Inductance of an audio transformer?

[anotherjim]

I'd be grateful if anyone can confirm my thinking on this.

I have a small audio transformer (Eagle LT44 or P631M). I can't find mention of it's inductance anywhere.

What I do know is it's specified impedance, which is 10k on the primary.

Can I assume that the impedance specified is at 1KHz?

In which case inductance should be 1.59mH. 1592mH

Cheers
Jim

[R.G.]

I have a small audio transformer (Eagle LT44 or P631M). I can't find mention of it's inductance anywhere.
What I do know is it's specified impedance, which is 10k on the primary.
Can I assume that the impedance specified is at 1KHz?
In which case inductance should be 1.59mH. 1592mH

Close, but not quite. You have to use the specified lowest frequency, not a midband 1kHz.

The primary inductance of a transformer is the determiner of its low frequency response. For a given low frequency requirement F, the reflected secondary load of Rsr, the inductance must have an impedance Zl greater than Rsr at F, or:

Zl = 2*pi*F*L = Rsr


which gives
L = Rsr/(2*pi*F)

I didn't find the spec for low frequency response on this transformer, but such trannies are usually 300Hz, so if that's true, for the 10K reflected impedance,

L = 10k/(2*pi*300) = 5.34H   

Actually, it's got to be AT LEAST that for a -3db response there. It may be larger and still meet spec, which is why they don't specify inductance.

And this is a lesson in why high-impedance transformers are tough to wind - you have to get five bleeping henries out of that tiny core and windings!

[anotherjim]

That was quick.
Many thanks R.G.
Just knowing that they usually use 300Hz for impedance spec will be a big help.

Cheers

[digi2t]

jim, if you can hang on till tonight, I have an LT44 at home, and an inductance/capacitance meter. I'll measure it up for you.

Your calculation does sound about right though, I used two of them to replicate the 8H TDK inductors that are found in the Honey Special Fuzz, and they worked fine.

[anotherjim]

Digi', thanks, that would be great 

[digi2t]

Well, it turns out that I have three of the blue ones. There are two versions of LT44's, yellow, and blue. Here is the info I found regarding the differences;

LT44 - Yellow
The primary side has a nominal impedance of 20K Ohms. The secondary side is nominally 2k Ohms and has a centre tap. The D.C. resistance of the primary is approximately 480 Ohms, and that of the secondary is 107 Ohms. The theoretical voltage and turns ratio is 3.2 and something close to this is seen when testing at 1kHz with the appropriate input and output impedances connected.  These have input and output impedances that are more useful for making ring modulators than the LT700s.

LT44 / P631M Blue 
This is the same as the one with the yellow tape but has the more usual blue winding tape and the untapped primary side has a higher d.c. resistance at about 1.08K.  If you're looking for low internal resistances to minimise losses, you would prefer the yellow type.

I measured all three, and I've noted the ranges for all my measurements;



I used the blue ones in my Ring Stinger, and the ring mod sounds good to me. Don't think using the yellow ones will make any difference there.

[R.G.]

As a side note, it is common for audio transformers to spec themselves as not -3db, or -6db at the low end, but instead at the frequency of -1db or -0.5db response.

These result in 2x and 4X the primary inductance needed for a given reflected secondary load. The numbers digi is measuring make sense for that. Also, the calculations for -3db are minimal - generally transformers will be designed to have some tolerance on the Lp as well, so that every unit, even the loosely stacked ones, will make the minimum frequency response.

[anotherjim]

Digi - many thanks for your help.

Interesting. Take your measured value for primary inductance (call it 16H) and the manufacturers Impedance (10k) and you have their test frequency = 100Hz.
For the total secondary, call the measured 1H and with the quoted impedance of 2K I get 320Hz.

So, I take from this that a "fair" idea of the inductance can be had by assuming somewhere 100 to 300 Hz as the spec frequency give or take tolerances.

I want to use the transformer as the output stage of a DI/amp sim' I'm "designing". I need another inverting stage to correct polarity plus an opportunity to have an isolated output and the transformer can do both jobs. I don't care about losses as the amp will hit 8volts p-p so I've got signal to lose.
I wanted to know the effect of the inductance in combination with a coupling cap and the amps output impedance. With the values I now have, I've got a low pass cut off about 40Hz, which will do me nicely I think. It should have a slight resonant peak just below that figure too.

[tubegeek]

I need another inverting stage to correct polarity

Can't you just reverse the relevant transformer winding to (from?) that section of the circuit? That'll reverse your polarity without additional components.

[anotherjim]

Tubegeek, that's what I meant. I'm using the transformer as an inverter and isolator. I intend to use it to correct the polarity following 3 inverting stages. Breadboard sound seems good, so it's pretty much build time.

[tubegeek]

Got it - the word "stage" threw me, I guess.

[PRR]

> common for audio transformers to spec themselves as not -3db, or -6db at the low end, but instead at the frequency of -1db or -0.5db response.

IME, this is true for "telephone" (modem) transformers. A telephone connection is hundreds of links of a chain, and must have known-gain/loss within a few dB overall. The specified low-limit will be fairly high (300Hz voice, 600hz data) and the loss there will be quite small (as you say).

If you accept -3dB you can often go well down in the bass. (The treble is often even better than that; they just don't measure it for telco use with sharp 4KHz cut-offs.)

The Argonne transformers (all those little things now made by anonymous asians) were designed for speech and low-fi music (pocket radios). Between rated impedances it is about 300Hz at several dB down (not really specified). This can be extended (as you know) by driving with very low Z, though the overload point is falling-off pretty quick (and wasn't that high to start with).

When measuring "inductance" on laminated iron cores, watch out. The effective inductance *changes* with frequency, and often a significant step around 400Hz. This is where eddy-currents start to throw the flux out of the iron. By 800Hz maybe only a quarter of the iron is realy working, and inductance has fallen-off about that much. So if you want a 100hz filter, measure at 100Hz (and an octave each way); if you want a 2KHz filter then measure at 2KHz (and explore a bit either way for oddities).