Interstage transformer Vs valve phase splitter for push-pull amp

[audion_1908]

Hi I could do with some advice I have built a few small basic valve class a amps before but not a valve push pull amp, it will be built as a learn fun and for use with a small radio.
So what are the pros and cons of an interstage transformer phase splitter Vs a valve phase splitter for a push-pull amp?
Any advice welcome

[Diabolical Artificer]

Pro's good/high voltage swing/gain, con's less than perfect frequency response, large phase shift, high price, that's it in a nutshell. The consensus nowadays is to avoid an IS tfmr, if you need big voltage swing use a choke coupled stage, otherwise use a LTP or cathodyne PS.

What are you trying to drive?

Andy.

[audion_1908]

I am a bit lost? I would like to build a low ht (12v to 90v) audio amp mostly out of stuff I have and that means tv valves, it will probably be used with a diy radio tuner.
the amp will be useful as a test bed for low ht audio for use in other radios

(What are you trying to drive?) ?

Wanted broken AC/DC Bakelite radio for conversion to 12v
https://www.vintage-radio.net/forum/...d.php?t=166701

[Radio Wrangler]

If you're going to be working with limited HT voltage, then transformer coupling has a big advantage of making the most of what HT you do have.

It will restrict the frequency response of the amplifier and add even more phase shift which will limit how much feedback you can use, if you choose to use any.

It will be hard to find one. Period parts will date back a long time to when interstage audio transformers were more common. They need a lot of inductance to give good frequency response at the low frequency end, given the high impedance environment in which they must work. This means a great many turns of very fine wire. Back in the day, varnish coating of copper wire was less reliable, moisture got in through cracks and you got 'Green spot' corrosion. Many transformers have died just in storage, so any remaining workers are sought after by people restoring old sets, plus they have a certain cachet amongst boutique amplifier people. This makes them expensive. All those turns and very fine wire on a small bobbin makes them a pig to wind. It also makes for large stray capacitances and low self-resonant frequencies, so you audio frequency range gets eaten away at the top end as well.

If you want to make a transformer with widest bandwidth, one old trick in the RF world is to make the circuitry around it as low impedance as possible. This is true also for audio frequency range stuff. Unfortunately, the high-Z world of the audio interstage transformer pushes it in precisely the opposite direction. The interstage transformer is a far harder thing to make work than the amplifier's output transformer, and even done to the best leaves you with worse remaining compromises.

Most designers were very glad to leave the things long behind.

You still find them in some modern boutique amplifiers. Feedback is pooh-poohed in these circles, so their phase shift isn't a problem. The low frequency roll-off and high frequency roll off seem to be liked because they focus listener's attention on the range of human voices. So they have an odd sort of following.

Oh, another problem; as the stage driving them will be single-ended, there will be a net DC flux from the bias current. To avoid the core being saturates, it will have to be gapped. THis brings the inductance down and spoils the low freq response further.

Interstage transformers were very common in push-pull transistor output stages, but their low impedance world made them a far easier task, and most of these things went in radios with small speakers and limited bass response.

I'd avoid them like the plague.

The concertina phase shifter needs high HT to work well, the long tailed pair uses an extra valve, but doesn't need so much HT.

David

[GrimJosef]

One advantage of a transformer, in the abstract at least, is that you can be sure of genuinely symmetric drive to the output valves. You can get close - many would say more than close enough, given the inevitable mismatch between the output valves - with transformerless circuits. But it's hard to get, and maintain, perfect symmetry without a transformer.

A second advantage is that the DC resistance of the transformer's secondary will be very much lower than a typical grid leak resistor. So if your precious vintage output valves are a bit gassy then with a transformer you won't have nearly as bad an issue with grid current raising the grid's DC voltage to dangerous levels. On the other hand, unless the secondary is split into two separate halves, you won't be able to balance the quiescent DC currents through the output valves by adjusting their individual grid bias voltages.

Cheers,

GJ

[audion_1908]

Ok lots to think about, so the phase splitter transformer is definitely out, so will use an extra valve, a phase splitter.

thanks, now I know what I need to concentrate on learning about

[PJL]

Not to forget an important benefit, there are no coupling capacitors to go leaky and wreck the amplifier.

[audion_1908]

Originally Posted by PJL

Not to forget an important benefit, there are no coupling capacitors to go leaky and wreck the amplifier.

Modern plastic film capacitor are great, and I always pick ones higher voltage than I need, at least 100v over, I am not worried, I suppose in the time the coupling transformers where common, they where probably cheaper than the wax paper capacitors

[kalee20]

Lots of good points made already.

Pro's:

* Can provide some voltage gain.
* Can give excellent drive balance.
* Can give low DC resistance in grid circuit of driven valve(s).
* Can give completely isolated drive, no ground loops (nobody mentioned this so far!).
* Allow driver valve to operate at near full-HT on its anode.
* Failure doesn't put +ve volts on driven valve grid.
* Can have unlimited life.

Con's:

* Expensive.
* Can give rise to distortion.
* Very difficult to get a wide, flat frequency response.
* Phase shifts make applying global negative feedback, to any great extent, tricky.
* In practice, 1930's transformers are more likely to be dead than 1930's valves.

Originally Posted by Radio Wrangler

The interstage transformer is a far harder thing to make work than the amplifier's output transformer, and even done to the best leaves you with worse remaining compromises.

...

Most designers were very glad to leave the things long behind.

...

Oh, another problem; as the stage driving them will be single-ended, there will be a net DC flux from the bias current. To avoid the core being saturates, it will have to be gapped. THis brings the inductance down and spoils the low freq response further.

Amen to all that, from someone who's designed and made both!

A compromise might be a huge inductance for the grid bias resistor, well grounded DC and open for AC, just what you need. A couple in series perhaps one for low frequencies one for high, saves all that layer winding. No current in non fault conditions either.

[audion_1908]

Originally Posted by merlinmaxwell

A compromise might be a huge inductance for the grid bias resistor, well grounded DC and open for AC, just what you need. A couple in series perhaps one for low frequencies one for high, saves all that layer winding. No current in non fault conditions either.

You have lost me?

[Ed_Dinning]

There were recently some small interstage transformer kits available on ebay. The looked like ex BBC parts and were ready to wind to your own designs, but the price was pretty high for one, never mind a stereo pair and you still then had to source the wire and wind them.

Ed

[audion_1908]

Originally Posted by Ed_Dinning

There were recently some small interstage transformer kits available on ebay. The looked like ex BBC parts and were ready to wind to your own designs, but the price was pretty high for one, never mind a stereo pair and you still then had to source the wire and wind them.

Ed

Expensive doesn't sound good, and I only need mono

[Diabolical Artificer]

The problem your going to face is the low voltage. Most low voltage amps like the one your on about use single ended OP stages, so no phase splitter. If you used a cathodyne PS gain is split between in and out of phase signals, a LTP is good but again your losing about a 1/3 of your HT for the tail. However if you could get a negative supply, -40 to - 80v, that solves that problem. with this you can use a CCS to get near perfect balance. That said I don't think you really care about all singing and dancing hi fidelity, so some of this won't matter.

"A compromise might be a huge inductance for the grid bias resistor...." "You have lost me? " I think
what Merlin was refering to was if you use a grounded cathode stage with grid bias, usually a good few meg, you could use instead an inductor, but it's not 100% clear as you say.

Hammond do reasonably priced inter stage tfmr's, about £30 - see - http://hammondmfg.com/124.htm this compared to a vintage IS tfmr that could cost well in excess of £100.

Lastly, have a read here - http://www.valvewizard.co.uk/ he's after designing guitar amps, but his articles can be applied to any valve amp.Also look on line, there are loads of free PDF's of old books about amplifier design by authors like Crowhurst, Morgan Jones etc etc.

Andy.

[audion_1908]

Thanks Diabolical Artificer

I now have a stack of used nicad cells so can up the HT voltage to 45 or 80v or so

Thanks for the valvewizard link I will take a wile to work through it all!

[Diabolical Artificer]

Just download and read the PDF on Triode Gain Stage, this gives you a good basis for valve amp design. Whilst reading it set up a test circuit, it'll make more sense. If you just use something like a 47k for the anode resistor, 1m or 470k for the grid leak, you only then have to work out the cathode resistor. Once you've drawn a load line a few times and calculated the cathode resistor, the rest should make more sense and fall into place.

Good luck, andy.