EMC conducted emissions pre-compliance testing with home-made LISN

[Jay_Diddy_B]

Hi,

I am not sure where your phase limits are coming from. If I analysis the circuit in LTspice:




LTspice gives a phase angle of 84.7^o which is pretty close to the 84.18^o that you measured.



There is really good correlation between the response predicted by LTspice and the measurements that you made. This indicates that the circuit is 'free' from parasitic components.

Regards,

Jay_Diddy_B

[EMC]

Phase is from the CISPR LISN Spec.   I have seen several LISNs and all are high at 150kHz; I don't know how to lower it.

[Jay_Diddy_B]

Hi,

This circuit matches the impedance curve:



These are the results:



This was achieved by placing a 1 Ohm resistor in the circuit.

The problem is, I can't find a 5uH LISN schematic with the resistor in it.

Regards,

Jay_Diddy_B

[T3sl4co1l]

Hmm, an FCC Part 15 LISN defines a 5 ohm resistor in series with the bulk capacitor towards the mains side.  Which seems reasonable to me as setting the low frequency line impedance and damping characteristics.

Tim

[EMC]

So they must be providing for the resistance of the inductor.   I guess the easiest solution would be to just live with it.

[Leo Bodnar]

Here are some pictures of a typical 50/250µH LISN.
Leo
P.S. I meant to post them in line LISN thread but for some reason they have ended up here.  It's journey that matters, not the destination.

[skipjackrc4]

A 2 dB difference between measurement facilities is actually pretty good.  That's well within normal component and setup tolerance for EMI measurements. 

When performing EMI measurements--particularly those that involve common mode currents--you need to make sure that the power leads between the LISNs and the EUT are of the correct length and height above the GND plane.  The common mode impedance is affected by conductor height, and changing it can easily result in 2+ dB of difference.  The height is probably specified in the standard, but I've never actually read any of the CISPR requirements, so maybe not.  Foam (watch out for ESD) or construction lumber works well for making cheap cable standoffs.

You also twisted the power leads coming from the EUT.  Did the test lab do this as well?  That will alter differential mode impedance.  You typically want your test leads between the LISN and the EUT (pre-LISN leads don't matter as much) to match as closely as possible to the kind of power leads that will be used in application.  If your end user won't use twisted power leads, then you shouldn't test with them twisted.

[D3f1ant]

Is there much to gained (or is that lost) by having the ground plane when doing quick verification tests on bench for pre compliance? I've not bothered, I tend to do a quick scan to make sure it's not obviously going to fail, and make sure I'm a few dB below limits and its always worked out so far. Guess it wouldn't take much effort to put some kind of plane on the bench, maybe a big sheet of pcb would do.
Radiated is harder, I don't bother putting things in the TEM cell until I have a point of reference from a EMC lab, every product gives unique results due to the near field nature of the TEM cell I guess.

[skipjackrc4]

Is there much to gained (or is that lost) by having the ground plane when doing quick verification tests on bench for pre compliance? I've not bothered, I tend to do a quick scan to make sure it's not obviously going to fail, and make sure I'm a few dB below limits and its always worked out so far. Guess it wouldn't take much effort to put some kind of plane on the bench, maybe a big sheet of pcb would do.
Radiated is harder, I don't bother putting things in the TEM cell until I have a point of reference from a EMC lab, every product gives unique results due to the near field nature of the TEM cell I guess.

If you're trying to do accurate measurements, then yes, you need the plane.  The presence or lack of a reference plane can alter the common mode currents by a quite a bit (or not at all, depending on the source of the current).  Lifting cables above the plane and waving them around can result in changes of over +/-10 dB in emissions (both conducted and radiated).  This is why it's so important to use standardized setups, and is probably the single biggest issue in EMC repeatability.

A few strips of Aluminum foil from the kitchen can work in a pinch, as long as the seams are well connected.  You want to plane to extend beyond the boundary of the EUT+LISN+cables. 

[D3f1ant]

Interesting, will have to do some experiments. +/-10db is a huge difference. I had a product last year that I measured as being a few db below and it failed anyway. I made further improvements and ended up way below on my setup and still only just passing, maybe the ground plane is the difference there. Sounds like I really should find/allocate some space for a better setup.

[skipjackrc4]

Interesting, will have to do some experiments. +/-10db is a huge difference. I had a product last year that I measured as being a few db below and it failed anyway. I made further improvements and ended up way below on my setup and still only just passing, maybe the ground plane is the difference there. Sounds like I really should find/allocate some space for a better setup.

The setup doesn't have to be permanent--just role some foil out on the bench.  That's obviously not as good as a solid copper plane that weighs 100+ lbs, but it should get you a lot closer.

[Smokey]

Bump for excellent compliance information.