EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: TimFox on January 09, 2019, 05:03:32 pm
-
I have been using a DE-5000 LCR meter for several years with good results, building several test fixtures for four-terminal measurements to binding posts or Kelvin clips. I included shorting switches (multiple poles paralleled of miniature toggle switches) to facilitate calibration.
Today, I was measuring the usual magnetizing and leakage inductances on the primary and secondary windings of a 4-winding 50 VA toroidal isolation transformer. With all the other wires open circuited, I measured the parallel inductance between the three taps on one primary winding and obtained reasonable values: 0.696 H from "100" to "120V", 10.35 H from "0" to "100", and 13.82 H from "0" to "120". Each of these parallel-mode measurements at 120 Hz had Q from 1.0 to 1.25, so the series and parallel inductance values are different.
However, I had a previously-unseen problem measuring one of the secondary windings. I could get a reasonable value of 16.7 H, Q = 1.24 in parallel-mode, but the display alternated between a good reading and several cycles of "-----". In series-mode, it only showed "-----".
The ratio of the secondary inductance to the full primary inductance agrees with the square of the ratio of no-load voltages.
When shorting all the other leads together for leakage-inductance measurements, the secondary winding gave a series of 1.81 mH with ESR = 42 ohms, comparable to the DC resistance measured separately, consistent with the tight coupling expected of a toroid.
At 16 H, the DE-5000 is nowhere near the maximum inductance value of 2000 H.
Could this be hum magnetic field pickup from the test bench environment? If so, why only on the secondary winding?
-
I don't own DE-5000 but transformers are sometimes tricky to measure . As they have hysteresis loses in core + parasitic capacitance in the windings and maybe other complex stuff when more windings are present , the measurement signal could be distorted somewhat to fool the bridge ( maybe only cheap ones ) .
-
I tried it several times, and always got good readings on the 13.8 H primary, but the strange readings on the 16 H secondary. I only tried 100 and 120 Hz frequency settings, but got the same results.
By the way, the open-circuit voltage from the DE-5000 is less than 1 V rms, which shouldn't saturate a 120 V transformer. I haven't checked for any DC component, however.
-
There is a small DC component, I don't remember the exact value, but IIRC, I did write a message in some topic in this forum. Order of magnitude was in the 1-2 V range or something as low. Enough to keep the signal always positive.
Ah, here, the 2nd last paragraph:
https://www.eevblog.com/forum/testgear/der-ee-de-5000-unboxing-and-teardown/msg1568356/#msg1568356 (https://www.eevblog.com/forum/testgear/der-ee-de-5000-unboxing-and-teardown/msg1568356/#msg1568356)
-
I’ll look at the DC voltage issue. From your post, it appears that the DC is only common-mode, but perhaps there is an imbalance giving a small differential DC voltage across the part-under-test. Toroidal transformers saturate easily on DC, but the manufacturer did not specify any DC limits. Does anyone recognize the “——-“ readout alternating with a believable value?
Next, I’ll try the other frequencies, but the transformer is probably useless above 1 or 10 kHz.
-
Curioser and curioser.
Measuring the secondary winding again with the others open-circuit.
The 120Hz measurements repeated the situation above, with maybe 3 or 4 “——“ readings for each valid measurement on parallel inductance and only “——“ on series.
Changed to 100Hz: stable readings. Series 11.19H, Q=1.205. Parallel 18.97H, Q=1.207.
1000Hz: stable. Series 2.32H, Q=1.43. Parallel 3.44H, Q=1.44 (looks like the transformer capacitance is affecting the apparent inductance).
Now it gets strange. I cycled the “LCR Auto” button through the choices. Auto didn’t work for me, but when I got back to “L”, I got stable readings at 120Hz:
Series 10.2H, Q=1.43. Parallel 15.0H, Q=1.43. The difference between the stable 15.0H now and the pulsating 16.7H before is probably settling time.
I was looking for a way to defeat auto-ranging, suspecting range hunting, but there seems to be no manual range function.
Before this, the only “legitimate” problem I’ve had with the DE5000 is that the unit treats all reactance as positive, so when measuring a device (such as a wire wound resistor) that is inductive at low frequencies but capacitive at high frequencies, the indication is ambiguous. For that, one must look at the phase angle, which seems to be the only variable that can change sign.
-
Final update: although the symptoms at 120 Hz persisted for two days, after the cycling of “LCR Auto” that I reported yesterday, the problem has vanished.
Today, with the same setup (transformer and meter have not moved), I get repeatable results for Ls and Lp at both 100 and 120 Hz on both the primary and secondary windings, with roughly 20 sec settling time after changing the connections or frequency. Since Ls and Lp are calculated from the same data, there is no settling time when switching from series to parallel.
I can only hypothesize that something made that auto-range hunt at 120 Hz, and that went away when I cycled “LCR Auto” button.