60 Hz Inductance Test

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

CJ

Well-known member
GDIY Supporter
Joined
Jun 3, 2004
Messages
15,652
Location
California
OK, I tried the Per Lundahl method of just using a variac off the power line to measure the inductance of a transformer winding.
It works! But please don't put 100 volts on a mic input tranny like I mentioned.
We did have a little of a language barrier at the show, when Per mentioned using 100 volts to get a good reading, he was probably talking about a big output transformer. It would burn up a mic input as the current would soar as the transformer reached saturation.

I did a graph of inductance vs voltage level. It is neat in that it shows where the transformer has a fairly stable inductance, where it saturates, and where there isn't enough excitation current to get the max inductance.

I checked a Cinemag mic input primary on the Sencore and Gen Rad meters and got about 8 henries on the sencore and ten henries on the Gen Rad. But the Gen Rad measures inductance at 1 khz.
I got about 24 henries with Per's test at 60 hertz, which makes sense, as the core is working better down there.

The most linear region was from about 1 to 4 volts ac rms.
You can see the transformer going into saturation as the voltages increases.
Here is the chart:


50_hz_inductance.jpg


To use this method, and please be careful as you are on the power line, hook up the variac to the winding you want to check. Use an ammeter in series with one of the ac wires going to the transformer. Bring up the voltage very slowly while watching the ammeter. Get about 5 ma going thru the winding and measure the voltage across the primary with an ac volt meter. Keep the current steady when you do this. Take down the voltage and current measurements. Divide the voltage by the current. This is your ohms, or reactance. Then use the formula for inductive reactance to get the henries.
Sample calculation:

Xc=2 Pi F L , Since we are using 60 hertz, we will have a constant of 6.28 times 60 = about 377.

So Xc=377 L for our test. So lets say you got a reading of 5 volts when drawing 5 ma.

5/.005=1000 ohms reactance.

So 1000=2 Pi F L , or 1000=377 L. L equals 1000/377=2.65 henries.

Just keep plotting values til you get some consistent inductances across a set of voltages.

I think I like this method better than any I have tried as it tells you so much more about the transformer than just a henries number.
Thank's Per!
cj
 
Hey CJ
I can see using a variac works pretty well. But why not just
use a good low impedance power amp (solid state) and
a good oscillator. Seems like it would be a little safer.
The oscillator would be better controllable. Most oscillators
have good attenuators on them for a more precise measurement.
Also you could measure at 20Hz if you needed to. Just an idea.

RonL
 
> please be careful as you are on the power line, hook up the variac to the winding

Please put a plain 120V:24V power transformer between your 0V-140V live-line variac and your test rig. (230V:24V in lands where 0V-280V Variacs are used.)

That isolates live line wiring, and also gives you a better range for audio-line input transformers (so you aren't cramped at the bottom of the Variac dial).

120V:6V or even :3V (6VCT) may be better for Mike inputs.

Hi-Z output windings may need more voltage than you can conveniently get from small power trannies. In 120V lands we can use the two 120V windings on some 115/230 dual-primary world-wide power transformers to get an isolated 120V supply. Tube-type power transformers like wall-plug-to-250V may be needed to test big tube output transformers.
 
I was looking at the output specs on my signal generator ant it will put out about 40 volts at 200 ma, plenty to do the above test!
Plus, you can chart inductance all over the audio band also.
Save the variac for the big stuff, as PRR says.
 
Here is an inductance chart I did with the signal generator used in place of the variac. It shows how the inductance really climbs towards the bottom of the audio band. at real high frequencies, like 20,ooo hertz, little or no current flows in the transformer, inductance drops to next to nothing, and frequency takes over as the main component responsible for the reactance.
I double checked it with the variac at 60 hz to make sure there was not something funny going on with the generator's impedance doing weird things, such as resonating with the large henries of the transfromer.
300 mv ac was input to the primary to get these values.
I got 43 henries at 1 khz from the Gen-Rad bridge which exactly matches the number I got from this method.
I have a whole family of curves like this only at many different input levels.



marinair_inductance_excel.jpg
 
OK, I posted a question on inductance at the Magnequest forum and got a response from a kind gentleman regarding the inductance boost at low frequencies:

It's quite common for non-airgapped transformers with "ordinary" steel or nickel lminations.

The inductance is a maximum near saturation. Saturation occurs when the signal level divided by frequency is high, so a low frequency acts like a high level.

Approximately, the inductance increases as the square root of volts per Hz, over a range of about 10:1 in inductance, i.e. 100:1 in volts per Hz. At small signal levels and high frequencies the inductance is more nearly constant.

This seems consistent with what I am seeing as far as inductance changing with level and frequency.
cj
 
Bringing this thread back, as I´m trying to test some transformers.

I think I understand all the process described above, but one thing is still not clear at all...

Let´s say I want to calculate the recomended or optimum source impedance of an unknown transformer for a good audio frequency response. So, using the following formula (I hope I´m going well...):

(6.28 * F * H) - DCR = Zsource

So, if I substitute 20Hz and 169 Henries (the marinair T1452 inductance at 20Hz, as you can see above in CJ´s tests) in this formula, I end with a ridiculous, high impedance number.

Even if I use the 34Henries (T1452 inductance at 2kHz), I end with a source impedance of ridiculous high value, 4k270 Z, assuming winding DCR=0ohms for now...

So, using the inductance test above, how can I come to a nice, realistic inductance value for a transformer winding at audio band? Should I consider the value measured at higher frequencies, where inductance readings starts to flatten into a straight line? Would THIS be the inductance of the winding or at least an aproximated value that would work for optimum source impedance calculations using the formula above?Is the formula I´m using ok for this purpose?

I know I´m doing something really wrong, but I can´t understand what it is. Hope someone comes for my help, as I have been reading many things about testing audio transformers and didn´t came with a simple answer. Althought, I kow transformers are no simple beasts I would like to be able to at least end with a fixed inductance number...

Thansk for all !!!
 
The inductane spec is most impostant at the lower bands to ensure good bass response. You want good results down to 20 hz. If you plug in the numbers at 20 hz for the T1452, you get about 21,000 ohms reactance. Remember, this is a line input transformer that is speced at 10k so 20k is not too unreasonable to believe. The T-1454 mic input has lower henries so as to match up with the mic impedance.
 
> if I substitute 20Hz and 169 Henries (the marinair T1452 inductance at 20Hz, as you can see above in CJ´s tests) in this formula, I end with a ridiculous, high impedance number.

-IF- you assume 20Hz, 169H gives 21K.

The Sowter equivalent is rated 10K:2.4K or 2.5K:600

So what you seem to get is -3dB at 10Hz (about -1dB at 20Hz), which is not unusual for a good line-in tranny.

> I would like to be able to at least end with a fixed inductance number...

You NEVER get any "fixed inductance number" with iron-core transformers. Inductance varies all over the place. But for our purposes, all we need is some minimum inductance. So you have to search for the -lowest- inductance, near the frequency of interest. This usually happens at very-low AC voltage. It rises 20% to 300% at some high AC voltage. When it starts to fall, distortion will usually be too horrible to consider. So you plot from sub-milliVolts up until inductance rises and just starts to drop: we don't care about past there. You note the lowest inductance for response estimates, you don't care how high it may rise, but the Volts where it just starts to drop again is the maximum level you want to work at. And you measure at roughly the frequency of interest, because iron-core inductance will be very different at 1KHz than at 50Hz or 20Hz.
 
Perfectly cleared and simple. thanks. I´m a dumb. It´s really a high impedance transrmer... Duh! Where was my mind?

So, optimum source impedance should be 1/10 of this value, aprox.

Thanks for reorganmizing my thoughts. Now, back to my transformer testing jig! :green:
 
Back
Top