Isolation transformer for LISN - is shielded bad, or not an issue?

[julianhigginson]

I'm about to buy an isolation transformer for my LISN, and I see the one basic differentiation between the options available (well, apart from power) is that I can have shielding between the windings, or not.

Now shielding isn't mentioned in the test standards or the generic descriptions of hooking up LISNs I've seen, but given the LISN is just a big filter that should be blocking any HF which could travel across the isolation transformer in either direction, I expect it's not going to make a difference either way.

But before I buy one, if everything else between two transformer options is the same, is there anything I might have missed that would mean I'd *not* want to have a shielded transformer for the purposes of using with a LISN?

[T3sl4co1l]

It only matters to the extent that your mains is noisier than the mains-to-EUT isolation of your LISN filter.

I don't see any reason shielding would be undesirable.  A shield changes the HF response of the transformer (namely, it lowers the winding impedance somewhat), but that's a complex circuit regardless, and the common mode response (presumably) will be much better with.

To clarify: consider each winding is a bunch of wire on top of wires, in other words, transmission lines.  So there is some high frequency impedance characteristic of the winding.  Obviously enough, this is well damped when matched, so put a resistor across it.  But a resistor would waste mains power, so put an R+C across it.  Or don't even bother, because your LISN is a couple big inductors anyway.  Which, to be clear, doesn't isolate the winding, because the inductor(s) are just more wires on top of wires -- both should be damped.

I personally wouldn't worry about it, and just get whichever is cheaper.  If the mains is measurably noisy already (you can test the LISN without an isolation transformer, just be careful), consider getting a shielded one, or adding a line filter.  That's about it.

Tim

[julianhigginson]

cool!

Yeah I was just wondering if anyone had tried already and had issues and had an "oh no, that's a bad idea because of x" that I hadn't considered.

buying one now.

[RFDUK]

Not related directly to your question re transformer options, but here is an unexpected 'feature' of using an isolation transformer during conducted emissions testing with a LISN.

I have a design engineer customer who brought along a new in development SMPSU, about 100W if I remember correctly.

My test set up used a 500W isolation transformer, one of those yellow enclosure types with a carry handle.

On the first visit the PSU blew up on switch on. Back at base he had worked on this PSU for some time without problems. We measured the AC supply voltage and that was fine.

He went away, talked to the silicon manufacturer apps support. The device was deemed to be virtually foolproof in the apps circuit as long as the input Vmax was not exceeded.

Came back for another test and again the PSU blew on switch on. Replaced the device and this time armed with a scope we captured an input over volts on switch on.

Long story short, the 500W isolation transformer was the cause. Replaced with a 3kW item, no over shoot, PSU working as expected and relieved engineer.

I don't understand the mechanism, maybe someone can add something,  but this was a definite feature of the small isolation transformer.

Martyn.

[T3sl4co1l]

Difficult to say without an equivalent model of the EUT's input circuit, but it could be that leakage inductance is resonating with input filter capacitance, with a high enough Q to cause problems.

The larger transformer would exhibit the same thing regardless, but likely has lower leakage, maybe leading to lower Q and faster overshoot, which didn't break the EUT?

Poor guy, he should've done surge testing first...

Tim

[CopperCone]

what would be a good way to test for a nasty resonance to determine if its OK to pair stuff with the LISN?


All you can do is shift the peak frequency of the resonance around right, to an area where the circuit does not produce energy in? I came across the power company saying this in regards to commercial line reactors. The LISN is a kind of line reactor but its weird. You can't really dampen it much. And I have no idea if you can make a notch filter that cancels it out? That would be a better solution because if you shift the resonance around you don't really make a universally compatible system, especially if the switching frequency of the thing being tested changes or some thing?

[julianhigginson]

hmm.. that's really interesting.

yellow with a handle sounds like a worksite safety transformer. but I don't know enough about them to know if they are really the same a as a generic isolation transformer or not.

Do you think that the issue was purely between the LISN and the transformer? Or had you used that same setup before with other things and no issues?  (that said I guess something previously tested could still have seen the same power event, but survived it, and you wouldn't know..)

[RFDUK]

Thanks to all for the thoughts coming through here.

The LISN was eliminated as one of the first experiments, so not LISN related. That makes the diagnosis options a whole lot simpler I guess.

The PSU blew on switch on when powered directly from the transformer.

PSU fine when powered directly from the wall socket.

The transformer in the yellow enclosure is potted. There is a mechanical re-settable fuse on the AC input. I linked that out as an experiment at the time, no change, PSU destruction on switch on.