this one) are done using a regulated bench power supply so hum is not an issue. The effective resistance of a 10M45 CCS IC is over 100K ohms so it will not cost much efficiency. A triode (the second 300B or 211) will eat up as much as half of your power, but it will match the driven tube exactly. A pentode will have a much higher Rp, costing less efficiency. The CCS IC has the highest "Rp" but some say that they don't sound as good as a tube.

I found out that these same transformers worked reasonably well as the OPT in a 45 SE amp just wired up as an SE transformer and run with 30 mA of current through it. Parafeed works pretty good too.