Measuring Inductance

As I await my samples of the new Edcor "X" series transformers, I've sketched up an idea for a bridge to measure their inductance with direct current applied. Comments and suggestions are most welcome.

sketch

As drawn, the DC current control will be hyper-twitchy. Put some resistance in that emitter. Switch it to give several ranges. I suggest trying about 0.2V to 2V resistor drop. SO a 2mA to 20mA range would be 100Ω, 20mA to 200mA would be 10Ω.

I'm not sure why a selection of ratios is useful. Or the calibrated variable resistor. Use 50% ratio and two clips where you can try all the resistors in your stock-box. Most of us have resistors in 5% accuracy every 20% up and down from 10Ω to 1Meg.

20Hz or 50Hz is useful, but also check at 100Hz and 400Hz. Iron inductance will sometimes drop there. This does not usually hurt bass response because inductance drops slower than frequency rises, but would be critical for midband calculations.

Signal level is VERY important.

I use my LMS card to plot inductance/resistance across the audio band ... and up to the limits of the card
100k I think.
and at the max current the card can handle.
relatively high for line level work but low for passive xover work.
I've never experienced a xover anomaly due to current/power ... apart from outright failure.

:shock:
wot was the question ???

oh
yeah ... I use the computer

Kev, the problem at hand was the measurement of inductance with a direct current flowing through the coil. I already have instruments to measure inductance with no DC.

PRR, what I really had in mind for the "calibrated variable resistor" was a decade box. But I just drew it as a rheostat 'cause I'm lazy :wink: I grant you that the ratio control is an unnecessary frill in that case, although it would be handy if you were actually using a pot. For the signal source, I planned on an external sine generator tunable across the band. My old tube HP, with its ballsy output capability, should do nicely.

Thanks for the suggestion on the current range control.

You can hook up the meter to the pri, then put some dc into the sec, the core does not know the diff.

allow for turns ratios

use hv dc and high res so you don't load it

CJ, the transistor current source takes the place of a crude resistive current source--which means you don't have to use high voltage. That's a great idea about energizing the secondary... but since the secondary is usually spaced further from the core, is it really "all the same" as far as the core is concerned?

[quote author="NewYorkDave"]Kev, the problem at hand was the measurement of inductance with a direct current flowing through the coil. I already have instruments to measure inductance with no DC.[/quote]

yeah I get the point
:roll:

I don't think my LMS card and software would be compfy with a steady DC flow while measuring.

as PRR said ... your circuit could be twitchy and may be a little hard to calibrate.
It is likely to be heavily reliant on the calibrated variable resistors.
what accuracy are you shooting for ... or is it just for comparison ?

must think on this more

the sec. (which winding?) can be anywhere, depending on the winding structure.

do it backwards to verify the non interaction of spacing.

want a really cheese ball way of doing it?
put a diode and a resistor across the secondary.

> meter to the pri, then put some dc into the sec

True and a useful simplification. I'd assumed it would be useful to have a gizmo that could measure just one coil, for chokes, for wacky-Z transfos, or getting a take on a failed transformer.

Also, shorting any secondary tells you the leakage inductance, often a handy number to know. (Though this rarely depends much on DC current.)

Either way, the hard part is making a DC current with very high AC impedance. Big resistor from a high DC voltage works, with small error that can be subtracted away. But this tends to need a power supply as big as the final project. That may be handy if you already started the project, might be awkward when just going through the junk bin.

> since the secondary is usually spaced further from the core, is it really "all the same" as far as the core is concerned?

Yes.

Feed a strong AC voltage to a low-volt winding on an open transformer, one with a little space between the outside of the coil build and the core. Get a yard of thin wire. Pass it through the space once, measure the induced voltage. Pass it through the space and almost to the other space, a half-turn. You get the same voltage. Pass through the other space, the voltage doubles.

It isn't the number of turns, or the location in the build. It is the number of passes through the core (two per "turn" on common US E-I cores).

Actually, location and proximity to other winding must make some difference, but for iron-core 20Hz 50Hz 60Hz coils, the effect is less than 0.001 or 0.1%. 999 flux lines flow through the core, and thus induce all other windings, for every 1 flux line that flows through air.

For the current limit part I would try the 2 Si diodes Base to + in your circuit picture with a resistor from base to ground for enought current for the diodes and base of the transistor.

This gives about a .6VDC difference between the diodes(1.2VDC) and base emitter voltage drop(.6VDC). Knowing this the math becomes
.6VDC/ emitter resistor for the current wanted. You would pick emitter to B+ resistor values for the currents wanted.

Standard current limit circuit.

Nice circuit Idea Dave. I might build it to test transformers I use in microphones to see what happens with possable phantom resistor mismatches and the microphone output transformer. I like to keep phantom on with transformerless pres to keep the input caps formed even with external powred tube microphones.

you can estimate the efects mathematically also, if you know a few things, like the lam type, approx turns, you get into a catalog that has all the charts and graphs for the engineers who need to know this stuff before they build, they usually have graphs that show what happens with varoius levels of flux, dc, all that good stuff.

cj

I've been making measurements of a transformer using the circuit I posted earlier (with an emitter resistor added as per PRR's suggestion). One thing I noticed is that the inductance value plummets as I start drawing DC through the coil, but beyond a certain value of current, it starts to level off. Is this normal, or should I be suspicious of my test setup?

thats normal. a better way to do this would be to put a sine wave into the input, which is kept constant at 1 volt p-p. then measure the output voltage p-p as you add dc. this will give you a real world idea of what is goint to happen to your signal as dc is added. see that it will vary with requency. 20 k cps does not care about dc, because the core is no longer inside the coil at this point.

I posted the results in the "Class X transformers from Edcor" thread on the Lab. Thanks for the tips, fellas!

I can't help but wonder, though: is it a safe assumption that the collector of the transistor is presenting a high impedance as I vary the current through it? Looking at the dismal results of the test, I have to wonder if some shunting effect is involved.

[quote author="NewYorkDave"]I posted the results in the "Class X transformers from Edcor" thread on the Lab. Thanks for the tips, fellas!

I can't help but wonder, though: is it a safe assumption that the collector of the transistor is presenting a high impedance as I vary the current through it? Looking at the dismal results of the test, I have to wonder if some shunting effect is involved.[/quote]

Measure it. Vary the voltage on the collector and see what the delta I is.

With no emitter degeneration, power transistor collectors aren't terribly high Z, and the Z is reciprocally dependent on current. But I would still expect it to be high relative to d.c.r.'s of most inductors.

When you add emitter R the output Z goes up: Z out (new) = Zout (old) x ( 1 + ReGm). This holds until you get to up into the collector-base resistance zone, which will probably be a few hundred k for a power transistor, several meg for a garden-variety small signal.

Beware of thermal shifts if your transistor is heating up, which in this circuit will cause the collector current to rise due to negative Vbe mag. tempco and positive beta tempco.

A more elaborate current regulator could be contrived, with an error amp sensing the drop across the emitter resistor. Use a P-channel FET and it could be very accurate indeed, since there will be no loss due to base current.

It was a 2N4403 with a 100-ohm emitter resistor, connected as in the sketch I posted earlier, +12VDC supply, with a 100K pot at the base and a 33K fixed resistor between the pot and common (giving a collector current range of 0 to 40mA). As long as the impedance looking back into the collector stayed over 100K, I'm golden. :thumb:

With that high an impedance in the base, the current source will be quite a bit lower Z. My formula was assuming a voltage source between the base and the other end of the emitter resistor.

Circuitmaker says about 5.8k at maximum (~43mA). This rises to about 73k with a much stiffer (x100) divider. Your impedance may vary.

What if I bypassed the base to common (at AC) using a large cap? Or used a 10K pot and a 10K resistor, which would give about the same range of collector current?

Hmmm... I don't trust simulators, but my "free" version of Circuitmaker sez that simply bypassing the base with 1000uF will raise the collector impedance at 40mA to over 250K.

I'm glad I asked; I might have to re-do these measurements.

The cap will do about that according to my full version sim.

Note also that the impedance is not constant with current.

Thanks for the warning about the high impedance in the base. Lowering the impedance at the base did make a difference on the bench. I made BOTH of the changes described above--adding the bypass cap, and changing the base voltage divider to 10Kfixed/10K pot. I ran the test again (and updated the results in the other thread), and you can see the difference at the lower end of the DC current scale, where the fairly high inductance of the DUT was being shunted by the collector impedance and giving a false low reading.