Output inductor buzzing in a high power UPS

[Pentium100]

Hi,

I have a Borri B8033 UPS rated at 40kVA. The UPS is working, but the output filter inductor of one phase is buzzing rather loudly, it probably is indicating some kind of problem. The noise is lower when the load is higher.

My LCR meter shows higher inductance for that inductor, compared to the other two phases (it also shows lower Q at 10kHz).

The driver module for the output IGBT shows signs of getting hot (electrolytic capacitors were kinda bad, the board itself is slightly discolored), but I have no idea which came first - something happened to the driver board and that messed up the inductor or the driver board is getting hotter because of the inductor (or it may be not related). The output filter capacitors are OK.

The buzzing inductor measures:

f,Hz    L,uH    Q       ESR
100     860     6       0
1k      790     12.5    0.39
10k     706     7.69    6.76
100k    510     2.54    126


Another inductor measures:
f,Hz    L,uH    Q       ESR
100     730     5.8     0
1k      680     11.67   0.33
10k     615     9.05    4.25
100k    470     2.83    103.8

The third inductor measures:
f,Hz    L,uH    Q       ESR
100     737     5.79    0
1k      662     13.2    0.33
10k     600     8.7     4.26
100k    451     2.75    103

So, the buzzing inductor has much higher inductance than the others. It looks exactly the same, so I think that there is some kind of fault.

What kind of fault could cause the inductor to buzz and to measure higher inductance value?

[wraper]

If you are measuring this in circuit, measurement difference probably is caused not by inductor. I suggest looking for faulty film capacitors.

[Pentium100]

If you are measuring this in circuit, measurement difference probably is caused not by inductor. I suggest looking for faulty film capacitors.

I measured the inductors out of circuit, well, with the output module removed and the inductor still in the chassis, there could have been some sense resistors, but I do not think so.

[wraper]

If you are measuring this in circuit, measurement difference probably is caused not by inductor. I suggest looking for faulty film capacitors.

I measured the inductors out of circuit, well, with the output module removed and the inductor still in the chassis, there could have been some sense resistors, but I do not think so.

Swap 2 inductors and see how it works.

[TheMG]

I measured the inductors out of circuit, well, with the output module removed and the inductor still in the chassis, there could have been some sense resistors, but I do not think so.

Are the output capacitors part of the output module that you removed? I'm not familiar with that specific model UPS but a lot of UPS that size that I've worked with, have the film capacitors mounted in the chassis usually in close proximity to the output inductors.

It would be very unusual for an inductor to increase in inductance, in fact I don't really see how that would be possible. The more likely failure scenario would be a decrease in inductance as the result of failing insulation on the windings causing shorted turns.

[Pentium100]

Are the output capacitors part of the output module that you removed?

Yes. Everything is on the module, only the input and output inductors are on the chassis. I measured the output capacitors and they measure fine.

I've been wondering what can cause the increase in the inductance as well, or at least the inductance to be measured higher. Maybe the core is a bit loose? But then an airgap would probably decrease the inductance, not increase it. It also could be that this inductor was always higher inductance for some reason, but it's too much of a coincidence IMO.

The inductors have air flow cooling them, but still get quite warm. The buzzing one did not seem to be hotter than the other ones (then again, I checked this some time after shutting the UPS off). I waited for them to cool down and then measured again, with the measurements staying pretty much the same.

I wonder how else I can test them. I will try swapping two of them and see if the buzz stays with the inductor.

[Pentium100]

I managed to remove the inductor and found out that a couple of core laminations are loose and can move a few mm (see attachment). The coil former (not sure about the term, the plastic part with the winding) can also move up and down on the core.

I guess now I have to figure out how to glue the inductor so that its parts do not move.

That inductor is rated at 700uH, but because of the loose lamination, it measures as higher inductance. If I put the lamination back to its place, the inductance drops.

[T3sl4co1l]

Heh, that'll do.  Hold it in place and squirt some superglue under it?

Tim

[Pentium100]

I used epoxy. Glued the few places where I saw the laminations were loose, put the inductor back in the UPS and now it's silent.