doing in house EMC pre-compliance testing - where do I start?

[nctnico]

I don't feel antennas are all that helpful, outside of a lab setting; you aren't going to have the same gain calibration (due to antenna curve, environment, orientation, etc.), and ambient sources get in the way.  Just a good set of probes (E and H field, current clamp), and decoupling networks (LISN, CDN), covers quite a lot.

Agreed. Without a shielded environment, the noise from the environment will swamp any signal you measure. Any antenna is just a waste of money. Even bench measurements with E/H field probes and LISN can be affected by outside sources.

A 'cheap' way to get some kind of shielded environment is a TEM cell and/or a shielded tent but you still won't get anything that is like an EMC lab grade measurement.

[fourfathom]

So if I put 600µV into this 600mm dipole antenna, is that 1mV/m?

Probably not.  An E field of 1 V/m refers to a potential difference of 1 V between two points 1 m apart.  There is an equation for free-space field strength given an isotropic (ideal, zero size) radiator, and the antenna factor is usually relative to that.  Also, your 600mm dipole is probably not matched to your cable at the frequency of interest, so you will have reflections and frequency-dependent cable attenuation (which will be greater than the spec'd cable attenuation.)

A short antenna is still a good tool.  If you make it very short and put a 50 Ohm resistor at the coax connection you will have a fairly well-matched antenna with a linear antenna factor (below the antenna resonance frequency.)  It's not going to be calibrated unless you do the math, but it shouldn't have any surprising behavior.  Use it for relative measurements.

[Simon]

So if I put 600µV into this 600mm dipole antenna, is that 1mV/m?

Probably not.  An E field of 1 V/m refers to a potential difference of 1 V between two points 1 m apart.  There is an equation for free-space field strength given an isotropic (ideal, zero size) radiator, and the antenna factor is usually relative to that.  Also, your 600mm dipole is probably not matched to your cable at the frequency of interest, so you will have reflections and frequency-dependent cable attenuation (which will be greater than the spec'd cable attenuation.)

A short antenna is still a good tool.  If you make it very short and put a 50 Ohm resistor at the coax connection you will have a fairly well-matched antenna with a linear antenna factor (below the antenna resonance frequency.)  It's not going to be calibrated unless you do the math, but it shouldn't have any surprising behaviour.  Use it for relative measurements.

The antenna was plugged straight into the signal generator.

[fourfathom]

The antenna was plugged straight into the signal generator.

Was it decoupled from the ground and power wiring of the generator by a choke or something?  Otherwise the generator and power cable become part of the antenna.  Even decoupled, the generator chassis in close proximity to the dipole will affect antenna impedance and pattern..  And how do you even connect a dipole to a signal generator without using a cable?

I once made an "active dipole" signal source using a simple transistor oscillator and a 9V battery on a scrap of circuit board.  The whole thing was much smaller than a credit card, with the dipole wires directly attached, but I assume that's not what you did.  It was just an experiment, I was bored.

[fourfathom]

Without a shielded environment, the noise from the environment will swamp any signal you measure. Any antenna is just a waste of money. Even bench measurements with E/H field probes and LISN can be affected by outside sources.

I've done lots of useful compliance work with measuring antennas in the company parking lot, in the middle of the city.  Yes, there's noise, but if you already know what you're looking for you can usually work around the ambient crud.

[Simon]

Yes well I know I have this 70MHz peak, I need to see that on my analyser, I now know that I can keep reducing the RBW to get the noise floor down and this will eventually have my signal stand out from it hopefully. It looks like moving to an unused upstairs floor of the building has done it in terms of noise floor.

[nctnico]

Yes well I know I have this 70MHz peak, I need to see that on my analyser, I now know that I can keep reducing the RBW to get the noise floor down and this will eventually have my signal stand out from it hopefully. It looks like moving to an unused upstairs floor of the building has done it in terms of noise floor.

Reducing RBW is not always a good idea. If the interference signal is wideband noise, then the amplitude will drop as well. If you really want to drop the noise floor, then set the attenuation to 0 and turn on the pre-amplifier.

[mikeselectricstuff]

I don't feel antennas are all that helpful, outside of a lab setting; you aren't going to have the same gain calibration (due to antenna curve, environment, orientation, etc.), and ambient sources get in the way.  Just a good set of probes (E and H field, current clamp), and decoupling networks (LISN, CDN), covers quite a lot.

Agreed. Without a shielded environment, the noise from the environment will swamp any signal you measure. Any antenna is just a waste of money. Even bench measurements with E/H field probes and LISN can be affected by outside sources.

A 'cheap' way to get some kind of shielded environment is a TEM cell and/or a shielded tent but you still won't get anything that is like an EMC lab grade measurement.

But even with a shielded environment, reflections mean that moving anything even slightly could easily change your readings substantially.   You can't win....

[fourfathom]

If you really want to drop the noise floor, then set the attenuation to 0 and turn on the pre-amplifier.

That may work for the Spectrum Analyzer noise floor, but will make no difference if the noise is from the RF ambient environment (as appears to be the case).  Reducing the measurement bandwidth will reduce *some* of the ambient junk, but (as mentioned) will also reduce the measured level of your own signals if they are wideband.  For narrowband signals such as normal clock leakage then reducing the bandwidth shouldn't affect measurement accuracy.

[fourfathom]

I suggest that you don't worry about not having an anechoic or shielded chamber.  They are great if you can find one (and I have), but you can still make useful measurements in a noisy environment:

• Scan in close, to see what is radiating.  If you have the test report use that to guide you.
• Be able to switch the DUT on and off, to identify your signals from the ambient hash.
• If possible test outside rather than inside next to racks of other noisy gear.
• Accept the fact that you will have to personally identify and measure the offending signals, and personally ignore the ambient ones.  You can't just push a button and have your test equipment do a pass/fail scan for you.

[jonpaul]

Bonjour à tous

EMI and other compliance is a deep field, the equipment design at the start needs to be cognizant of requirements for conducted and radiated EMI, safety and transients susceptibility for te regions contemplated for market, and equipment class, eg consum, industrial, medical, avionics

Failing one test by a huge margin, you will not be able to put a bandage on a faulty design eg noisy SMPS.

I reccomend a complete système design review by an experienced EMI, Comlliance engineer or laboratorys equipped with the cal traceable SA, LISN, Antennas.

Expect $20k..150k and 6 months to finish the job.

Beware of Chinese competition that will do no compliance and merely counterfeit the CE, UL, VDE mark.

Bon courage

Jon

[nctnico]

I suggest that you don't worry about not having an anechoic or shielded chamber.  They are great if you can find one (and I have), but you can still make useful measurements in a noisy environment:

• Scan in close, to see what is radiating.  If you have the test report use that to guide you.
• Be able to switch the DUT on and off, to identify your signals from the ambient hash.
• If possible test outside rather than inside next to racks of other noisy gear.
• Accept the fact that you will have to personally identify and measure the offending signals, and personally ignore the ambient ones.  You can't just push a button and have your test equipment do a pass/fail scan for you.

Even then you might still end up short handed. Currently I'm working on improving a PoE power supply based on an existing design that barely passes, so I wanted to add some extra margin. I've added filters that basically seperate the PoE supply from the rest of the board and which makes the power supply look super quiet on a LISN. EMC lab result: worse than before for radiated emissions. And the biggest problem is that I can't really reproduce the radiated emissions test in my own lab so I'm basically flying blind as whether additional modifications I made based on H field probe measurements have a positive effect. I suspect there is some influence from the metal casing as well. Only going to the EMC lab for a short test will tell if I'm on the right track or not. At an EMC lab they have the proper test setups for the test (shielding, ground plating, proper probes like LISN / ISN, etc) which also affect the results.

[tautech]

We had a product at work fail EMC in CE marking. So we have bought a spectrum analyser and will be getting the required licence on it for the required filters etc. We have an aerial, and that is all I know.

How do we go about setting this up for the pre-compliance the spectrum analyser claims it can do? The analyser itself has some sort of output as well as input. Is this so that it can self characterise the setup and make the required adjustments to get a roughly good metric against the standard?

I am sorry to say that Telonic have been pretty useless in advising on what we need to the point that I have an aerial but no cable for it, they must be doing so well that dealing with customers in order to make a few more quid on some cable is totally not worth their while.

It seems you're just after some generic advice even after deciding to commit to a SA purchase ! 
Did you have no idea of what you were getting into or is not revealing the brand of what you purchased didn't want to be a brand advertisement ?
Most manufacturers have App notes on this stuff however without knowing which of these you selected or if it included a EMI promotion we don't know which App note to direct you to. 
https://telonic.co.uk/product-category/test-measurement/bench-spectrum-analysers/

First steps into EMI detection are usually with near field probes, yes the sort of stuff that is often offered in promotions.

[nctnico]

IMHO having an EMI option on a spectrum analyser does not bring much to the table where it comes to hunting down causes of the excessive emissions.

[T3sl4co1l]

IMHO having an EMI option on a spectrum analyser does not bring much to the table where it comes to hunting down causes of the excessive emissions.

Indeed, the standards use average too, and 9 isn't much less than 10kHz bandwidth.  QP is less than peak, but if peak is low enough, QP certainly will also be.

Doing it with a scope, would be... more of a challenge(!), but nonetheless possible.  The main downsides are signals in the ~mV (preamps are a must), and working out averages vs. frequency is really hard.  Can also be troublesome just to even trigger on a given signal, when there are multiple frequencies in play; when this happens, a local probe probably helps (trigger on the source you're looking for), then averaging on the trace (DSO FTW).

Tim

[nctnico]

Agreed that an oscilloscope isn't a substitute for a spectrum analyser. An oscilloscope is useful though to measure ringing in any power related circuit that uses some kind of switching / PWM topology.

[T3sl4co1l]

Indeed, the combination is powerful: correlating a spectral peak to a specific waveform and location, may take some investigation to trace out, but tells you, well, basically all there is to know about the case.

Tim

[alm]

Can also be troublesome just to even trigger on a given signal, when there are multiple frequencies in play; when this happens, a local probe probably helps (trigger on the source you're looking for), then averaging on the trace (DSO FTW).

Would you need to trigger? Couldn't you just acquire a few mega samples of data and run FFT on that? Of course this doesn't solve the dynamic range issues.

[nctnico]

During radiated / conducted emissions tests average or quasy-peek measurements are used. So a pulsating / sporadic signal won't cause much trouble to pass EMC testing; the continuous tones (or frequency bands with continuous emissions) are the problem to pass EMC. There isn't much to trigger on.

[T3sl4co1l]

Can also be troublesome just to even trigger on a given signal, when there are multiple frequencies in play; when this happens, a local probe probably helps (trigger on the source you're looking for), then averaging on the trace (DSO FTW).

Would you need to trigger? Couldn't you just acquire a few mega samples of data and run FFT on that? Of course this doesn't solve the dynamic range issues.

But that just tells you what the spec already tells you; which is nothing about how they're correlated in time.  Usually you'll have (semi-)periodic signals in the circuit, and you can look at each, e.g. SMPS switch node, or CPU clock (if they differ), to tell which is which.  But if they have similar amplitudes at the point of measurement (e.g. probing a cable or LISN), you can't trigger on one or the other, you get both.  Sometimes the amplitudes differ enough to catch just the taller one, but if you're trying to track the smaller one, that's no good.

Put another way: you see some peaks on the spectrum.  How do you correlate those with some impulsive noise in the circuit, and ultimately to an offender?

And, say, placing a finger on the offending switch node, ought to pull enough common-mode current to raise the respective peaks on the spec (assuming the switch node is safe to handle..!), so you can figure it out in that case, without having to track through a scope.  But that might also be hard to do (say with everything buttoned up), but say you've got access to some related signals, that might be a use case for the scope.

...Oh, and about nonperiodic stuff; if you can trigger on it and see that that's the offending problem at the LISN or whatever, that helps, too.  Spread spectrum, or variable-frequency (e.g. QR flyback, LLC) converters, the spectrum is just a blur, easy to hide many overlapping sources within.  Scope can help you separate those too.

During radiated / conducted emissions tests average or quasy-peek measurements are used. So a pulsating / sporadic signal won't cause much trouble to pass EMC testing; the continuous tones (or frequency bands with continuous emissions) are the problem to pass EMC. There isn't much to trigger on.

So no one's ever had an SMPS fail radiated (>30MHz)?

Trouble is, impulsive noise also tends to have high peak amplitudes.  A 400V switch edge, even with considerable attenuation, can still be say 10s of mV peak on the outside, and QP won't quite save you enough to pass.

Tim

[Simon]

We had a product at work fail EMC in CE marking. So we have bought a spectrum analyser and will be getting the required licence on it for the required filters etc. We have an aerial, and that is all I know.

How do we go about setting this up for the pre-compliance the spectrum analyser claims it can do? The analyser itself has some sort of output as well as input. Is this so that it can self characterise the setup and make the required adjustments to get a roughly good metric against the standard?

I am sorry to say that Telonic have been pretty useless in advising on what we need to the point that I have an aerial but no cable for it, they must be doing so well that dealing with customers in order to make a few more quid on some cable is totally not worth their while.

It seems you're just after some generic advice even after deciding to commit to a SA purchase ! 
Did you have no idea of what you were getting into or is not revealing the brand of what you purchased didn't want to be a brand advertisement ?
Most manufacturers have App notes on this stuff however without knowing which of these you selected or if it included a EMI promotion we don't know which App note to direct you to. 
https://telonic.co.uk/product-category/test-measurement/bench-spectrum-analysers/

First steps into EMI detection are usually with near field probes, yes the sort of stuff that is often offered in promotions.

Che'? Don't worry it's one of yours, and my next purchase will not be from telonic, they just shift boxes and can't even explain the promotion clearly or what they are selling. They are about as useless as the $5 meters they used to embarrass themselves with in sending one with an oscilloscope that cost $700 as though you did not have a meter you actually trust already before spending that much on a scope.

I don't understand why you think I would intentionally not say the brand, a machine is a machine and how you use it is about the same whoever made it.

If I understand correctly the EMI options are indeed a waste of time. when I use the QP filter something that I could pick up before literally vanishes, all noise including the SA's is a flat line along with a signal that stood 20dB above the noise. Could well be I do not know what I am doing yet, I have mostly read the manual.

My first thing will be to run the offending machine and see what that lab test looks like in my environment. I will in fact try to get a second machine to modify so that we can always do a side by side test of the offender and the new candidate, if the new one is 25dB quieter we have some sort of chance.

[nctnico]

Having two machines is a good option but be sure they are equally bad before doing any modifications.

For the project I am working on, I have 3 boards available and one of them seems to behave better than the other two. Typically I go back & forth between changes on the same board to make sure the change I'm making really has some effect. After that I use a 2nd board and also measure the difference between unmodified and modifed as a verification.

[tautech]

I don't understand why you think I would intentionally not say the brand, a machine is a machine and how you use it is about the same whoever made it.

Really, so you don't think we might have some experience in this field as many of our SA sales are for exactly this and many on the forum are doing this work too however generic advice won't get you far with the different UI's, feature sets and core capability of the many different devices on the market.

Don't worry it's one of yours

Mine, don't you more correctly mean the same brand as we sell ?
Exact model please ?

Good, there's lots of guidance in US App notes:
https://siglentna.com/application-note/electromagnetic-compliance-pre-compliance-test-basics/
https://siglentna.com/application-note/electromagnetic-compliance-pre-compliance-conducted-emissions-testing/
This content needs updating if you purchased a later SSA3000X Plus analyzer as it was written for the earlier SSA3000X that doesn't have the EMI mode or the inbuilt range of CISPR-filters the later units do.

Another good guidance App note:
https://siglentna.com/application-note/pre-compliance-basics-analyzer-configuration-settings-and-protection/

[Simon]

None of my questions have been about model specific stuff, it's literally been about, having this kit, what da hell do I do with it now? how does the, let's say, physics of it work?

It's this one: https://telonic.co.uk/product/siglent-ssa3015x-plus-1-5ghz-spectrum-analyser-free-tg-licence/#reviews How the tests are done I would expect to be the same with any unit which is why the standard tells you what settings to use. Having realised that the spectrum analyser will have a noise floor of it's own I assume that this is what would make the most difference to measurements when you have a standard that wants signal below a certain level that the analyser may not be able to do.

I'll read the app notes and see if they clear anything up for me.

[tautech]

None of my questions have been about model specific stuff, it's literally been about, having this kit, what da hell do I do with it now? how does the, let's say, physics of it work?

The clue is in the terminologies used; EMI and RFI, namely the interference emitted and how you might detect it and measure it to prove that it's inside the standard.

It's this one: SSA3015X Plus
How the tests are done I would expect to be the same with any unit which is why the standard tells you what settings to use. Having realised that the spectrum analyser will have a noise floor of it's own I assume that this is what would make the most difference to measurements when you have a standard that wants signal below a certain level that the analyser may not be able to do.

For low levels you will need the Preamp engaged and especially with the HH RFI probes if emissions are somewhere down where they need be.

I'll read the app notes and see if they clear anything up for me.

They'll help for sure and there's also a lot of videos on EMI/RFI work here:
https://siglentna.com/videos/ssa3000x-plus/

Becoming conversant with the User manual Chapter 7 EMI Measurement starting on P112 will help too:
https://siglentna.com//wp-content/uploads/dlm_uploads/2022/08/UserManual_UG0703P_E02D.pdf

One of the first things with SA's and RFI/EMI probes is to leave your phone a few yards away.
Do NOT be tempted to reduce the automatic input attenuation and if the overload screen warning is displayed or the overload buzzer goes off rapidly remove the DUT, detection device or kill the DUT power.