Output Transformer - what do I have?

[frankmcvey]

Bought this beautiful beefy NOS Haddon output transformer and its matching 10H choke years ago, and it's just seen the light of day again during my current clear-out. Problem is, there are no markings on the transformer other than tag numbers. The label on their (looks and fits like original) box says 30W Ultralinear Transformer, 10H Choke and that's all.

Measuring the transformer primary windings between tags 1-5 with a Fluke 25, we have:

1 70R 2 50R 3 50R 4 70R 5

On the secondary we have 1.4R between tags 1 and 3,

1.2R between tags 1 and 4

I can't see the windings to estimate what gauge wire was used. The transformer weighs 2.4Kg and the laminations are 80 x 95mm. Can any of you transformer gurus guesstimate what output valves this would suit from this meagre info, or perhaps ID it from an old catalogue?

Cheers,

Frank

[Radio Wrangler]

It cannot be done with any certainty. All the resistance readings tell you is that the wire gauge is quite generous and that the winding plan keeps things well balanced.

You will have to drive one of the windings with some signal or low voltage mains so that voltages can be measured, and they will give yiu tutns ratios and from those impedance ratios can be reckoned.

The 10H choke mention is a mystery in terms of what it means.

David

[frankmcvey]

Sorry if I didn't make that clear, David. The transformer and it's matching smoothing choke came in the same box, hence the slightly misleading label. By matching I mean that it's cosmetically identical in appearance and paint finish, other than the underside tag layout/numbering.

[frankmcvey]

Hi again, following David's advice, I fed the txfmr primary tags 3 and 5 (CT to end of winding) with 1KHz 30Vpp; appearing at one of the outputs (tags 18-20) I get 1.5Vpp.

A 55 year old memory is muttering something about turns ratio squared, but I'll have to look it up!

Cheers,

Frank

[trobbins]

It's likely a 6.5k PP primary to 4 ohm secondary.

Best to take measurements between primary 1-2, 2-3, 1-3, and secondary 1-3, 1-4. And then swap to drive primary 1-3, and measure voltages on 3-4, 3-5, 3-5. That ensures the OT is good working health, and the UL % can be determined, as that would identify the valves that it was designed to work with.

How accurate is your meter? Does it only measure to 0.1VAC, and is it ok at 1kHz ?

[frankmcvey]

Thanks for the reply. I'll make your measurements in the morning - it's 1am here! The ac measurements were pp, made on the scope, so it's eyeballing, but pretty close, I'd say. Just spotted that I've made a typo in the first post. The secondaries are 1-3 and 2-4.

Cheers,

Frank

[Radio Wrangler]

Voltage ratios will be equal to turns ratios, so that's easy.

Square a turns ratio to get the impedance ratio of one winding to another.

David

[trobbins]

Originally Posted by frankmcvey

The ac measurements were pp, made on the scope, so it's eyeballing, but pretty close, I'd say. Just spotted that I've made a typo in the first post. The secondaries are 1-3 and 2-4.

Yeh, eyeballed pk-pk measurements are fine if that is all you can do, just as long as all measurements are the same 'units'. Due to your measurement tolerances, there will be a bit of tolerance on the calculated impedance for primary, given that the secondary impedances (for the two windings) are used to 'match' a common speaker impedance value (eg. 4 ohm with both in parallel and 16 ohm for series connection).

Calculation spreadsheet attached if you have MS Excel, and are unsure of the calculations required.

[Diabolical Artificer]

From the size and other info I'd say Mullard 5-20 clone two x EL84/6V6. It's not big enough for EL34's.

Andy.

[John Caswell]

Contrary to what everyone else seems to do I always measure my output transformers with low volts into the secondary and measure high volts out. Struck me as easier. Haddon used to make the transformers for Leak.

John

[Radio Wrangler]

Measuring the voltages with a scope doesn't look terribly accurate, but then a lot of people use many-digited multimeters, but well above their roll off frequency for AC volts and are unaware of the errors they get. They get a 4 digit number and trust every last digit of it. So I'd go with the scope.

Driving a low voltage winding with a low voltage is very appropriate and likely will yien voltages more comfortably within the ranges that can be measured.

David

[Skywave]

Originally Posted by frankmcvey

I fed the txfmr primary tags 3 and 5 (CT to end of winding) with 1KHz 30Vpp; appearing at one of the outputs (tags 18-20) I get 1.5Vpp.

For testing a mains xfmr. - which is designed to be used at 50 or 60 Hz - I wouldn't use 1kHz. That relativity 'high' freq. could give misleading results. If you have the kit, try 100 Hz; ideally 50 Hz. (For audio O/P xfmrs., 1 kHz is O.K.)
I'll also add that measuring the resistance of inductors - which, therefore includes chokes and transformers - with a DVM can also give misleading results. If you have an AVO 8 (or similar analogue meter), use that: the differences in resistance readings might be significant.

Al.

[frankmcvey]

Hi, All, thanks for the replies and advice. So here are my results:

Input Voltage at primary tags 1-5 (ie HT end to HT end) 32Vpp

Output Sec winding 1 (tags 18-20) 0.75V
Output Sec winding 2 (tags 19-21) 0.75V

This would give us a voltage, hence turns, ratio (T) of 32/0.75 = 42.66

Impedance Ratio = Tsquared = 42.66 x 42.66 = 1820.

Assuming that each of the 2 (separate) secondaries windings are for an 8 Ohm load (which would give the possibilities of either a 4, 8 or 16 Ohm speaker)

Primary Impedance = 1820 x 8 = 14560

Since the recommended load for an EL84 is 6-8K (valve museum data) it looks as if I have about double that. Am I correct in using tags 1-5 as my primary input - should I be using tags 1-3 ie CT to HT instead of HT to HT?

[trobbins]

My calcs indicate you have a nominal 7.2k PP to 4 ohm (parallel) or 16 ohm (series) OT.

Did you get to measure the primary taps, in order to identify the UL % ratio - as that may narrow down the intended valves to be used with the OT?

[frankmcvey]

Thanks - PP being plate/plate? Yes I did take measurements at the screen taps - With 32Vpp fed to tags 1-2 I see 2.3V on each secondary
With 32Vpp fed to tags 4-5, I see 2.3V as well.

Thanks for your help.

Cheers,

Frank

[trobbins]

The screen tap to plate end therefore provides a nominal 13.9:1 turns ratio to a secondary, whereas the B+ to plate provides a nominal 21.3:1. So the screen tap to B+ would provide about a 7.4:1 TR, or 35% turns ratio, or '% common winding'.

[GrimJosef]

If they meant it when they said 30W then that'll be for EL34s (or possibly KT66s/6L6s pushed a bit), not EL84s.

The Mullard 5-20 output transformer is spec'd as 7k primary impedance with 20% screen taps or 6k6 with 43%. The choke for that amp is indeed a 10H one.

Cheers,

GJ

[Diabolical Artificer]

"If they meant it when they said 30W then that'll be for EL34s" Yes but look how big it is, looks about 1 1/4" wide lamination stack, possibly 30w but on the small side.

Andy.

[Radio Wrangler]

Power ratings for audio transformers aren't absolutes. There is a factor of how much flux density the designer wants to run it to. Linearity suffers. No transformer has zero distortion, so for a spec number when the answer to the question is really a curve, where do you pin the tail on the donkey?

There's only a limited number of valve types it is likely to be intended for, and for each of them a limited number of designs that it would be worth bringing out a transformer as a product for. So it's likely to either have been done on contract for an amplifier manufacturer, or to suit a published design.

Radford was primarily a transformer manufacturer.
Quad and Leak transformers will be instantly recognised on here.

Who else was there? Rogers, Armstrong and the guitar/stage PA firms.

For published designs, they were dominated by 'Applications department' output from Mullard and GEC. It could be for something in the pages of wireless world, or a kit provider.

But, for someone with one transformer in his hand, all the parameters can be measured. Some easily, some more difficult, and then a valve pair and bias setting chosen. It would be easier if the original intent was known, but without it, it's not impossible.

David

[GrimJosef]

Originally Posted by Diabolical Artificer

... Yes but look how big it is, looks about 1 1/4" wide lamination stack, possibly 30w but on the small side ...

It's a lot bigger than the Quad II's 15W unit, when you get that out of its can. The Mullard 5-20 chassis drawings show a hatched footprint for the whole of the output transformer which is just over 3" by just over 3.5". Although they call it a 20W amplifier they actually test it up to 27W. I'd wager a small sum that you were right when you identified this as the intended use back in post #9.

Originally Posted by Radio Wrangler

... Radford was primarily a transformer manufacturer. Quad and Leak transformers will be instantly recognised on here.

Who else was there? Rogers, Armstrong and the guitar/stage PA firms ...

The guitar/PA people cared more about power and less about objective fidelity, so ultralinear was used less often there.

Smaller hi-fi players included RCA (again with very distinctive transformers), Pye, Beam Echo, Pamphonic, Lowther, Sugden and several less well-known (these days) others.

Cheers,

GJ