Measuring Transformer Inductance

[13engrsapper]

There are other ways to measure inductance.

For example, you have to figure the inductance is in the ball-park of 10H. So if you use a noise / fn generator make an LC with a 1nF capacitor, you should get a big peak in the 1kHz range. Find the exact peak frequency and then plug that into a calculator like this one:

(Sample)RLC Low-pass Filter Design Tool - Result -

Using the DCR for R, 1n for C, fiddle with L until it matches the measured peak. Then you know L.

you could make a high or low pass with a precision resistor as well.

 

[CJ]

If you use 120 VAC on the primary it will help to "wake up" the core of an old transformer that has not been used in a while.

You can see this by using the milliamp meter, the current will gradually come down a bit over time with 120 applied. If you try to measure a dead core with an LCR meter you might not get an accurate reading.

Some winders will do this to break in a new transformer before they ship it.

 

[trobbins]

CJ, that seems to imply heating up the winding so that DCR increases, and/or ensuring the BH curve is centred on the origin (but I guess only if the 120Vac is wound down at the end with a variac) - although the OPT is for SE use so some off-normal BH may be default.

 

[CJ]

It gets the domains moving in the core steel. It is a strange phenomenon .

I have had old transformers on the bench that measured very low inductance. And after a while they just snap back to life. Magnetic aging process gets reversed.
.

 

[andia]

I am playing around with my new lcr meter (gwinstek lcr 916) and I am wondering what mode is to use for inductance measurement of transformer windings. there is an Lserial and an Lparallel mode. Which is the right one?
thanks.

 

[ccaudle]

I am playing around with my new lcr meter (gwinstek lcr 916) and I am wondering what mode is to use for inductance measurement of transformer windings. there is an Lserial and an Lparallel mode. Which is the right one?
thanks.

Have you tried the auto mode to see what it will pick? If you are measuring an isolated transformer winding then series mode should be what you need.

 

[andia]

Have you tried the auto mode to see what it will pick? If you are measuring an isolated transformer winding then series mode should be what you need.

thanks, i just now saw, that it jumps to Ls in automode, if I get it right.

 

[trobbins]

It appears that the gwinstek lcr 916 measures with a 0.6Vrms test signal - but the datasheet is a bit ambiguous. 0.6Vrms is likely higher than other handheld LCRs (mine measures at about 0.23V), which is good imho, but is still quite low wrt nominal operation of the transformer. You may want to inspect the values at a few test frequencies. Also note that your stated accuracy will vary with Q (ie. D, series resistance), so that needs to be appreciated.

 

[ach]

If you use 120 VAC on the primary it will help to "wake up" the core of an old transformer that has not been used in a while.

You can see this by using the milliamp meter, the current will gradually come down a bit over time with 120 applied. If you try to measure a dead core with an LCR meter you might not get an accurate reading.

Some winders will do this to break in a new transformer before they ship it.

This is really interesting. I'd love to hear more about this approach. Do you use a variac or anything to ramp-up?

 

[dougsta]

Gave the CJ method a try, also compared it to a low cost tester.

I'm on 240vac here so I used a small mains isolation transformer (as used in UK shaver sockets, has a 0-120-240vac secondary, 85mA).
Black meter is reading the Vac at input to DUT, blue meter is set to mA range

Reading 132vac on the secondary 120 tap.

So, using Pri Inductance = Vac / 6.28 x f (Hz) x I (A), I get:

132 / (6.26x50x0.038) = 11H

On the mini tester it reads 7.8H

Who wins...?

The manufacturer measures it at 10H, so a high 5 to CJ

https://primarywindings.com/product/pwse10w5k0g/

 

[dougsta]

Time to test a few more bits n bobs...

I used the Vac in as a reading while under the test load and not the before test, any reason to use the unloaded Vac?

A small Hammond 4H 90mA DC choke: AC test = 4.9H and mini tester = 3.6H
A push pull OT from a Jet City amp: AC test = 136H and mini tester = 'unknown component'
A 'home brew' push pull OT using microwave oven transformer laminations: AC test = 105H and mini tester = 'unknown component'
A small push pull OT from and old 1950s amp: AC test = 30H and mini tester = 20H

Another win for CJ

 

[trobbins]

dougsta, any measured value has to be interpreted for the application of the part - the value is only an indicator - and may actually be misleading depending on how the part will be used.

If a part is used in an application where 132Vac at mains frequency is across the winding, and with no dc (or dc imbalance as can easily occur in a PP output stage), then the measured value stands. However I've measured a 25% drop of primary PP inductance in a 15W hi-fi output transformer when the dc imbalance was 46mA on one side and 54mA on the other (ie. 4mA of DC current through the primary).

As another example, if the choke is sandwiched between filter caps (as in a CLC) filter, and has dc current through it then the working inductance may well be quite a lot lower than measured using 132Vac, as the choke likely has a low AC voltage across it.

 

[dougsta]

dougsta, any measured value has to be interpreted for the application of the part - the value is only an indicator - and may actually be misleading depending on how the part will be used.

If a part is used in an application where 132Vac at mains frequency is across the winding, and with no dc (or dc imbalance as can easily occur in a PP output stage), then the measured value stands. However I've measured a 25% drop of primary PP inductance in a 15W hi-fi output transformer when the dc imbalance was 46mA on one side and 54mA on the other (ie. 4mA of DC current through the primary).

As another example, if the choke is sandwiched between filter caps (as in a CLC) filter, and has dc current through it then the working inductance may well be quite a lot lower than measured using 132Vac, as the choke likely has a low AC voltage across it.

Thanks Timbo, I might try an old 100W solid state amp fed with a sine wave from an old android phone, measure the Vac in and I at various frequencies.

Doug

 

[dougsta]

Just tried it with an amplified sine wave from an android phone, kept the output of the 100W amp well below max just in case.

I used the 10W SE OT from Primary Windings.
Result at 440Hz

Vrms	I(A)	Frequency (Hz)		Inductance H
16.1	0.0007	440		8.323682687

Scope photo shows the input to the power amp on ch1 top trace 0.2V / div. Lower trace is the output from the power amp 10V / div.

 

[dougsta]

Tested same OT with a little bit more juice and a 1KHz sine

Vrms	I(A)	Freq (Hz)		H
20.65	0.00089	440		8.396875793
29.7	0.00058	1000		8.153964419

 

[trobbins]

Spreadsheets are great, but perhaps better to realize (or state) your measurement accuracy or uncertainty for Vrms and I(A) and then curtail your result's resolution, as they may turn out to be effectively the same (within error).

If your measured resolution is not too good then you may not easily identify a change in inductance with excitation voltage (for a constant frequency). That change can indicate how much the low frequency stability can be affected if global feedback is used around the OPT.

 

[dougsta]

Spreadsheets are great, but perhaps better to realize (or state) your measurement accuracy or uncertainty for Vrms and I(A) and then curtail your result's resolution, as they may turn out to be effectively the same (within error).

If your measured resolution is not too good then you may not easily identify a change in inductance with excitation voltage (for a constant frequency). That change can indicate how much the low frequency stability can be affected if global feedback is used around the OPT.

Tim, I was a bit lazy with the formatting of that sheet, tidied the calculated H field now to show 2 decimal places.


AC Volts measured on a true rms digital multimeter with +/- 1%. AC current draw was measured on the mA setting on 2nd true rms digital multimeter with +/- 1.5%. I didn't eyeball the sine Vp-p on the scope, esp at 10V / div.

Measured				Calculated
AC at input to DUT (V)	AC Current Draw (A)	Freq (Hz)		H
20.65	0.00089	440		8.40
29.7	0.00058	1000		8.15

Doug