My brain was toast.
Heathkit W-5M.
Half above the equator, half below the equator.
Now that is balanced.
Heathkit W-5M.
Half above the equator, half below the equator.
Now that is balanced.
Saturation at a certain amplitude of a given low frequency occurrs when there is insufficient winding inductance in combination with insufficient core cross-section. Taking these factors into consideration, a toroid transformer can be designed for a given power output at defined low frequency that will tolerate some DC imbalance without approaching saturation. In principle, it is no different from a EI type transformer.
Its so simple to balance the currents why wouldnt you do it ?
People have said nulling for DC will unbalance the AC signal and introduce distortion - but I have yet to see anyone demonstrate this in real life. People also claim that having a small amount of DC imbalance in an EI transformer moves the transformer crossover point reducing transformer distortion - but again noone has ever demonstrated this as a real effect.
My advice is to use LM317 (good to 120v) with a zener for turn on protection - simples.
Also take note of the fact that different toroidals have widely different primary inductances so there are no useful generalities when discussing toroidals without measured specifics. Also parasitic capacitances are your primary concern and have the potential to ruin a project, manufacturing design will make different products behave wildly differently.
Shoog
People have said nulling for DC will unbalance the AC signal and introduce distortion - but I have yet to see anyone demonstrate this in real life. People also claim that having a small amount of DC imbalance in an EI transformer moves the transformer crossover point reducing transformer distortion - but again noone has ever demonstrated this as a real effect.
My advice is to use LM317 (good to 120v) with a zener for turn on protection - simples.
Also take note of the fact that different toroidals have widely different primary inductances so there are no useful generalities when discussing toroidals without measured specifics. Also parasitic capacitances are your primary concern and have the potential to ruin a project, manufacturing design will make different products behave wildly differently.
Shoog
Last edited:
Definitely enjoying the discussion, but can I follow up on something with the original question?
Let's say I'm using a dual tube with a shared cathode (829B, 832, 815, 5894, 6DZ7, etc.). Garter bias is out, because of the shared cathode.
With regulated screen and bias supplies, adjusting the bias individually per side can get it very close, to within 1mA at any given time, and drift will happen pretty slowly over time.
My thought is to use a 50VA toroidal power transformer with a split primary, for an output stage that tops out around 15W. With that trafo, for casual listening, is just using individual bias like that enough? Do I need to pop for one of those magic bias boards? Or is there another way to balance the current that works with a shared screen and cathode?
Let's say I'm using a dual tube with a shared cathode (829B, 832, 815, 5894, 6DZ7, etc.). Garter bias is out, because of the shared cathode.
With regulated screen and bias supplies, adjusting the bias individually per side can get it very close, to within 1mA at any given time, and drift will happen pretty slowly over time.
My thought is to use a 50VA toroidal power transformer with a split primary, for an output stage that tops out around 15W. With that trafo, for casual listening, is just using individual bias like that enough? Do I need to pop for one of those magic bias boards? Or is there another way to balance the current that works with a shared screen and cathode?
If you don't want bass, it'll work fine. If you want bass, get two 50VA coils and connect then like in post #8.
Example - one 25VA VPT12-2080 on a 6P1P output 15W @50Hz, but ~4W @25Hz. With two you get full power @25Hz. In my experience, 1mA mismatch is negligible.
Example - one 25VA VPT12-2080 on a 6P1P output 15W @50Hz, but ~4W @25Hz. With two you get full power @25Hz. In my experience, 1mA mismatch is negligible.
I used a dual tetrode with shared cathode in one project. Drift happened relatively quickly and significantly impacted the circuit. Don't recommend this solution.
Shoog
Shoog
I made an amp with GU29 (829B). I had to rebias it almost daily. Hence why I'm a big fan of the AB-4 and AB-Q. I have 7 GU29 tubes available BTW lol
Shared cathode tubes like the GU29 / 829B were designed for RF service, Class C push pull RF.
And they also were rated for Class AB1 push pull AF Modulator service (as in AM modulation for Communications, not for a Hi Fi AM transmitter).
Communications service is 300Hz to 3000Hz. Those frequencies typically do not saturate the push pull transformer laminations.
Now, if you use a common cathode tube that is designed for such a restricted frequency range modulator service, and want to increase the frequency range from 20Hz to 20,000Hz for Hi Fi service, you probably better be able to sense the plate current, and be able to adjust the quiescent current in each plate.
RF push pull is not a problem, nobody ever saturated an Air Core RF transformer (not Air Gapped laminations), like those tubes were meant to drive.
A higher power common cathode tube, the 5894, was designed for push pull RF service (Air Core transformer),
and for Class B push pull modulator service, where both plates have very low quiescent plate current.
By the way, neither the 329-B nor the 5894 are capable of Ultra Linear Push Pull service. They use a common screen, not just the common cathode.
The common screen is intended to be used with a Regulated screen voltage supply.
Pay attention to details, and then you can use common cathode RF tubes successfully for Audio output service.
And they also were rated for Class AB1 push pull AF Modulator service (as in AM modulation for Communications, not for a Hi Fi AM transmitter).
Communications service is 300Hz to 3000Hz. Those frequencies typically do not saturate the push pull transformer laminations.
Now, if you use a common cathode tube that is designed for such a restricted frequency range modulator service, and want to increase the frequency range from 20Hz to 20,000Hz for Hi Fi service, you probably better be able to sense the plate current, and be able to adjust the quiescent current in each plate.
RF push pull is not a problem, nobody ever saturated an Air Core RF transformer (not Air Gapped laminations), like those tubes were meant to drive.
A higher power common cathode tube, the 5894, was designed for push pull RF service (Air Core transformer),
and for Class B push pull modulator service, where both plates have very low quiescent plate current.
By the way, neither the 329-B nor the 5894 are capable of Ultra Linear Push Pull service. They use a common screen, not just the common cathode.
The common screen is intended to be used with a Regulated screen voltage supply.
Pay attention to details, and then you can use common cathode RF tubes successfully for Audio output service.
Last edited:
Or just parallel both halves and use twice as many 🙂 13k5 plate to plate is hard to find especially at decent power... paralleling makes transformer selection easier 🙂
BTW, I just found an open 10R cathode resistor in one of the monoblocs. It sounded a little off when pushed to high power yesterday and now I know why. 240mA on ole side, 120mA on the other. At low power it wasn't even noticeable - we're talking 20+ watts. YMMV, but these ones function alright with some mismatch.
BTW, I just found an open 10R cathode resistor in one of the monoblocs. It sounded a little off when pushed to high power yesterday and now I know why. 240mA on ole side, 120mA on the other. At low power it wasn't even noticeable - we're talking 20+ watts. YMMV, but these ones function alright with some mismatch.
Last edited:
Would this work properly if I parallel the secondaries on each transformer? The impedance ratio I'm after is 12k:8, which is really close to the 230V:6V on the Antek AS-0506 (11.7K:8).
From what I can tell, if I connect ala post #8, I end up with the equivalent of 460V:24V, or about 3k:8. But if I run all the secondaries in series-parallel, it's 460V:12V, or my original 11.7k:8 impedance ratio.
I could be way off, though. I'm still getting my head wrapped around transformers.
Those filaments take lots of power:
829-B 6.3V @ 2.25A (14.175 Watts).
The filaments and cathodes were made to allow very high peak cathode current (like when g1 is very positive, and the amp is in Class C).
30 Watt total plate dissipation (15 Watt per plate).
Parallel RF tubes filament power adds up very quickly.
30W plate/14.2W filament = ~ 2:1
6L6GC 6.3V @ 0.9A (5.67 Watts)
30 Watt plate dissipation
Parallel Audio tubes filament power adds up less quickly.
30W plate/5.7W filament = ~ 5:1
But then who is worried about efficiency, because I really like the looks of so many RF tubes.
829-B 6.3V @ 2.25A (14.175 Watts).
The filaments and cathodes were made to allow very high peak cathode current (like when g1 is very positive, and the amp is in Class C).
30 Watt total plate dissipation (15 Watt per plate).
Parallel RF tubes filament power adds up very quickly.
30W plate/14.2W filament = ~ 2:1
6L6GC 6.3V @ 0.9A (5.67 Watts)
30 Watt plate dissipation
Parallel Audio tubes filament power adds up less quickly.
30W plate/5.7W filament = ~ 5:1
But then who is worried about efficiency, because I really like the looks of so many RF tubes.
Last edited:
Toroidal cores saturate with very little DC current through them. Keep DC out, or be sure balance is very good in push pull applications.
It is all very simple really:
B = Bac + Bdc
If you have a toroid with practically zero airgap and a dc unbalance you will very quickly run up Bdc to astonishing high levels.
Calculating Bdc would take guessing out of the picture
B = Bac + Bdc
If you have a toroid with practically zero airgap and a dc unbalance you will very quickly run up Bdc to astonishing high levels.
Calculating Bdc would take guessing out of the picture
Last edited:
You can connect the secondaries as you like. All in series, all in parallel, or series/parallel. Post #8 really just demonstrated how to wire the primary side 🙂
Last edited:
Cool. I mocked it up in ltspice and I was right, I needed to run the secondaries in series-parallel to get the right ratio.
At $17.50 a pop, that's $70 for the whole amp. I think that's definitely worth a shot to see how well these tubes can stay in balance once they're burned in. And if it doesn't work out, I can swap for Edcor EIs, and reuse the toroidals for something with individual tubes that I can garter bias.
Thanks, everybody! This was really enlightening.
At $17.50 a pop, that's $70 for the whole amp. I think that's definitely worth a shot to see how well these tubes can stay in balance once they're burned in. And if it doesn't work out, I can swap for Edcor EIs, and reuse the toroidals for something with individual tubes that I can garter bias.
Thanks, everybody! This was really enlightening.