Connecting series pass valves in parallel usually requires measures to equalise the currents and that may complicate the biasing in this topology.
I don't think you are off thread at all exeric!
I don't think you are off thread at all exeric!
I guestimate about 7watts, but that maybe a bit on the high side.
My speakers are probably at leat 97db/Watt so this amp is probably capable of burning them.
Shoog
My speakers are probably at leat 97db/Watt so this amp is probably capable of burning them.
Shoog
From my feeble understanding the different transfer function of pentodes over triodes in driving current into a fairly low resistance load is just perfect. There is really only one reason to be prejudiced against pentodes in this application and that is partition noise. To eliminate that you would have to go to a high impedance output triode pair, such as a cascode topology. However you would then end up defeating the streamlined design Shoog has come up with.🙁
Others may feel differently,but I find it hard to see how you could do with less than about 20W or so for 83db speakers. If you are keen to keep Shoog's elegant 4 valve design then you might be able to use 3.2k anode to anode and swing about 140V RMS from a 250V operating point on a 6528. 6S33S would be relatively easy but you would then have 4 very big output valves.
I'm thinking it might not be that difficult. If one continues to use that historical thread with Brian Beck the circuit for one push-pull had constant current sources on each cathode linked together with capacitors between them at the cathodes. One's first impulse is to try to somehow connect any parallel tube's cathodes to what is already there. I could be wrong, but I don't think you have to do that. In fact at the cathodes the push-pull pairs should be completely disconnected
Instead, just create independent push-pull pairs and connect the only common point at the plates where all the different tube pairs combine current through the OPT. It seems to me this would work. Technically you wouldn't even have to have the same currents going through the different push-pull pairs. In fact you could have one pair biased to pure class A to eliminate crossover distorion through the OPT and the other pair biased lean into class AB. You still wouldn't have to mess with bias adjustments as the tubes age just as in Shoog's original motivation for constant dc current sources in the cathodes. (Which I absolutely love.)
I'm glad you don't consider this off topic. Someone should just give me a poke if it happens. I can tend to do that.
You want to build about 12 constant current sinks for the output valves! With the ones I use that would be 36 transistors,LOL! Actually, I am perfectly capable of doing something that crazy myself. I can't think why it wouldn't work but I am working at the moment so I can't think carefully about it now. You are going to need a mother of a power supply. Great!! Ha,Ha,Ha....
Yeah, I agree. 7 watts doesn't quite make it with my preferred speakers. And the changes needed to accomodate it sort of defeats the whole purpose of Shoog's elegant design. I'm rethinking trying to do this because I prefer smaller power tubes like the 6as7 to the monsters you mentioned that are used in power supplies. I'm probably going back to my original plan of using the differential aikido design driving "conventional" amplifying tubes with cathode feedback. If its not enough power I can either parallel them or wire one end of the capacitors in the cathode directly to ground for AB1 bias. Its going to be complex anyway so why screw up a good design like Shoog's!
Actually I've learned a lot from this thread that will be useful later on. Mostly it is about the different uses and topologies that involve constant current sources in the cathodes of output tubes. I'm definitely going to be using these idea in the future.
For now though I've decided that I want to get my feet wet with something that puts out close to 20 watts and is fairly simple. I'm going to do that with Pete Millets "Red" amp using sweep tubes. It even has, coincidentally, the complement of 4 output tubes and 4 input tubes. I take that as a karma sign. Also should be cheap experience and good sound. I can take the experience from that to go to infinity and beyond! he, he.
For now though I've decided that I want to get my feet wet with something that puts out close to 20 watts and is fairly simple. I'm going to do that with Pete Millets "Red" amp using sweep tubes. It even has, coincidentally, the complement of 4 output tubes and 4 input tubes. I take that as a karma sign. Also should be cheap experience and good sound. I can take the experience from that to go to infinity and beyond! he, he.
You can adjust the design of any push-pull stage to a power differential stage. It is only if you want to use these mains toroids that you probably need to use low impedance tubes. You talk of sweep tubes; I am sure that EL509 would work very well in a power differential stage. Or what about Wavebourne's favourite GU-50? Cheap,easily available and very nice triode curves,I am sure that would give you the power you need.Remember,with CCS you can swing more volts so you should get a bit more power than with regular push-pull. I have learned a lot from this thread too,particularly with DF96 filling in details of things I had guessed at and tried and him going to the trouble of running those simulations. The real experts on this forum are very generous with their help! Good luck and get building!
My last post is misleading(late at night!!) It is ,of course, in the driver stage that, assuming same differential topology, you get a bigger swing with less distortion. The AC loadline of the output stage depends on the transformer impedance in just the same way as always. However the better driver conditions may enable more power at lower distortion.
Remember that differential operation means that you are restricted to class A. There is absolutely no way for this topology to stray into class B of any kind.
The Tabor clone is capable of about 30Watts of Class A pentode power. There are things I would do to refine that design and feel that it could deliver better performance than I got out of it. Gain is not an issue with that design.
Shoog
The Tabor clone is capable of about 30Watts of Class A pentode power. There are things I would do to refine that design and feel that it could deliver better performance than I got out of it. Gain is not an issue with that design.
Shoog
I understand that with differential CCS it is only class A. So no one has to always return to a previous thread to know what we're referencing here's the map of different CCS cathode schemes courtesy of Brian Beck. Pic A,B,C,D, and E is class A only. Pic G, and H is the "standard" CCS scheme that allows limited AB. Not sure about F. Is this correct?
The tabor amp has always been one that interested me but I could never find a lot of documentation beside the schematic itself. Class A is good. I also realize Pete Millets amp is not class A and is fixed bias. However it seems to be a good way to introduce myself to partial feedback/anode follower/schade feedback (take your pick).
The tabor amp has always been one that interested me but I could never find a lot of documentation beside the schematic itself. Class A is good. I also realize Pete Millets amp is not class A and is fixed bias. However it seems to be a good way to introduce myself to partial feedback/anode follower/schade feedback (take your pick).
Attachments
Last edited:
Hello Shoog, I have been thinking quite a bit about your bias arrangement and can see many advantages; one of these, it seems to me, is the possibilty of running a preamp input stage on dual CCS's. Normally I use high gm valves for input which are very hard to match and drift very quickly with the result that common mode rejection is far from maximised. I think that your biasing scheme would solve this straightaway.I also took a little time to read some of the older threads on these topics(I only joined this forum 2 months ago) and was thinking about the toroidals;there is no getting away from the fact that low frequency power is limited by the limitation of 115V AC on each primary at 50Hz. Either one has to use transformers with higher rated primaries-do they exist?-or do what you have done to really get the output impedance really low. What about differential cathode followers with CCS's in the anodes? 50 ohm rk might be achievable with various single tubes,indeed,cathode to heater voltage specs allowing, there might be quite a good choice of possibilities. I may try this rather than just a higher power version of my last build(which,after all,I am very satisfied with). Can you forsee any particular problems other than fairly tough driver requirements?
Exeric; sorry there has been no reply to your last question. It is hard for me to see with the large idle currents in the designs we have been discussing how,other than in extremis,how this kind of stage could be driven into class B and I don't know what the behaviour would be under such circumstances.
Exeric; sorry there has been no reply to your last question. It is hard for me to see with the large idle currents in the designs we have been discussing how,other than in extremis,how this kind of stage could be driven into class B and I don't know what the behaviour would be under such circumstances.
If you want reliable high common-mode rejection then you really need an input transformer. Valves will always drift whatever you do. Fixing the current does not necessarily fix the gain.
Yes, I did wonder if my reasoning was correct about this and I am aware of the formula mu1*mu2/(mu1-mu2) which I often use to estimate worst case CMR. Is there no benefit in the dual arrangement in this situation then? Would you then recommend an input step-up transformer for a MM cartridge? I cannot really use MC as I need to change tips for 78's, early Russian LP's etc.
Last edited:
Actually my reasoning centred on cathode emission and how good CMR is achieved in shared cathode valves-I was not focusing on the different amplification factors.Is that totally wrongheaded?
Not a step-up transformer, just an isolation transformer.
I have not really been following the later parts of this thread so I am probably missing something. Where does your formula come from?
I have not really been following the later parts of this thread so I am probably missing something. Where does your formula come from?
OK. I'm not certain why MJ uses mu, as I would have thought Ra was more appropriate. I need to think about that. One thing to remember is that for finite Rk the CMRR depends on whether you take differential or single-ended output from the LTP. With differential it doesn't matter if the anodes move, as long as they both move the same. I think MJ assumes differential output, but he doesn't seem to say.
- Status
- Not open for further replies.