Output transformer - when needs must

[ChristianFletcher]

I have recently discovered an open circuit primarily of an output transformer built into the back of a old speaker of unknown origin that was rescued from the hoard. I’m just in the process of restoring an old cossor battery set missing speaker and transformer.

The speaker is in ok working condition and has a 2.2 ohm impedance. I’m going to be matching this to a 220HTP. The optimum load impedance for the output pentode is between 17-19Kohm depending on which tube data sheet is referenced. Which looks very high to me based on past experience?

After raiding the hoard again I find that I have a few spare output transformer but these are from more modern sets likely of the EL84 flavour with a much lower impedance ratio than I’m looking for probably 5-7kohm. I also have a new transformer from RS that may get me to 15k

Based on my experience I have found that perfect matching really isn’t necessary and wasn’t perfect even back in the day when these were built, but I still want to do the best I can.

If I want to use what I have which is really my only choice at the moment I’m thinking of either looking for a higher speaker impedance or wondered or if I can just increase the speaker impedance by adding some Additional R to the voice coil.

Just wondering if I’m likely to be walking into any traps here.

Many Thanks Regards Chris

[barrymagrec]

It seems unlikely that the speaker is actually 2.2 ohms impedance, DC resistance possibly - more likely 3 ohms nominal and such figures are frequency dependent and far from exact.

Adding series resistance to the speaker will reduce power to the speaker and probably affect frequency response.

[ChristianFletcher]

Thanks Barry

The impedance is 2.2 and the dc resistance is 1.72. Applying the 1.3 rule of thumb also gives 2.2.

I measured the impedance using an ac impedance bridge and an audio impedance analyser. I wouldn’t say this reading is particularly unusual. I don’t think I have a shorted voice coil this would normally show on the impedance bridge and the speaker has normal volume running off a signal generator.

I could be wrong!!

You may be right on the speaker output response. I guess I’m just going to give it a try and look for alternative output transformers.

Chris

[ChristianFletcher]

I just had a thought and wonder how many speaker have lower impedance than report because person making the test doesn’t null out the test leads! 0.6 ohms makes all the difference when your only measuring a couple of ohms.

Could there be an error in the optimum load valve I am referencing 17k sounds high?

[kalee20]

What is the output valve, triode or beam tetrode/pentode?

For beam tetrodes or pentodes, the optimum load is approx. 0.9 x Va / Ia where Va is the steady-state anode - cathode voltage (don't forget to allow for cathode bias resistor drop) and Ia is steady state anode current. Triodes are a bit more complex because generally the input signal has to be larger and you are constrained by not taking the grid positive.

For a mains output pentode, yes I agree 17k does sound high.

[barrymagrec]

220HTP / Mullard PM22A 2 volt battery pentode.

Mullard give 19k.

[electrogram]

If the secondary winding is on the outside you could remove some windings to increase its ratio

[RF Burn]

Chris,

I assume you mean the output valve is a 220 HPT?

According to the Cossor data sheet the optimum load is indeed 17K Ohms, which is not unusual for a battery pentode, where the anode current is kept quite low (for obvious reasons) and therefore a matches to a higher impedance load than a thirsty mains pentode. Indeed some battery tetrodes/pentodes where specified for optimum loads of up to 24K.

As the load impedance is pretty much a compromise anyway and specified to be at a point on the 2nd/3rd harmonic distortion curves where both are 'reasonably low' the actual value is not critical. It is unlikely (unless you have 'golden ears') you will be able to tell the difference between a 17K or 15K load, the size of the transformer and its construction are more likely to be noticeable though in terms of frequency response!

p.s. The RCA AR88 receiver is designed for a speaker impedance of 2.5 Ohms IIRC, so 3.0 Ohms is not always the answer, in any case it is only nominal (measured at 1KHz?) since speaker impedance varies greatly with frequency, especially at resonance!

Adrian

[ChristianFletcher]

Many thanks for the replies. Brilliant ideas and much food for thought. I really just wanted to pull in some ideas before getting started. The anode current is stated as 4mA in the radio service data for the 1937 TRF melody maker. Great idea about removing some windings on the secondary that could be an interesting experiment as I have some output transformers here in a poor state that wouldn’t be much of a loss.

I think I will have ago with a new transformer I bought from RS components it had is multitapped designed for valve amps and has a tapping for 15kohm to 4ohm. Really nice that RS are now stocking transformers again

Just a final question does the dc resistance of the primary winding affect the anode current. The set is biased using a interstage transformer and a -bias from a battery?


Many thanks for all the brilliant comments and idea. Just nice to talk these modifications through with someone first.

Regards Chris

[RF Burn]

Originally Posted by ChristianFletcher

Just a final question does the dc resistance of the primary winding affect the anode current. The set is biased using a interstage transformer and a -bias from a battery?

Regards Chris

It should not do unless the HT rail is very (I mean very) low. The DC anode current is controlled by the grid bias voltage.

Adrian

[Herald1360]

It'll have more effect on a triode than a tetrode/pentode.

[Diabolical Artificer]

Use a Zobel R/C, something like a 10r and 0.1u, across the OPT sec, this will make the primary Z less dependant on the speaker, whose Z varies with frequency anyway. The 10r needs to be 5W or so. Play about with the value's with 4r7 to 20r and 0.1u to 10n ish.

Andy.

[ChristianFletcher]

Thanks guys I haven’t tested the RS transformer yet in the radio, but it’s a perfect turns ratio for what I wanted. Anyway here’s where I got to as some people maybe interested. I’m no valve expert but just enjoy fixing any kind of stuff.

Thanks Regards Chris

Link to my transformer test video

https://youtu.be/7vVlvPiefpg