Are universal output transformers ok?

[Pixel]

Are universal output transformers ok? Or is it better to have a transformer that only has windings you are going to use?

[Helmholtz]

The only disadvantage of a universal OT I can think of, is that additional windings and taps use up some winding space.

So, everything else being equal, a dedicated OT could be wound with heavier gauge wire, resulting in lower DCR and somewhat lower copper losses.

Example: A 40W classAB amp has an RMS plate current of around 0.2A per tube at full power.
So every 25R increase of OT primary DCR per side increases OT losses and lowers available output by 1W.

BTW, the highly revered RS OTs of original Marshall JTM45s were universal type.

[Mick Bailey]

There are different types of universal output transformers - one with just a selection of output taps, which I've used quite often and just ignored the unwanted taps with no problems at all. Another type has additionally a selection of primary taps. With this type you get a table of input/output impedance values to suit a broader range of tube and speaker types but they're usually limited in output power and are larger and more expensive compared with a dedicated transformer. I've only seen these up to about 25W rating (and a good number of years ago), so I'm thinking they have limitations and are perhaps less efficient in their use of available bobbin space. There are some small 3.5W SE transformers available from OEP that have a range of 4-15K Ohm primary and 3-15 Ohm secondary that also work as P-P. Hammond list some universal types (125 series) that have additional secondary windings to give primary/secondary impedance choices but they only go to 15W, possibly due to the reasons of winding space.

[trobbins]

To add some more comment, it can influence the bandwidth - which may or may not be a concern for a guitar amp. If the primary is tapped or has settings like for 6k, 8k, 10kPP connections, then there is a tradeoff between PP DCR (as Helmholtz points out) and inductance which supports low frequency extension and lower distortion. If there are a lot of 'dark' windings (ie. unused) then that can increase the leakage inductance (because the windings are not as closely packed as they could be) and could increase the coupling capacitance (depending if the dark windings are connected to a 'hot' or 'cold' tap) and cause high frequency bumps (but for a good OPT they should only be a concern if the amp is using feedback). If the transformer has quite good specs for frequency range then it is likely well designed with interleaving - if it has no specs then you may want to check if the windings are interleaved.

[J M Fahey]

By definition a "universal" anything will make even greater compromises than usual; as mentioned above lamination window area is *strictly* limited, so thinner wire must be used.
A usually ignored but real problem: unused windings are still connected and increase unwanted winding to winding capacitance, loading signal out at higher frequencies.
Of course, universal transformers are problem solvers, when proper dedicated ones are unavailable and single unit custom winding is not justified ($$$$ + time).
Also for some higher primary impedance wirings, they do not have the required inductance to reach low, compared to a dedicated one.

[Enzo]

With un-defined terms like "higher frequencies", it might behoove us to determine just what higher frequencies we might be discussing. Surely the guitar amp won't care if there is more rolloff at 1kHz than "normal".

[trobbins]

If there is no feedback then there is likely no concern as OPT response peaks and dips are very likely to occur above circa 20kHz, where the response may be already well suppressed from prior amp circuitry (including pentode Ra). If there is feedback then all bets are off as a peak at the wrong frequency or wild phase shift may lead to instability, although the frequency of response peaks are likely to be above 10-15kHz. So my references to high frequency are related to the very end of of audible range and above.

There are often just so many if's and buts wrt OPT response, and that extends to the interaction with output stage tube type and its operating mode, as well as the OPT itself (including core design and when it starts to increase distortion at low frequency, and how well balanced a PP output stage is).

[Mick Bailey]

I saw a large Parmeko universal output transformer that was about twice the height of an Orange OR80 PT and the same footprint. It was truly universal in that it was rated for SE and PP but the owner was certain it was only 20W. It had 4 primary taps and additional secondary taps apart from the regular speaker connections labelled 'Line 1' and 'Line 2' - does anyone know what these would be for?

Edit, I just realized the line outputs are probably for a PA distribution network, but I'm not clear why they're in addition to the regular marked speaker outputs.

[Helmholtz]

I measured primary to secondary capacitance Cps (both windings floating) of a Marstran RadioSpares JTM 45 OT (universal type) and a number of dedicated Zaa OTs of similar size, e.g.:

Marstran RS: 2.9nF
Fender '64 SR: 2.8nF
Mesa SOB: 3.8nF
Marshall JTM 45 RI: 520pF

I measured many more OTs and could see a general trend that Cps increases with better coupling (lower leakage inductance) at least for dedicated Zaa OTs of comparable size.
That makes sense. It's known and logical that interleaving increases Cps.

In an amp, part of Cps is shorted because primary CT and one end of the secondary are grounded. This typically lowers the effective capacitance between primary and secondary by a factor of 4.

Still wondering how Cps affects the frequency response and stability of a PP amp, considering that capacitive couplings between both ends of the primary and the secondary should cancel because of opposite phase, assuming symmetrical coupling capacitances.

I tend to think that Cps is much less relevant than leakage inductance and primary self-capacitance - but always willing to learn.

[Pixel]

I saw a large Parmeko universal output transformer that was about twice the height of an Orange OR80 PT and the same footprint. It was truly universal in that it was rated for SE and PP but the owner was certain it was only 20W. It had 4 primary taps and additional secondary taps apart from the regular speaker connections labelled 'Line 1' and 'Line 2' - does anyone know what these would be for?

Edit, I just realized the line outputs are probably for a PA distribution network, but I'm not clear why they're in addition to the regular marked speaker outputs.

I have a SS PA amp that has an output transformer with 4, 8 and 16 ohm as well as 25V and 70V outputs.

[Helmholtz]

I saw a large Parmeko universal output transformer that was about twice the height of an Orange OR80 PT and the same footprint. It was truly universal in that it was rated for SE and PP but the owner was certain it was only 20W. It had 4 primary taps and additional secondary taps apart from the regular speaker connections labelled 'Line 1' and 'Line 2' - does anyone know what these would be for?
.

Maybe for a 70/100V distribution system. 100V/70V sound distribution was quite common in the 60s. My '66 JTM50 OT has a separate 100V output winding.

https://en.wikipedia.org/wiki/Consta...speaker_system

[trobbins]

I've just been through some output transformer testing, as a way to make measurements that can be compared to vintage measurements, and to make quick automated impedance and frequency response plots using modern available software with a soundcard. I'll get around to extending that to look at how the frequency response measurement interacts with high frequency amp stability when using feedback.
https://www.dalmura.com.au/static/Wi...asurements.pdf

I also just used those measurements on an output transformer in a small vintage 8W PA amp that I have on the bench to restore. Usually the PA amps just have 50V, 70V and 100V speaker outputs, which for a low powered amp means that the effective speaker impedances equate to 300, 600 and 1200 ohm which is not that convenient for direct connection to an 8 or 16 ohm speaker, so some wrangling of secondary windings is needed or otherwise a full replacement of the output transformer.

[Helmholtz]

I'll get around to extending that to look at how the frequency response measurement interacts with high frequency amp stability when using feedback.

Sounds interesting. Please keep us updated.