Definitive transformer testing?

[Mark]

I was reading a post on Dumble transformers and voltages and I remembered when Merren talked about transformer testing he recommended to Ken Fisher to test the output transformers Fischer bought from Pacific.

Given that both Dumble and Fischer tested their transformers, perhaps we should know something about testing transformers?

What tests are you aware of and how do you carry them out?

[Stevem]

First off you need the test gear to input a high level of D.C. Into it and then also a pretty good level of AC drive signal at different frequencies all while reading the frequency seen on the secondary side while monitoring the distortion level seen on both sides of the OT.

Do you have such gear?

If you are looking to just prove out the OTs turn ratio and hence impeadance then that's a far far easier thing to do!

[R.G.]

Presumably you want to look at non-destructive testing.

Here's a decent set of non-destructive tests that tell you a lot about the transformers.

1. Wire resistances: accurately measure the DC resistance of each winding. For multi-tap windings, measure each segment.
2. Voltage and turns ratio: ensure that all leads are not connected to any external load, then apply a voltage to one winding. Accurately measure the voltages on each of the other windings or winding segments.
3. Primary inductance: this is the main determiner of the bass response of the transformer. Measuring is complicated by the fact that it's not one number - primary inductance varies depending on the magnetic field inside the transformer, and hence with DC levels, size of the applied signal and so on. What you're after is trying to produce a reasonable estimate of the minimum inductance at expected operating signal levels. This test critically requires that all windings other than the one being tested are open.
4. Leakage inductance: this is the main determiner of the high frequency response of the transformer and is related to some forms of crossover distortion. Measure by picking a winding or winding segment, then short circuiting the leads of ONE other winding segment and measuring the inductance seen on the selected segment. Repeat for every other winding segment that is accessible.
5. High voltage withstanding: Apply an increasing high volltage between the primary and all other windings shorted together. Look for any sign of conduction or arcing.
6. Thermal and power rating: accurately measure the DC resistance of each winding as in 1 above, but only after the transformer has cooled to room temperature for some days (yes, days). Then hook up the transformer in its expected "normal" operating conditions and run it at maximum power, stopping to disconnect it every 30 minutes or so and re-measure the winding resistances, noting time and new resistance. Do this until it stops increasing temperature, or until your measurements produce a good enough temperature rise curve to estimate the final temperature.

Test 1 tells you if the trannie is in fact hooked up the way you think it is. That's not always true.
Test 2 tells you if the thing is going to do the transformation ratios you expect.
Test 3 tells you if the bass response is likely to be adequate.
Test 4 tells you about the high frequency response.
Test 5 tells you if it will live in the high voltage environment you expect. For OTs, it's probaby good enough to test it to 3kV. For AC power line transformers, you probably want to test to 4200Vac.
Test 6 tells you if it will transfer the amount of power you expect it to without burning up.

[martin manning]

I typically look at DCR, voltage/impedance ratio, and phase relationship to verify basic specs and lead identification. WIth an LCR meter, you can measure primary inductance (indicator of low frequency response), and primary interwinding capacitance and leakage inductance (indicators of HF response). You can measure all this stuff and more, but how do you know what you are looking for without hearing it? Could Merren have been referring to a listening test? From other discussions we have heard KF auditioned transformers with varying specs from the same manufacturer, perhaps this is what he was doing? An A-B comparison set up would be easy to arrange, and enable an isolated comparison to a known standard.

[Mark]

Thanks for the replies and advice. Much appreciated.

[Mark]

Presumably you want to look at non-destructive testing.

You're a funny man RG.

Here's a decent set of non-destructive tests that tell you a lot about the transformers.

1. Wire resistances: accurately measure the DC resistance of each winding. For multi-tap windings, measure each segment.

This is measured with an ohmmeter.

2. Voltage and turns ratio: ensure that all leads are not connected to any external load, then apply a voltage to one winding. Accurately measure the voltages on each of the other windings or winding segments.

This would be done with a variac and a true RMS multimeter. I seem to remember trying this test with voltage on the primary and getting a ratio, and then putting a smaller voltage on the secondary and getting a different ratio. Not the reverse/inverse of the first ratio.

3. Primary inductance: this is the main determiner of the bass response of the transformer. Measuring is complicated by the fact that it's not one number - primary inductance varies depending on the magnetic field inside the transformer, and hence with DC levels, size of the applied signal and so on. What you're after is trying to produce a reasonable estimate of the minimum inductance at expected operating signal levels. This test critically requires that all windings other than the one being tested are open.

I assume Martin has answered this one, in that an LCR meter is required.

4. Leakage inductance: this is the main determiner of the high frequency response of the transformer and is related to some forms of crossover distortion. Measure by picking a winding or winding segment, then short circuiting the leads of ONE other winding segment and measuring the inductance seen on the selected segment. Repeat for every other winding segment that is accessible.

So this is perform on the transformer with the LCR meter?

5. High voltage withstanding: Apply an increasing high volltage between the primary and all other windings shorted together. Look for any sign of conduction or arcing.

I assume a Megger or insulation tester will be required for this test.

6. Thermal and power rating: accurately measure the DC resistance of each winding as in 1 above, but only after the transformer has cooled to room temperature for some days (yes, days). Then hook up the transformer in its expected "normal" operating conditions and run it at maximum power, stopping to disconnect it every 30 minutes or so and re-measure the winding resistances, noting time and new resistance. Do this until it stops increasing temperature, or until your measurements produce a good enough temperature rise curve to estimate the final temperature.

Okay, this test has me wondering, I wouldn't have thought the transformer would heat up much after the testing with the variac and the voltmeter. Are you referring to the amp being previously used?

When you mean maximum power are you referring to the amp operating with a sinusoidal waveform, or operating with a squarewave, which will be harder on the transformer?

The temperature would have to be measured with something like a Fluke 561 IR thermometer.

Test 1 tells you if the trannie is in fact hooked up the way you think it is. That's not always true.
Test 2 tells you if the thing is going to do the transformation ratios you expect.
Test 3 tells you if the bass response is likely to be adequate.
Test 4 tells you about the high frequency response.
Test 5 tells you if it will live in the high voltage environment you expect. For OTs, it's probaby good enough to test it to 3kV. For AC power line transformers, you probably want to test to 4200Vac.
Test 6 tells you if it will transfer the amount of power you expect it to without burning up.

Martin, I can't be sure, but I couldn't see Chris Merren telling Ken Fischer to do a listening test, sounds like something Fischer would do himself. I'm not sure it is very definitive test as there are a lot of variables in such a test.

[martin manning]

It's true there are a lot of variables, but if one were looking to compare several transformers with varying electrical specs and or transformers with the same electrical specs form different manufacturers, listening would be the ultimate test. I think I am remembering correctly that KF ordered transformers with a number of differing specs from Pacific. That was from the guy from Pacific who had access to KF's spec cards who was here a few years ago. That says to me that he was comparing the sonic effects of those variables.

Yes, you can make measurements 3 and 4 above using an LCR meter.