Weird square wave when measuring transformer

[JohnRoberts]

Try not to be blinded by all the science.

To be comfortable making measurements you need to know your test bench measurement noise/resolution floor. Namely measure how clean your source signal is.

Inexpensive digital conversion of difficult analog signals can result in spurious results.

JR

 

[rubinstu]

You might find it interesting take the output of your transformer (the squarish signal) as shove it back into the mic input of your PC (making sure you don't overload it). Make a recording and run it through a spectrum analyzer (i.e. Fourier transformation). You should see a big spike at your fundamental frequency, then smaller spikes at the odd harmonics. A true, perfect square wave would have smaller and smaller harmonics ad infinitum. Of course, your sampling frequency is probably 44 KHz, so you'll only see artifacts to 22 KHz. For extra credit, do the same thing with different shaped signals - triangles (should see only even harmonics), sawtooth, etc.
Then if you really want to see cool stuff, get some samples of instruments playing single tones and see what you get in the spectrum analysis. (Spoiler alert: it's lots of weird harmonics!)
If memory serves me correctly, a violin is very close to pure even harmonics.

 

[Jen]

It is a 2.2K resistor, which would give a proper 10dB attenuation. So the secondary load is 3.2K. Give that a try.

 

[Jen]

This whole thread shows the importance of using real devices to create test signals, and not some phone app, a "pad" or laptop utility. Sine and square wave generators that actually produce functional calibrated test signals with low impedance source can be purchased for very little money these days. One might cheat on a scope if you can tolerate exceptionally poor bandwidth and range of input attenuation, but test signal source should be left to devices designed to do that job. With new signal generators available in the $75 region, there is no reason to be burdened with the uncertainty of questionable signal sources.

 

[Khron]

This whole thread shows the importance of using real devices to create test signals, and not some phone app, a "pad" or laptop utility. Sine and square wave generators that actually produce functional calibrated test signals with low impedance source can be purchased for very little money these days. One might cheat on a scope if you can tolerate exceptionally poor bandwidth and range of input attenuation, but test signal source should be left to devices designed to do that job. With new signal generators available in the $75 region, there is no reason to be burdened with the uncertainty of questionable signal sources.

And/or it helps to be aware of the parameters and limitations of one's already existing equipment.

That wonky "squarewave" at the output of the tested transformer would've been a lot less alarming, had the INPUT signal been also viewed on an oscilloscope (even if not simultaneously; although that would have conveniently showed any input-to-output differences, with the signals appropriately scaled).

Yes, using the "proper" tools can and does lead to more reliable (and especially repeatable) results, but even with lesser tools (or especially in those cases), it's very much a case of "it's how you use it". Even with $50k worth of test gear, you can still screw up measurements and end up chasing red herrings down rabbit holes...

 

[Brian Roth]

My go-to I built 35 years ago.

Bri

 

[Mr. Moscode]

My Eico 427 scope. Photo not available.

 

[Matt Syson]

When testing a transformer you must also consider it's likely SOURCE impedance that will normally be fed into the transformer, meaning (usually mics) that can have 'source' impedance as seen by the transformer of anything from say 50 to several hundred Ohms. THIS will affect the way a tranformer behaves with real signals. Test gear usually has the ability to alter the 'output' impedance to 50/150/600 Ohms as each setting will change the overall performance not least of which the load of the tranbsformers secondary will alter the total load now placed on the signal generator.