Edcor trafo preliminary tests

I just ordered some WSM600:600 from Edcor and ran my usual tests. .

First impressions: The transformer is small, considering it's only 0.5Watt . It fits on a 1u rack chassis though, so that's nice.

I connected the primary to a signal gen, and the secondary to an oscilloscope. No load resistor.

1. With the signal gen at 10Hz, the trafo shows some distortion in the negative half of the waveform. Raising the signal gen to 15Hz made the distortion disappear. 20Hz is fine.

2. Running sine waves, I noticed that the output waveform started increasing (getting bigger) compared to the input waveform around 10Khz. It continued to do this until 20Khz, then the output waveform started getting smaller again as the frequencies got higher, then started increasing again (forgot to note... maybe around 50Khz). Also at higher frequencies above 20Khz, the phase shift was so severe that on the oscilloscope, it looks like the phase got inverted.

This behavior is unusual from all the other brands of trafo I tried. What do you guys think of it?

I haven't done listening tests yet.... just want to post my first impressions.

> 20Hz is fine.

At what level???

> output waveform started increasing (getting bigger) compared to the input waveform around 10Khz. It continued to do this until 20Khz, then the output waveform started getting smaller again as the frequencies got higher, then started increasing again (forgot to note... maybe around 50Khz). Also at higher frequencies above 20Khz, the phase shift was so severe that on the oscilloscope, it looks like the phase got inverted. This behavior is unusual...

The peak at the top of the audio band, and phase shift, is quite normal, especially with low-Z drive and hi-Z load. A second peak at 50KHz is odd, but probably meaningless.

Yeah, the tests aren't really meaningful unless one of the windings is terminated in its rated impedance... and the source impedance and drive level must be specified.

I'm assuming your scope has an input impedance of 1M shunted by, say, 20pF plus the capacitance of your probe lead. What's the output impedance of your signal generator? Try this: use an X10 scope probe (with compensation trimmer properly adjusted) and a signal generator having a 600-ohm output impedance, or add resistance to the output of the signal generator to build its output Z out to 600 ohms. OR, if your signal generator has an output Z much lower than 600 ohms (50 ohms is common for lab equipment), skip the "build-out" resistor and strap a 600-ohm resistor across the secondary, in parallel with the X10 scope probe.

The idea of using the X10 probe is to isolate the transformer secondary from excessive additional capacitance that might be resonating the winding and giving you a false peak at some high frequency.

Check with frequency sweeps at low level, nominal level (presumably +4dBM), maximum level (+27dBM), and if you like, various points in between. Chart the results.

My signal gen has an output impedance of 600ohms.

Sorry, it's not peak at 50Khz... (I should write these things down instead of remembering them).

Dip at 70Khz, after a peak around 33Khz, then 70Khz and above it climbs up again.

15Hz at 6.7Vrms input sine signal looks okay. (sorry, 6.7VRms at 15Hz is as high as my signal gen can go... can't output any higher than that).

10Hz at 6.7Vrms input sine looks a little distorted... lowering to about 3.5Vrms the output looks okay.

.... did more note taking...

1Khz
-----
Vin = 7.25 Vrms
Vout = 7.24 Vrms

10Khz
-----
Vin = 7.44 Vrms
Vout = 8.27 Vrms

15Khz
-------
Vin = 7.23 Vrms
Vout = 9.40 Vrms

20Khz
------
Vin = 6.7 Vrms
Vout = 11.16 Vrms

25Khz
-------
Vin = 5.59 Vrms
Vout = 13.65 Vrms

30Khz
--------
Vin = 3.2 Vrms
Vout = 9.74 Vrms

33Khz
-------
Vin = 2.11 Vrms
Vout = 16.55 Vrms

40Khz
-------
Vin = 4.24 Vrms
Vout = 14.16 Vrms

50Khz
--------
Vin = 6.76 Vrms
Vout = 9.17 Vrms

Of course, anything above 25Khz could be irrelevant. But from 10Khz to 20Khz, the measured output voltage is more than the input. Is this a matter of loading the secondary with a 600ohm resistor for a more linear response?

Since your signal generator has an output Z of 600 ohms, don't terminate the secondary with a 600-ohm resistor. The transformer is already properly terminated by the source Z of your signal generator. But do try it again with a properly-compensated X10 scope probe; you may be surprised by the results.

> My signal gen has an output impedance of 600ohms.

Many "will drive 600Ω", but are not 600Ω output. Test it. When you put 600 ohms across the output, does voltage drop to half?

If, like many "600Ω" sources, the real source impedance is more like 100Ω, then the 33KHz resonance will reduce when you drive it with a proper 600Ω.

The rise at 33KHz will also vanish if you load the secondary. Unloaded, you see resonance. About 10K load may tame it. If both the source and load are 600Ω, the response will be quite tame (and you'll get 3.6V instead of 7.2V).

Distortion is critically dependent on source impedance. If both source and primary resistace were zero, distortion would be zero. Worst-case would normally be source impedance equal to rated impedance, unloaded secondary. But that's no longer common. Check distortion with (roughly) the source impedance you will be using.

BTW: here are the dB:
1Khz 7.25 Vrms 7.24 Vrms -0.01dB
10Khz 7.44 Vrms 8.27 Vrms 0.92dB
15Khz 7.23 Vrms 9.40 Vrms 2.28dB
20Khz 6.7 Vrms 11.16 Vrms 4.43dB
25Khz 5.59 Vrms 13.65 Vrms 7.75dB
30Khz 3.2 Vrms 9.74 Vrms 9.76dB
33Khz 2.11 Vrms 16.55 Vrms 17.9dB
40Khz 4.24 Vrms 14.16 Vrms 10.47dB
50Khz 6.76 Vrms 9.17 Vrms 2.65dB

During my lunch break, I ran my own tests on the same model of transformer (WSM600/600).

Source impedance: 50 ohms
Level: +20dBM
Load impedance: 620 ohms

Results:

Distortion: I didn't have a distortion analyzer handy, but did not see any OBVIOUS distortion of the waveform above 8Hz or so.

Insertion loss: 1dB at 1kHz

Frequency response:
Essentially flat from 10Hz to 19kHz
-1dB at 20kHz
-2dB at 30kHz
-3dB at 40kHz
-6dB at 60kHz
and so on, until the slight peak at self-resonance: -4.5dB at 143kHz

Phase shift:
Negligible within the audio band
38 degrees at 30kHz
43 degrees at 40kHz
60 degrees at 55kHz
69 degrees at 75kHz
104 degrees at 143kHz

All in all, it ain't looking bad for $8

PS: It's very illuminating to compare my test results with Owel's. It's not so much about whose test method was "better" as much as it illustrates how critical an audio transformer can be regarding terminating impedances. Even when you allow for manufacturing tolerances between Owel's sample and mine, the behavior of the transformer with and without proper termination is markedly different.

This should answer the question that came up here a while ago, about why transformer makers will usually specify a certain model as "600:600" rather than just "1:1". And it also clarifies why you'll often be told that your transformer-coupled output must be terminated in a certain impedance, even though it may result in lower maximum output voltage compared to running into a higher-Z load.

My Edcor tests were done unterminated. And I think that was the fault. But I do agree that the frequency response from 10Hz(20Hz) to 20Khz is decent for this transformer and I wouldn't hesitate to use this in my projects.

On a related note, I had a post some time back about Sowter transformers oscillating with my preamp at max output.... I terminated the secondaries last night, checked it with the oscilloscope and it is now stable.

So lesson learned here, is terminate transformer secondaries.

...unless the primary is terminated, as in a mic preamp input transformer.

In other words, the primary OR the secondary must be terminated in its rated impedance.

:guinness: :thumb:

[quote author="NewYorkDave"]
In other words, the primary OR the secondary must be terminated in its rated impedance.
[/quote]

Thank You. This is going into my lab notebook.

Tamas