Yeah, I have a couple of those down in the cellar.
I think I´ll try another driver stage, i tried SRPP in my 6B4G amp and I disliked it alot. Thats a shame, I think the SRPP stage i a really clever circuit so I was terribly disappointed with the sound it produced together with my 6B4Gs.
Well, that was a completely different amp so maybe I should give the SRPP another chance. After all, the output stage in this particular schematic is SRPP-coupled so why not the same for the driver stage
I have to say that I´m still thinking about making my own choke plates, i already have the cores and bobbins so all I need is some wire and a truckload of patience. I´ve made a couple of OPTs before and that was a hell of a job, but I suppose chokes are easier to wind.
We´ll see where this ends...
Cheers!
I think I´ll try another driver stage, i tried SRPP in my 6B4G amp and I disliked it alot. Thats a shame, I think the SRPP stage i a really clever circuit so I was terribly disappointed with the sound it produced together with my 6B4Gs.
Well, that was a completely different amp so maybe I should give the SRPP another chance. After all, the output stage in this particular schematic is SRPP-coupled so why not the same for the driver stage
I have to say that I´m still thinking about making my own choke plates, i already have the cores and bobbins so all I need is some wire and a truckload of patience. I´ve made a couple of OPTs before and that was a hell of a job, but I suppose chokes are easier to wind.
We´ll see where this ends...
Cheers!
Nuh. sorry, won't work.Fuling said:
If you have the windings phased so the flux cancels the signal cancels as well. What will work though is get a toroid with 4 (!) primaries and connect it as you say for one tube, and the other 2 windings to a second tube next to it but with the winding phase reversed WRT the first tube. Then you will have good dc balance. Drive the grids anti-phase.You can also connect a reasonable size cap from one anode to the opposite side cathode and vice versa to make the coupling really tight. Should work really well.
That's right. The bottom of the top winding gets pulled down; the top of the bottom winding gets pulled up. The *start* of both windings would *both* have to be pointing up or pointing downward. If the tube starts to increase conduction then to make an output signal, the top winding flux and the bottom winding flux both have to increase together *in the same direction*. If they oppose each other in dc (good) then they will oppose each other in ac signal as well (bad), so no output. If they aid each other in dc they will aid each other in ac signal too. Dem's the sad troof.Fuling said:
Maybe we´re not talking about the same topology, coz I once talked to a guy who told me he had managed to build an SE amp with a PP transformer. He had split up the centertap of the primary to get two separate primary windings. One winding was connected between B+ and plate, the other between cathode and ground. One output tube only, and a driver stage capable of quite some swing since the power tube will act partly as a cathode follower. Maybe I´ve misunderstood the guy completely but this is how I remember his explanation.
Cheers!
Cheers!
My copy of MJ is the third edition, which up until today I thought was the latest edition. I haven't seen the fourth as I believe it just came out. But the reasoning looks solid. The DC component is out of phase and therefore canceling while the ac component is in phase and therefore additive.
There isn't any reasoning, just a statement that because the windings are antiphase the quiescent current cancels. Says nothing about what happens to the signal current. If there was reasoning about it it would go something like this:
Look at the arrows on the diagram Fig. 6.11 on page 392 (3rd edition). The arrows go the same direction for changing signal current as they do for quiescent current. If quiescent current cancels so must signal current. It's not at all like push pull where the signal currents are driven "antiphase".
There's nothing to "resolve" it's obviously a mistake or at least a serious misunderstanding.
"Sadly the drive requirements are almost as severe as for the cathode follower, and this technique has not been widely used. See Fig. 6.11"
LOL I thought everyone comes up with this at some point until they realize :facepalm: I remember when I did...
Thanks Michael Koster and Smoking Amp, I needed a face palm moment! I've never built an SE power amp, only PP. Just goes to show that what they say is true: believe nothing of what you hear, 10% of what you read, and only 50% of what you see with your own eyes! Good thing I tried to type a rebuttal. All it took was a closer look. Thanks again!
Ron
Ron
Yep, that scheme catches almost everyone. We all like to think optimistically I think.
There is a way to fix it up though. Either the external inductor scheme I mentioned above (not so hot).
Or search on "anti-triode". A number of threads on that, doubles output power, uses a conventional P-P Ct'd OT, and still preserves the single ended signature. But has an issue with SE distortion current unbalancing the OT some. Various ways to fix that too. Simplest version looks like a differential output stage with a CCS tail, but uses a much higher gm device on one side of the pair (a low capacitance Mosfet or a high gm frame grid pentode with some fixup for its screen current to preserve DC balance). Can use SE drive to either side, or even conventional phase split P-P drive to both sides. KenPeter came up with a simple fix for the SE distortion current unbalance: just use two CCS tails, one for each output device and connect the tails with a big cap for the AC signal (low impedance AC there, so a big cap is needed).
There is a way to fix it up though. Either the external inductor scheme I mentioned above (not so hot).
Or search on "anti-triode". A number of threads on that, doubles output power, uses a conventional P-P Ct'd OT, and still preserves the single ended signature. But has an issue with SE distortion current unbalancing the OT some. Various ways to fix that too. Simplest version looks like a differential output stage with a CCS tail, but uses a much higher gm device on one side of the pair (a low capacitance Mosfet or a high gm frame grid pentode with some fixup for its screen current to preserve DC balance). Can use SE drive to either side, or even conventional phase split P-P drive to both sides. KenPeter came up with a simple fix for the SE distortion current unbalance: just use two CCS tails, one for each output device and connect the tails with a big cap for the AC signal (low impedance AC there, so a big cap is needed).
Last edited:
A bit late, but for the sake of the underlying issue...
When you say "cheap" I assume you mean a common toroid intended for mains power supply use?
If so, the two limiting factors are likely to be flux density and primary capacitance.
I understand that such transformers are usually very close to saturation density when unloaded and driven by the rated mains voltage. That gives you a known absolute limit to the ratio of drive voltage over frequency. So for half mains frequency the max drive voltage would be half of mains voltage, etc, pro rata. Er...I think. That would be unloaded, so for the loaded case you should take into account the drop in effective primary voltage caused by the current, the winding resistance, and ohm's law.
I haven't tried it myself, but I have seen reports that winding capacitance can be more of a problem. Always worth measuring this, and leakage, and primary inductance while you're at it, before you spend time designing an amp. Especially if you intend to use feedback.
All of which is why purpose-designed toroidal output transformers are not cheap.
Ian
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.