Ordering guitar amp OT

[Helmholtz]

Look at those meager primary inductance numbers.
That's because they are measuring at 1V/1kHz, which IMO doesn't make much sense.
It just gives the minimum inductance, corresponding to initial permeability.
(1V across the primary corresponds to an output of 0.25mW)
As soon as you increase the measuring voltage and/or lower the measuring frequency the inductance increases.
Values might be 10 times larger at 240V/50Hz.
Unfortunately the factor 10 is not a fixed number for all OTs but depends on core, turns number and effective airgap.
I'm typically seeing an impedance maximum around 230V to 270V at 50Hz.
At even higher voltages the inductance starts to drop again indicating beginning LF saturation, but never gets as low as with 1V signal.
This means that bass response is typically best at some medium output and somewhat worse at full output.

The Hammond table also shows primary and secondary impedance data (Zpri, Zsec).
Again those numbers are rather useless as they depend on actual inductance and thus vary strongly with measuring current.

[Helmholtz]

As the path of leakage inductance is mainly through air, leakage inductance should be rather constant.
But primary inductance strongly varies with magnetizing current.
As will the ratio Lpri/Lleak.
So what's the correct Lpri/Lleak ratio?

[R.G.]

I'm thinking that the variation of primary inductance might be one of those things that get more complex the closer you look, and eludes a simple answer.

Yeah, initial permeability is lower, maxes out on the middle of the BH curve, and falls off again as you get to higher flux densities. Exactly how much the inductance varies depends on the core material, the core stack pattern, the degree to which the core is jogged together and squared up, the size of the excitation Volt-time, and probably the phase of the moon.

It's a fact that the bass response falls off at small signal levels to some degree, and picks up as you use more output. An OT used mostly at 1W has less bass than if it is used at 10W or 30W. For accurate comparisons, you have to specify the test level. But in general, you can't know the variable primary inductance until you get your hands on the OT, or just work with the number the manufacturer gives you.

In a way, that's conservative. Bass response at higher sound levels will always be better than at low power levels. If you plan for bass response at the low level inductance, it will get better at louder volumes. It also depends on how much is enough. Guitars in standard tuning don't go under 80Hz even if slightly mistuned. If you're going to play bass through it, you need another octave, and so twice the primary inductance at a given impedance level.

That "impedance level" gets tricky too. The bass cutoff on a transformer gets better as you drive if from a lower impedance. Bass rolloff is determined by the ratio of primary inductive impedance ( Xl = 2* pi * F * L) to the paralleled driving impedance and reflected secondary impedance. Power tubes not used as triodes are "high" plate impedance, so the bass cutoff happens when Xl = reflected load. You can finesse this a little if you can drive the primary with a lower impedance, like triode mode for output tubes. I once got a 60Hz low cutoff from a nominally 300Hz/600 ohm coupling transformer by driving it from an opamp through a 10 ohm resistor.

I guess that if you can't get hands (and instruments) onto an OT to do your own measurements at higher levels, about all you can do is take the manufacturer's word about Lpri as being a conservatively low inductance, and if it gets you to a low enough bass cutoff point, then it's going to be good enough and get better as you drive it harder.

[nuke]

Do you happen to recall which era of Vox AC30 OT you used? Haddon, Albion, Woden? Just curious...

I don't recall, as the last time I saw that amp was over 20 years ago. It needed a lot of work, plenty of people had been in it before. I'm not 100% sure it was a Mercury, just going off memory, and I think back then, there was probably just one model from Mercury. The others were probably Korg/Marshall (which I think this is) and maybe Hammond.

I do remember that amp though. Made me remember why I both hate vintage AC30's and love vintage AC30's. They're a terrible PITA to repair and built to fall apart, like a Lucas electrical system in a Jaguar of the same era. But when they work right, they just sparkle like nothing else.

[Helmholtz]

I already described a well working DIY method to determine Lpri at operating conditions somewhere above.

A practical example:
Original 50W Marshall Drake 784-139 OT.
Lpri @
- 1V/1kHz (LCR meter): 3.4H
- 1V/100Hz (LCR meter: 4.3H
- 230V/50Hz (impedance method): 94H
- 270V/50Hz (impedance method): 98H
- 460V/50Hz (impedance method): 50H

Lleak @1V/1kHz: 11mH.
Now, what's the relevant Lpri/Lleak ratio?

[R.G.]

I would approach it from the "how much is enough" direction.
I know I am the one that brought up Lpri/Lleak as a figure of merit, but that came from the hifi world where sheer bandwidth is the goal. Guitars have more modest needs, especially on the bass end.
Lpri is what sets bass rolloff. Lleak is mostly not involved there; it's a high frequency thing.

For a 50W amp, probably with a pair of 6L6-ish tubes, the common plate to plate impedance is 4400 ohms. For a -3dB point of 80Hz the inductance needed will be around
L = 4400/(2*pi*80Hz) = 8.76H. So the 1V inductance readings are only mildly anemic. Hundreds of henries is way, way sufficient.
Primary inductance is determined by the core dimensions and by number of turns; a primary with no interleaving and the same number of turns on a given core will have the same primary inductance as one that is many-interleaved. Not-interleaved primaries will have high frequency losses (and crossover issues, parenthetically) compared to a well-interleaved winding.

Maybe the thing to do is to calculate how many henries you need ( by L = Zpri/ (2*pi*F)), divide by two as a hard cutoff to take into account the low-level inductance losses, and then look for the lowest leakage spec or the highest Lpri/Lleak.

[Helmholtz]

I've seen at least one european OT manufacturer who specifies Lpri at 240V/50Hz.
E.g.: https://www.tubetown.net/ttstore/de/tub ... jtm45.html
Here's even a Hammond spec. sheet specifying Lpri at 240V/50Hz :
https://www.tubetown.net/ttstore/de/ham ... -watt.html
I think this is more useful than 1V/1kHz data.