Conducted immunity and emissions with in-house equipment (EMC)

[Gabiruman]

Hello there.
I'm not sure if EMC compliance fits in this category, but I couldn't find a better one, in case there's a better one please relocate me.

Currently, in my job, I'm trying to make an EMC test plan for our product, that is Conducted and Radiated Emissions and Immunity.
The problem lies with the equipment used in Conducted testing, professional current probes and transformers for monitoring and injection are quite an expensive purchase for something we won't use that often.

So I was wondering if there's a cheaper way of building our own probes with sufficient injection and monitoring capability.
We already have a LISN (TEKBOX TBOH01 5uF), and we've measured some emissions with it and I also know it's possible to inject power with it, the thing is how much and if it's enough. Also the standard specifies we have to separate Common mode from Differential mode, and for what I've seen online that's not possible without a current probe.

I'll attach the standard we're following, and I'll write down some of the specifications that were provided by our contractor:

CS on power leads – Sine Wave Injection – Differential Mode
- Vinjection = 1 VRMS, 9 kHz to 50KHz
- Vinjection = 1 VRMS, 50KHz to 50MHz, with a 50% AM envelope modulation 1 KHz square wave.

CS on power leads – Sine Wave Injection – Common Mode
Vinjection = 1 VRMS, range from 50kHz to 50MHz, 50% AM envelop modulation 1 KHz square wave.

CS on signal lines – Sine Wave – Common Mode
Subject the unit to the following “Bulk” CM sinusoidal injection superimposed on the signal interface lines:
P incident limited to 1W/50? (CW):
- Iinjection = 86 dBuArms,
- Frequency range: 50kHz - 50MHz
- Modulation: 1 kHz pulse repetition rate, 50% duty cycle

As for monitoring emissions, the frequencies vary from around 30Hz to 100MHz, but what's really important is between 10kHz and 50MHz
?
We have no experience in this area so any help would be appreciated, in the meantime I'll be making some research and updating accordingly.
Thanks in advance,

Best regards,
Gabriel

[T3sl4co1l]

First of all, find what standards you need to comply with.  Purchase a copy of them.  They will contain instructions on how to test equipment, how to build the coupling networks (well, on a basic level), and what signal levels and waveforms are needed.

It sounds like you're partway there already: you've got a good idea of what to apply, and how.

For conducted emissions, a LISN or CDN is more common than current probes, but I know some standards prescribe that instead, so YMMV.  These usually measure normal mode (i.e., each line is decoupled independently and measured), or common mode (pairs are decoupled together and one signal is measured).

Equipment can be rented, or networks can be built (and tested to verify they don't have quirky response).  A ground plane is easy to set up, but a shielded and dampened test chamber is rather more difficult to make or buy.  For emissions, make sure to measure background levels, so you can ignore peaks due to nearby transmitters and whatnot.

Tim

[Gabiruman]

We need not to purchase any standards, the only standard we have to sort of comply with is the one in the original post (ECSS-E-ST-20-07C), and even in this one there are several specs that can be tailored in our way. So this is as much descriptions as we're gonna get.

Like I mentioned the problem is the equipment used in conducted (emissions and susceptibility) tests. We already have a small semi anechoic chamber that houses the DUT (which is itself shielded from radiation, so the only susceptible parts are the cables and antennas) and we've used the chamber for preliminary tests, to properly test radiated emissions and immunity we have access to a "big" anechoic chamber from a University nearby.

We also have a LISN, but I'm struggling to find a way to perform the tests with it only, I can measure conducted emissions as a sum of CM and DM. And perhaps I can split both components with a power splitter and combiner, but I'm not sure if that'll do for injecting power.
The thing is current probes are used in almost all of the tests in the standard either for monitoring or injecting power and I won't be able to that all that with LISN only, at least I don't think so, thus the original question, how to make an in-house decent current probe or similar equipment.
We've already asked for quotes for this equipment, and even though they didn't rent any and we won't use the probes every day, we'd like to have a solution for those measurements at hand, that's why we're trying to make them and every other expensive equipment involved.

[T3sl4co1l]

Ew, yeah, it doesn't define probes, and it doesn't use LISNs for injection and measurement (at least most of the time, at a glance).  I guess you depend on the probes being within calibration.  So, Pearson and Tektronix (or other suitable) AC and DC current clamps are called for.

Pearson CTs aren't especially remarkable, they're just well made and calibrated.  The key is a toroidal winding with as few turns as possible -- 100 or less, say, so the resonant frequency is high.  Toroidal resonant modes are mostly shunted away, by enclosing the winding in a metal case, a torus that is slitted in the plane (but only the inner ring).  They also use nanocrystalline core material, to extend LF bandwidth.

You can do much the same by winding a single layer on a high-mu ferrite toroid, and wrapping that with foil tape so that the outside is a solid conductive layer, but a gap is left in the middle.  Connect the tape to the winding so it's grounded.  Ideally you'd have a balanced output (so, the tape would be grounded to a center tap, and the winding connects to shielded twisted pair cable), but that sucks, so you'd probably just ground one end of the winding instead, and use coax.  (This suggests possible improvements, like a balun, which might be as simple as a CMC ("current balun"), or a full on transformer.)  Also, don't forget the burden resistor, and an attenuator if needed.

To calibrate such a beast, you need a consistent current source.  You might set up a large coax or stripline model, where a conductor is surrounded on all sides, sizing things to get a 50 ohm characteristic impedance.  Use tapered sections to go between the input connector and the large model.  Put the current probe on that.  Like the, oh, I ran across that, uh, 5.2.8.3, make that thing and calibrate your probe against a 50 ohm termination on one end and a known source on the other.



Tim