Detect damaged turns in the primary of a low power mains transformer

[RoGeorge]

I need a coil in the ballpark of 20mH and 10ohms, for a Chua Diode based Chaotic Attractors Oscillator, something like these:
https://www.eevblog.com/forum/beginners/tek-475a-vs-tek-2445/msg4401871/#msg4401871
https://www.eevblog.com/forum/projects/t20347/msg3622373/#msg3622373
for an ad-hoc weekend project, built for fun with whatever lies around.  Won't go making custom coils or trying to order anything.

The only coil I have that comes closer to 20mH and 10ohms is from a former very small mains transformer, probably a 220V to 5-10V/100-200mA or so, but it has a burnt primary.  Most of the heavily burnt wires were removed, yet some primary turns are still showing traces of burnt wire isolation (not very visible in the pics).  There might still be damaged turns left underneath the visible turns.



I have a 50 ohms DDS and a digital oscilloscope, but no proper LCR instrument, only one of those very small ATmega48 universal identifier/measure BJT/MOS/D/L/C/ESR/R/etc. like the one in the pic:



The secondary coil I want to use seems OK.  It measures 9.2ohms and 24.5mH (with the magnetic circuit closed, not open like in the first picture with the transformer alone), and the wire in the secondary has 0.2mm diameter, but can't see how many turns.

- What losses to expect from such a transformer when in good condition?
- Any idea how to quick check if there are still shorted turns in the primary?

[ledtester]

Have a look at this video on building a "ring tester":

DIY transformer/inductor tester - DiodeGoneWild
https://youtu.be/QBbEYYWiBI8

or you can look at the ring-down on a scope.

[mag_therm]

Using a sharp blade with gloves on, cut across the unwanted primary at 90 degrees to the turns , and peel out the all turns layer by layer.

[RoGeorge]

Have a look at this video on building a "ring tester":

Thanks.  The problem is that I don't know what quality factor (Q, or how long of a ringing) to expect for the primary of a similar known good transformer. 

I suspect, if there are any shorted turns in the primary, then they are not perfect short-circuited turns (like the ones tested in the video).  The burnt turns I've already removed were showing some carbonized wire-isolation with low resistance, but they were not zero ohms shorts.

[EEVblog_Fan]

   You can build or buy a Ring Tester go to this page for details
  (this will also answer most of your transfomer Q and similar stuff:
  http://bobparker.net.au/Basic%20Digital%20Ringer/index.html

[Le_Bassiste]

- Any idea how to quick check if there are still shorted turns in the primary?

a short circuit on the primary side transforms as a short circuit on  the secondary side. this is of course only valid for AC voltages and currents. to test that, simply make a winding of one turn on  the primary side and short it. then measure the secondary winding again again with your LCR meter. if your meter reading isn't changed down to some 10% of your inductance(!) reading w/o the short, then the primary is still short circuiting somewhere.
theory behind it: U2/U1 = n2/n1, I2/I1 = n1/n2, then U2/I2 = n1^2/n^2 * U1/I1.

because applying a sufficiently high voltage on the secondary winding may induce flash-over on the hobbled primary side, i'd suggest to remove the entire primary winding, just for your peace of mind.

[MrAl]

Cut the primary turns off with a hacksaw and eliminate any potential problem.  Once cut they peel off with a little effort.

[RoGeorge]

Indeed the only way to be sure was to remove the primary turns entirely.

The wire was damaged in many places underneath what was visible, looked at it with a 30x on the optical microscope.  Some isolation portions where black carbonized, in other region it was just cracked (with black carbonized thin rings around the wire, pretty strange pattern - have no pics from the microscope, sorry), other places the wire was melted and found at least 3-4 small blobs of melted copper, like a solder blob, but very very small, smaller even than the tip of a ball-pen.

It was nothing to recover or reuse from the primary, and the wire was very, very thin, I've measured about 0.02mm diameter (about AWG52  ).  A liability for further short circuits to develop with time or with humidity.

However, in theory it should be possible to detect short circuits (even faint - high R - shorts) by looking at the phase at 2 different frequencies.  If R of the good coil is the main dissipation mechanism, the R of a good coil remains about constant at very low frequencies (e.g. at 50 and at 100Hz), while the X_L will double from 50 to 100Hz.  This should be possible to measure as a phase shift on a dual channel oscilloscope, something like in the attached LTspice simulation.

Thank you all the advice, left only the secondary section of the transformer, the secondary coil seems fine, no short-circuits in the secondary turns.  Primary turns were completely removed.

[RoGeorge]

It's intriguing how much the L varies with the core, anywhere between 5-50mH  , depending of how tight the E+I core sheets are tighten together.  Never tried this before in practice.

1.41mH - no core
5.52mH - only E sheets (opened magnetic circuit), coil at the outside half of the E sheets of Fe-Si
11.3mH - only E sheets (opened magnetic circuit), coil at the inside half of the E sheets of Fe-Si
24.6mH - E+I closing only one leg of the magnetic circuit, I sitting on top of E (not interleaved)
50.6mH - E+I closed magnetic circuit, I sitting on top of E (not interleaved)
57.5mH - E+I closed magnetic circuit, I sitting on top of E (not interleaved), pressed down by hand

Didn't properly closed the core yet, as it was originally (originally, the E sheets were inserted from both sides, alternatively, interleaved with the I sheets).

[MrAl]

Indeed the only way to be sure was to remove the primary turns entirely.

The wire was damaged in many places underneath what was visible, looked at it with a 30x on the optical microscope.  Some isolation portions where black carbonized, in other region it was just cracked (with black carbonized thin rings around the wire, pretty strange pattern - have no pics from the microscope, sorry), other places the wire was melted and found at least 3-4 small blobs of melted copper, like a solder blob, but very very small, smaller even than the tip of a ball-pen.

It was nothing to recover or reuse from the primary, and the wire was very, very thin, I've measured about 0.02mm diameter (about AWG52  ).  A liability for further short circuits to develop with time or with humidity.

However, in theory it should be possible to detect short circuits (even faint - high R - shorts) by looking at the phase at 2 different frequencies.  If R of the good coil is the main dissipation mechanism, the R of a good coil remains about constant at very low frequencies (e.g. at 50 and at 100Hz), while the X_L will double from 50 to 100Hz.  This should be possible to measure as a phase shift on a dual channel oscilloscope, something like in the attached LTspice simulation.

Thank you all the advice, left only the secondary section of the transformer, the secondary coil seems fine, no short-circuits in the secondary turns.  Primary turns were completely removed.

Yes and with something damaged like that you could also get intermittent shorts which may show up some times and not other times, and one of those times it shows up could be at a completely unexpected time during normal use.