Pultec Inductor Measurement

[CJ]

A lot of people seem to be doing eq's right now, so I thought I would start a thread on measuring the inductors.
I think it is best to run the inductor at it's desiginated frequency to get a good measurement. I usually use this simple resistor-inductor technique:

So lets say you wanted to measure the 108 millihenry inductor for the 1000 hz section of a MEQ-5. Well, at 1000 hz, the inductor should have a reactance of 2 pi F L , so we would have 6.28 x 1000 x 0.108 H.
This comes out to be about 678 ohms. So if we hook up a signal generator and scope to the above circuit, and use a 680 ohm series resistor (close enough), and apply a 10 volt peak to peak signal, we should see 5 volts across the resistor and inductor, right?
What do you think? Is this a legitimate method?

If you wanted to calculate the inductance when you did not get an equal voltage drop, say you only read 1 volt across the inductor, then you would just work backwards and do some algebra.

So 1 volt out of ten would indicate the resistor is responsible for 90 percent of the voltage drop. So if the resistor is taking 90 percent, that means the inductor is taking 10 percent. So the inductor would have to be 10 percent divided by 90 percent, or 1/9 the resistance of the resistor.

So in our above circuit, 1/9 th of 680 ohms is about 75.5 ohms. Plugging this back into the inductive reactance formula, we would have Xl=2 pi F L or 75.5 = 6.28 x 1000 x L. Solving for L gives 75.5/6280 or about 12 millihenries.

If this is too tough for some of you, we could write a simple excel program where you input all the parameters and it would spit out the answer.

This approach assumes a perfect inductor with zero ohms dc resistance, so it won't be perfect.

cj

 

[fum]

Thanks CJ, but where did you get the MEQ5?

:grin:

ju

 

[CJ]

Yard sale! 50 cents!
:razz:

 

[alk509]

[quote author="cjenrick"]This approach assumes a perfect inductor with zero ohms dc resistance, so it won't be perfect.[/quote]

You can always measure the DC resistance with an ohmmeter and do some math...

I like this idea. Keep it coming!

Al.

 

[owel]

I read somewhere that measuring the DC resistance of sensitive mic transformers might be bad? Something about magnetizing the core? Is this true? I don't think a VOM can put out that much voltage when measuring DC resistances?

 

[PRR]

> So lets say you wanted to measure the 108 millihenry inductor for the 1000 hz section of a MEQ-5. Well, at 1000 hz, the inductor should have a reactance of 2 pi F L , so we would have 6.28 x 1000 x 0.108 H. This comes out to be about 678 ohms. So if we hook up a signal generator and scope to the above circuit, and use a 680 ohm series resistor (close enough), and apply a 10 volt peak to peak signal, we should see 5 volts across the resistor and inductor, right? What do you think? Is this a legitimate method?

Uh, not quite. The inductor (assumed perfect) has a complex reactance. When the magnitude of reactance (ignoring phase) equals the resistance, you actually get 7.07V of your 10V input.

You know this from R-C circuits, which are -3dB (0.0707) down when "C=R".

> about 75.5 ohms

In most audio coils expected to give around 680Ω reactive impedance at a given frequency, the copper resistance will be 50 or 100Ω. So a 75Ω test resistance will be badly skewed by copper resistance.

Since you also have significant copper resistance in most audio coils, the cleanest way to measure "just" the inductance is to use a resistor much larger than the expected reactance. In the case above, try 6K8 and expect around 1V across the coil. The error from measuring a complex reactance against a large simple resistance will be under 10%. And if the coil is any good at the test frequency, the error from copper loss should be under 10%. (However iron-loss may confound you.)

Often, and especially when aiming for a band-pass, the best way to find the inductance is to parallel a cap across it to make a tuned circuit, drive with a high impedance, and look for the resonant frequency. Preferably pretty close to the freq you plan to use the tuned circuit at.

Iron-core coils will also show shifts of inductance with signal level. In an EQ, that is usually bad. Still you measure at the intended level, then try other levels to see if the inductance stays pretty constant, or where it goes high/low enough to matter. Any iron-core coil will go to negligible inductance if you give it a super high level. And for some high-H low-level EQ coils, the 25V out of an H-P 200 sig-gen may be too much. (I can show various inductances on the "120VAC" winding of a 25VA power transformer, using my hi-output 200 model.)

 

[CJ]

There is a good discussion on that at the forum where Klaus and Oliver hangout. Anyone have a link? The guy deliberately tried to magnitize the core and then took distortion measurements.

Here is another method you can use to measure your inductor if you have the exact same cap you are going to use in the circuit:

You would run the generator up and down the frequency range that the circuit is supposed to cover and look for the point of resonance. This can be found by connecting a scope or voltmeter across the resistor and looking for the biggest voltage across it. This would indicate the point where the reactance of the cap and resistor are the same, which is where maximum current will flow in a series resonant circuit.You could also use this method to plot the curve of the circuit. Well that is, if you used the same resistance as the Q resistor of the circuit, or use the same dc resistance of the inductor if no Q resistor is used.
cj

 

[clintrubber]

Does anyone happen to have those Edcor PC-type trafo's and knows what the prim.-L is ?

I'm looking for two ~2.45H inductors and since joining the Edcor-grouporder anyway I wanted to use the prim. of a Edcor PC-10k:600 (sec. not-connected).

But since I don't have these trannies yet I don't know yet if these are close enough in value.

(I've asked Edcor themselves but haven't heard anything yet.)

Oh, yes, the use of those inductors is for the nice pickup-sim circuit of Jack Orman: http://www.muzique.com/lab/pickups.htm



Thanks,

Peter

 

[Guest]

[quote author="cjenrick"]I usually use this simple resistor-inductor technique:

[/quote]
May be interesting if varies frequency in around 100 Hz.
You can measure critical frequency of mumetall lamination.
Inductancy is constant bellow and falls above it.
xvlk