Welcome! Need support, you got it. Or share you ideas and experiences.
Return to MagneQuest/Peerless Forum
141.151.17.78
In Reply to: RE: Parafeed PP transformer? posted by twystd on May 21, 2009 at 10:55:31
Hi Roe:
I have written extensively on this subject--- the import of primary impedance and it's effects on loadlines and, hence, the behaviour of
the amplifier. If you do a search across the different audioasylum forums with my name and inductance--- you'll have plenty of material to read and absorb.
The article written by Voltsec (published on our magnequest.com site) is quite comprehensive--- and uses the 845 tube and circuit application as a the working example. What is shown is that the tube generated distortion itself goes down dramatically as the primary inductance increases. This is because the tube is seeing and working into a larger EFFECTIVE load impedance.
In essence what it required is that to have a stable (relatively non-reactive load for your tube to work into) you must have enough primary inductance to adequately support your nominal reflected primary impedance.
Years ago (and perhaps yet to some extent today) folks would take the nominal or advertised pri impedance of a transformer as it's effective impedance. Which can lead to a lot of mischief--- and misleading specsmanship.
How much inductance--- there is no hard and fast rule--- unless we were living in the Platonic world of forms--- in which case, of course, we would have transformers with infinite inductance and no resistances of any kind--- and no core losses of any kind----
but, here on earth--- we must juggle how much L we can aim at or achieve against a host of other design factors and etc. As a very rough rule of thumb--- the most excellent designs extant (that I know of) are ones that can produce approx 100 henries of L per thousand ohms of reflected pri impedance while keeping the leakage L at or below 1 millihenry per thousand ohms of the nominal reflected pri impedance.
The above level of performance is near rave.... the Peerless 20-20 Plus series will generally meet this spec... and there are perhaps a few other tranneys made worldwide to this high level of specification. But these units are tough to design and quite intricate to wind. They will have a ton of interleaving--- the insulation materials must be carefully chosen--- the lead positions, starts, and stops of the windings must be laid out near perfectly--- it is really the state of the art that goes into these guys.
Is this level absolutely required to enjoy music? Nope. It's an IDEAL. Quite, quite nice if you can acheive it design wise--- and even better if the customer can afford to buy and use it.
And, again--- my "rule of thumb" simplifies the shit out of the design process---- and how you must really optimize for a very wide range of factors in design--- stuff in addition to the amount of L that the primary has--- but the primary L is none-the-less one of the most critical design parameters.
hope this helped---
MSL
Builder of MagneQuest™ & Peerless™ transformers since 1989
Michael- Thanks for taking the time to explain this, and I realize there are trade offs and no one spec makes or breaks a transformer. The intricacies of all of this is obviously over my head. Let me just ask you to advise me on this.
I think you've seen the circuit I'm trying to build, It's something I want to try, but if it doesn't work out, I'll try a more conventional design. Whatever I end up with will be a PP class A amp, running either 2A3-40s or 300BXL tubes. The dual center taps are required.
If you were building a beast like this which would be your choice, in the $400.00 (possibly a little more) a transformer range. I could go slightly more, but would have a tough time affording the price of the Peerless 20-20 plus.
twystd
What aspect of the circuit demands a split primary?
I was thinking of a parafeed PP with CCSs on the plates. The center taps have caps that return to the cathodes.
twystd
You know there is no particular reason to put caps to the cathode in a shunt feed PP. If you had a center taped primary you could put a cap between the two sections to insure no unbalanced DC in the windings. But it would probably sound even better to balance the current some other way and eliminate the caps all together. Then you do not need a center tap at all.
Google Marcel and Clovis by Phil Sieg. He has built an amp along these lines.
Michael
Yes Marcel and Clovis is a really interesting design. The design I'm thinking of uses the garter circuit to balance the tubes. I can't have a shared cathode resistor for the garter circuit to work, so it's a PP stage rather than a differential stage.
PP amps usually require a cathode bypass cap for full power. By putting the cap returning to the cathodes, I don't think I'd need a cathode bypass cap, plus they block any DC on the transformer. The thought is to have a really balanced PP stage, with minimal capacitors, and taking advantage of no DC, unbalanced or not, on the transformer.
No telling what it will sound like, it's just an experiment I want to explore. I haven't seen a schematic like it before, and neither have my local DIY friends, so thought it might be interesting. If it shows promise, I will release the schematic to the DIY community. If it sucks oh well, I'll just drop back and punt to something more proven.
twystd
Post a Followup: