UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Components and Circuits (https://www.vintage-radio.net/forum/forumdisplay.php?f=40)
-   -   Circuits for toroidal push-pull output transformer (https://www.vintage-radio.net/forum/showthread.php?t=173339)

qualityten 21st Nov 2020 1:31 am

Circuits for toroidal push-pull output transformer
 
Please could someone point me in the direction of a circuit for a push pull amplifier using toroidal output transformers? The pair I have are 6K6 primary impedance into 8 ohms and were intended to be used with EL34 valves. They are rated at 30W and have 49% screen taps.

I’ve tested them on a Mullard 5-20 and they sound good to me, but I understand that the circuit should ensure balanced DC current to avoid saturating them and I don’t know how to do this. I have read discussions about including a constant current source.

My aim is to build a stereo power amplifier on a single chassis. The power transformer is from a Leak stereo 50.

Any advice or suggested circuits would be very welcome. Thanks.

Diabolical Artificer 21st Nov 2020 9:38 am

Re: Circuits for toroidal push-pull output transformer
 
AFAIK no special circuitry is employed using toroid OPT's, are these from Toroidy? Quite a few builds over on DIYaudio have used them.

To get DC balance in an OP stage two pots are used to tweak the fixed bias,look up GEC An approach to AF amplifier design. On my EL34 amps I used a 10k dual gang pot wired so that as one side of the OP goes up in bias,the other goes down, very simple. To dial it in you put your DMM one lead on one side g1, one lead on the other g1,turn the pot to you null,EG as close to 0v as poss.

This is with fixed bias though, easy to do, however as mentioned earlier AFAIK folk just build bog standard Williamson type designs with toroid OPTS,where did you read about the need for DC balance?

Re Toroidy,they do SE OPT's too and as you know SE OPT's can saturate, I've tried to find if Toroidy use gapped cores or some other wrinkle to no avail. DIYaudio's search engine is terrible, so no joy there.

Lastly a good way to check is to purposely unbalance the OPT's and see if it makes a difference at LF. If your using a cathode biased OP stage just put in a different sized cathode resistor on one side, or better a fixed R and pot (which BTW will balance an OP stage) then do a sweep from 10/20hz to 100hz ish noting any distortion of the waveform.

Lastly, lastly El Paso on youtube did a video on using pots to balance an amp in terms of AC & DC, it does reduce THD.

Andy.

trh01uk 21st Nov 2020 10:23 am

Re: Circuits for toroidal push-pull output transformer
 
I rarely comment on anything touching on audio, hi-fi, etc, however this thread has me intrigued. Why would the construction style of the output transformer require any particular circuit?

I am aware of the need to balance the DC currents through the two output active devices to avoid pushing the transformer into saturation (as the OP notes), but as far as I know that applies equally to standard C-core, I-E core transformers just as much?

Noting Andy's answer to the way to get that DC balance, I would suggest that another traditional way to do it was to have the same bias on both valves, but ensure the valves are a matched pair.

Richard

GrimJosef 21st Nov 2020 10:36 am

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by trh01uk (Post 1312692)
... I am aware of the need to balance the DC currents through the two output active devices to avoid pushing the transformer into saturation (as the OP notes), but as far as I know that applies equally to standard C-core, I-E core transformers just as much? ...

It does apply but not, if I understand it correctly, equally.

The geometry of toroids means that the flux density in the iron is much more uniform than in, say, an EI core. And the flux leakage is much less. These details mean that saturation occurs much more suddenly in a toroid. Cost pressures being what they are, commercial manufacturers tend to operate their transformers as close to the saturation limit as they can get away with. If something like DC imbalance pushes the transformer into saturation the sudden onset will make the consequences in a toroid worse. So it's more important to make sure that it doesn't happen.

All that said, as Andy and Richard have pointed out, the circuit techniques for achieving DC balance are the same whichever core geometry we might choose.

Cheers,

GJ

trh01uk 21st Nov 2020 11:17 am

Re: Circuits for toroidal push-pull output transformer
 
Thank you GJ - very informative. Just taking this a little further, it seems to me that achieving balance at the quiescent current point (i.e. no signal) is one thing - and easily done by the techniques already mentioned.

But if we then ask what is happening instantaneously, if the two valves have different gm values, then the instantaneous values of anode current could well be quite different. Obviously getting worse the nearer you get to the peak of the driving waveform.

If toroids are more tricky in this regard, would it be fair to say that you really want valves that are matched for gm?

Richard

GrimJosef 21st Nov 2020 12:00 pm

Re: Circuits for toroidal push-pull output transformer
 
Ideally yes. But you can compensate for that too of course, either with a pot controlling the relative drive signals to the two output valves or by choosing a phase-splitter valve (if it's a long-tailed pair splitter) where the two halves of the valve have an equal and opposite discrepancy in their signal gains to the output valves. In an amp like the Quad II, where the two halves of the phase-splitter are swappable EF86s, it can be remarkable how much the distortion can be made better or worse by moving the valves between sockets.

Cheers,

GJ

Diabolical Artificer 21st Nov 2020 7:07 pm

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by trh01uk (Post 1312704)
Just taking this a little further, it seems to me that achieving balance at the quiescent current point (i.e. no signal) is one thing

All you can do normally is set the amp for the best conditions at quiescent/0v IP/DC, then tweak the AC balance under various operating parameters to get the best THD reading, but we are usually only talking about a few 1/100th's of a percent, mostly in-audible.

There are circuits allowing for dynamic bias,see - http://www.turneraudio.com.au/300w-5...tabilizer.html ,but it's pretty nigh impossible to achieve ideal conditions. If you match valves at what parameters do you match them? Within 5% Ik? 10%? If you then try to match gm as well as cathode current it gets more tricky, you need about 20 valves to get a good matched pair to within 10%,obviously it would need more if you tighten the matching specs. you can go further to match caps as well as resistors, it gets silly quickly.

As I see it there are at least two approaches to valve amp design 1) As best as possible hifi - low THD, excellent frequency response from 10hz to 100khz,all that jazz, or 2) as most folk like valve amps for their "sound", design for high-ish 2H distortion; the best "sound" and not worry too much about adding colour.

Back to the OP's question, rather than trying to find a schematic/design that fits the OPT, it might be better to design an amp around the OPT, the OPT in a valve amp is the component that's the most variable, every OPT has different leakage inductance,inter-winding capacitance etc. It's pretty easy to knock up an OP stage, once that's done the rest of the circuit is even easier.

Andy.

Ed_Dinning 21st Nov 2020 8:17 pm

Re: Circuits for toroidal push-pull output transformer
 
Hi Folks, given that the flux density in a toroid is more linear than in a stacked core, by operating at high flux densities this advantage will be negated by going further up the curve into a non-linear region.
The only advantage may be a more compact or less costly transformer.

However, on the windings side, it is then very difficult to arrange sectionalised windings for closest coupling and reduced leakage inductance.

So probably not worthwhile.

If it was an expensive item and you have the money, possibly better to go for "C" cores.

Ed

qualityten 23rd Nov 2020 10:45 pm

Re: Circuits for toroidal push-pull output transformer
 
Thanks for the replies to my query. I'm really busy at work, so haven't time to pursue this for a while, but am hoping to gather ideas and resources for a build later. The transformers themselves were made by Canterbury Windings in 2008 and are big, compared with their EI counterparts: 13cm in diameter, 7cm high and weighing 2.5kg each. The screen taps are at 43%, not the 49% I stated earlier.

As I say, they sound pretty good when slotted straight into a Mullard 5-20 amplifier, with no hum. Perhaps a little lighter in the bass than the other EI transformers, but subjectively greater clarity. Perhaps the lighter bass is due to some saturation?

I've done a bit of reading on the web. Apparently all toroid roads lead back to Menno van der Veen in Europe and the Plitron company in the USA.

joebog1 24th Nov 2020 7:54 am

Re: Circuits for toroidal push-pull output transformer
 
Better bass would mean non saturation!!. Saturation would mean better "fuzz".
I have Menno van der Veens book, and there is no magic soup contained therein. The circuits are very standard indeed. In fact driving 8 yes eight 6CA7/EL34 with a measly 12AX7/ ECC83 is in fact pretty rude. ANY decent push/pull circuit will suit. In fact some of the very classic English amplifiers ( read your own post!! ) will work superbly. Ed Dinning makes a very valid point!! Coupling on a toroid is pretty easy to do with 50 Hz power supplies, but is an absolute nightmare with a toroid.

Joe

Diabolical Artificer 24th Nov 2020 8:07 am

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by qualityten (Post 1313677)
Perhaps the lighter bass is due to some saturation?

Saturation shows up as a distorted sinewave, not really sure how it sounds.

Quote:

Originally Posted by qualityten (Post 1313677)
Apparently all toroid roads lead back to Menno van der Veen in Europe and the Plitron company in the USA

By that do you mean all toroidal cores are made by these big companies?

Andy.

daviddeakin 24th Nov 2020 9:41 am

Re: Circuits for toroidal push-pull output transformer
 
Toirodal transformers are no more or less sensitive to DC offset than regular EI transformers. The idea that they 'saturate sharper' is a bit of an old wives tale. Maybe it was true for one particular manufacturer, based on their particular construction details, and the notion got passed around, but it's not true in general. Neither of them like DC very much (a couple of milliamps DC can be enough to trash the primary inductance), and both of them saturate in a broadly similar manner.

There are various ways to buy balance in your outout stage, but by far the easiest is to use individual bias pots for each valve. You could adapt most any circuit to have individually adjustable bias if need be. Matched tubes help of course, but that is a secondary issue.

GrimJosef 24th Nov 2020 10:31 am

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by daviddeakin (Post 1313741)
... The idea that they 'saturate sharper' is a bit of an old wives tale ... but it's not true in general ... and both of them saturate in a broadly similar manner ...

Sorry to be an old wife, but as I understood it the flux density in a toroidal core is typically more uniform than in an E-I core. The transition to saturation will therefore be sharper because the whole core will saturate at once rather than the high-density bits first, then the mid-density bits, then the low-density bits. Is some part of that explanation untrue (I realise it's oversimplified, because obviously as parts of an E-I core saturate the flux distribution inside the core will be altered) ? If it is an untrue then I'm sure there are measurements somewhere confirming that. Could you point me to any ?

Cheers,

GJ

woodchips 24th Nov 2020 11:33 am

Re: Circuits for toroidal push-pull output transformer
 
I have to ask, how can an ECC83 drive two, let alone eight EL34's?

With only 1.2mA anode current and the capacitive loading of the valves and wiring it will be struggling. If distortion is what you are after then fine, get a string of triangular waves, but for reasonable hi-fi it seems a non-starter.

GrimJosef 24th Nov 2020 11:44 am

Re: Circuits for toroidal push-pull output transformer
 
An ECC83 drives a pair of EL34s in the classic Mullard 5-20, the grand-daddy of many commercial EL34 amps http://www.r-type.org/articles/art-003d.htm. Mullard were proud of the EL34's high gm, which it inherited from the EL37. A Mullard engineer is reputed to have said that the latter need just "a whiff of signal" to drive it.

Cheers,

GJ

kalee20 24th Nov 2020 12:49 pm

Re: Circuits for toroidal push-pull output transformer
 
I'm watching this with interest, as I'm making a push-pull amplifier with a toroidal output transformer.

It seems obvious to me, is that there should be two balance adjustments.

One will balance the DC currents, so the core is not biased one way or the other, and caters for the valves having slightly different emission characteristics.

The other balances signal drive to the output valves so that the AC voltage on the grids is the same, due to phase inverter lack of balance - or even better, having the AC component of the output valves equal (which would take into account slightly differing gm's in the output valves as well as phase inverter unbalance).

Small current-sense resistors (10Ω) in the cathode return would allow monitoring, for setting balance of both.

Quote:

Originally Posted by GrimJosef (Post 1313763)
Quote:

Originally Posted by daviddeakin (Post 1313741)
... The idea that they 'saturate sharper' is a bit of an old wives tale ... but it's not true in general ... and both of them saturate in a broadly similar manner ...

Sorry to be an old wife, but as I understood it the flux density in a toroidal core is typically more uniform than in an E-I core. The transition to saturation will therefore be sharper because the whole core will saturate at once rather than the high-density bits first, then the mid-density bits, then the low-density bits. Is some part of that explanation untrue

It's largely true, although a slight oversimplification, TBH. A toroid's saturation curve is slightly softer than the parent material.

The reason for this is that the iron around the inner circumference is a shorter path than nearer the outer diameter, so flux density will be higher. It will therefore hit saturation first, and drop out of the picture. The available core area will then be reduced, so inductance will fall slightly... and so it goes on until the whole thing saturates.

Quote:

Originally Posted by woodchips (Post 1313796)
I have to ask, how can an ECC83 drive two, let alone eight EL34's?

Well... Quad used puny EF86's to drive KT66's!

qualityten 24th Nov 2020 4:33 pm

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by Diabolical Artificer (Post 1313727)
"Perhaps the lighter bass is due to some saturation?" Saturation shows up as a distorted sinewave, not really sure how it sounds.

"Apparently all toroid roads lead back to Menno van der Veen in Europe and the Plitron company in the USA" By that do you mean all toroidal cores are made by these big companies?

Andy.

@Andy: I will look at the output sinewave in my present lash-up, using a single transformer and single 5-20, when I get a moment.

My other statement was just that all my searches for information on and circuits of push pull amplifiers with toroidal output transformers ends up referring to v d Veen's books and articles, or references to Plitron products, though there is a well-regarded Polish company that also makes toroidal OPTs and Canterbury Windings is still in business.

None of the manufacturers, to my knowledge, offer sample circuits for hobbyist constructors.

@Kalee20: have you decided on a circuit you could share?

joebog1 25th Nov 2020 12:06 am

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by woodchips (Post 1313796)
I have to ask, how can an ECC83 drive two, let alone eight EL34's?

With only 1.2mA anode current and the capacitive loading of the valves and wiring it will be struggling. If distortion is what you are after then fine, get a string of triangular waves, but for reasonable hi-fi it seems a non-starter.

I have absolutely no idea, woodchips, but Menno van der Veens circuit does it!!
I did warn you there is no magic soup contained in the book. In my opinion the book is the largest piece of advertising for his toroids. Chapters in the book have a bit of basic theory on voltage amplifiers ( a single twin triode) then a phase splitter, the details of which come directly from Mullards book, and following the phase splitter are circuits for 2, 4 or 8 6CA7's.
ALL the circuits require van der Veens transformers however.

If one cuts n pastes 2 or 4 or 6 etc 6CA7,s, and keeps pasting them in parallel you have van der Veens circuits. I have stored the book because it doesnt have any real information, but if you insist I will dig it out and scan a few pages.

Joe

joebog1 25th Nov 2020 12:23 am

Re: Circuits for toroidal push-pull output transformer
 
1 Attachment(s)
I dug out the book and I was wrong. He uses a 12AU7/ecc82 as the driver for 8 6CA7/EL34.
If anybody wants the circuits, just ask.

Joe

Diabolical Artificer 25th Nov 2020 7:50 am

Re: Circuits for toroidal push-pull output transformer
 
Quote:

Originally Posted by qualityten (Post 1313905)

@Andy: I will look at the output sinewave in my present lash-up, using a single transformer and single 5-20, when I get a moment.

you'll see the positive going + peaks get a chunk taken out of them, THD goes up, phase changes as well as amplitude. you need to keep an eye on whats going into the OP stage though to to make sure the signal loss isn't due to RC rolloff. It's even worse with no load, don't do this test at high OP though.

Ahh, penny drops, that Van der Neen, thanks for the PDF Joe.

Back to you OP, just build a 5-20, if you have any issue's tweak the design. Build a 5-20 circuit,don't connect the FB, run some sweeps going from low OP to full whack, then do a few sweeps at low OP with no load, with 100n load, then 1u load, record all this using REW or better paper. F sweep should be from 10hz to 100khz, REW and PC based software will struggle here at 100hhz.

This will show any saturation and instabilty, you can then add or tweak the design, i can help you here as will others.

I agree what others said about using an ECC83 as PS/driver, an ECC83 makes a poor LTP, a LTP generates more 3H than a cathodyne PS, due to a Z mismatch of each side. A LTP using a 12BH7 or similar with a CCS as the tail takes a lot of beating, but this requires a negative rail. That said Mullard and others used the ECC83, but then they were trying to flog new valves. I'm going on a loads of tests I ran trying to get very low THD, but I was also driving 6 EL34's, not 2. i got to less than 0.3% at 120w, 0.02% at 10w - not bad.

Andy.


All times are GMT +1. The time now is 7:29 pm.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.