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

[uski]

Hi folks!

I recently had the opportunity to do pre-compliance testing for conducted emissions (CISPR 25).
I'd like to share some feedback.

Initially, I have used a purely home-made LISN (courtesy of Idpromnut).
Then, I could use a Tekbox TBOH1 LISN. Results were similar to the home-made LISN.

I'm using a Rigol DSA832-TG (well the TG is irrelevant here), with the EMI option (I use that for the CISPR 9kHw RBW and I can also do quasi-peak measurements).

However I only had one of each of those LISNs and I wanted to do the measurements with two matched LISNs. I believe that's the proper way of doing things and it allows to differentiate between common-mode noise and differential-mode noise (by comparing the noise on the + and - power supply wires).

So I have used the design available here (thanks Jay_Diddy_B ) to build two identical LISNs :
https://www.eevblog.com/forum/projects/5uh-lisn-for-spectrum-analyzer-emcemi-work/msg404662/#msg404662

Here is my test setup. The LISNs are located at the top-left of the picture (you need to be logged-in to view the pictures)


Screenshots :


Here's another one (measurements are not as good, because of different input settings and lower sweep time - but I wanted to show the pass/fail limits feature) :



And here are the measurements from the certification lab :


As you can see the results are actually pretty close. It looks like the home-made setup underestimates the noise by 2dB but I wouldn't submit the design to the certification lab again with a 2dB margin only, so it's still pretty useful (plus, I know the bias now).

NB: I have enabled the input corrections to compensate for the 10dB attenuator built into the LISN, so that what's displayed on the spectrum analyzer matches the supposed readings. I like this embedded 10dB attenuator, it really makes a difference, previously (with other LISNs) I was getting warning from the spectrum analyzer when I was turning on the UUT ("IF signal out of range"), now it's silent, which is good.

Also there is a DC-block at the input of the spectrum analyzer as an additional protection (Minicircuits BLK-89S+), I could probably safely remove it, but it doesn't hurt, its insertion loss is negligible.

I have used a peak-detector for fast measurements, with a relatively long sweep time (I have chosen to use 50s to make sure I get all the noise). I could go much faster but by nature the noise of my device is transient so I need to make sure the spectrum analyzer spends some time on each frequency to get a good reading.

The span I used was 10kHz-15MHz because I know (from the pre-compliance measurements at the certification lab) that there is nothing interesting between 15MHz and 30MHz.

If anyone has suggestions for improving the setup, or if you have any question, just let me know !

PS: The black cable running from the bottom right of the picture is a USB UART for configuring the UUT. Don't forget to disconnect it before starting testing however it will screw up the grounding scheme (computer ground is safety ground, just like the copper plate and the LISN "-", which is different from the UUT "-")

uski

[uski]

Hi,

EDIT: This post is a reply to a post by someone, that has since been deleted (why? dunno! there's no stupid question!)
The question asked was : how do you explain the difference in frequency for the peaks, between the test house results and yours ? Also it looks like you have a serious noise problem around 650kHz !

Yes, the reason I got into all this was because we (clearly ) failed the conducted emissions test.
We were able to fix the issue by adding a common-mode choke and two capacitors, in line with the power supply wires, and test the effectiveness of this with the setup above. I thought showing the failed tests was funnier than the passed tests

The main responsible for the failure of the test is a DC/DC switcher on our device, which operates around (you guessed it) 650kHz. All the other lines in the spectrum are also caused by it.

The certification lab used a $150'000 EMI test receiver (R&S ESIB 40) which takes many more measurement points than my DSA832, over the spectrum I show here (150kHz-30MHz), so it's more precise.
I'm pretty sure I can get a better measurement if I reduce the span of the DSA832 around the problematic frequency.

Rigol actually provides (if you buy the EMI option) a software that does this, it first does a prescan of the entire band, then it "zooms into" the problematic peaks to get more data around them. But the software is buggy and no fix is scheduled to be released before several months according to their support (thumbs down Rigol on this).

Furthermore, the board I used for the tests is not the same than the one the certification lab received. And the switcher IC I'm using has a wide range of frequencies so a small (even just 0.5%) change in value could explain part of the frequency shift we see.

To confirm this, I just did some measurements again, with the same setup, but a 100kHz span around 650kHz.

With a 9kHz EMI RBW (you need to be logged in to view the pictures) :

The peak is at 648.6kHz

And with a 200Hz EMI RBW :

The peak is at 649.5kHz (it's normal that the amplitude changes if the RBW is changed)

Pretty close to the lab results which see the peak at 654kHz !
Those 5kHz could totally be explained by a slight shift in the resistor setting the switcher frequency and tolerances of the IC itself, plus maybe some slight differences from the test equipment themselves (I'm not sure what's the accuracy of the DSA832). After all it's a spectrum analyzer, not a frequency counter, and I'm only using its internal timebase.

I checked in the switcher IC datasheet :
- 649.5kHz : 207.25kohms
- 654.0kHz : 205.71kohms

I'm not concerned too much, what concerns me more is the 2dB difference in amplitude (when compared to the results from the lab, which should be lower because they're using a quasi-peak detector, not a peak detector like me !), but that isn't a big problem by itself.
EDIT: I think I found why there's a 2dB offset. See below.

If I have time I'll try to measure the insertion loss of the LISN (I have a tracking generator so I can make use of it !)

uski

[EMC]

Uski,

Well done, looks like a good result.    The main problem with the LISNs would probably not have been an issue in this test because of the low DC current.   The test lab used a LISN that still meets the impedance spec when carrying 100 amps.   The LISN inductors in the little guys you made would probably start dropping inductance at a few hundred milliamps DC; but, if thats all you need then thats great.   I would be interested to know if "verification through substitution" would have reduced or detected the error in your measurement (BTW who says you had error, it could be the test house that had the error!).   i.e.  1) remove DC power and ground the power supply end of the LISNs.  2) Attach an ARB/Sig Gen instead of the DUT at the other end of the LISNs.   Increase the ARB/Sig Gen until you get the same DSA832 deflection/amplitude.   I was wondering if the resulting ARB reading would be closer to 75dBuV rather than 72 dBuV.   If closer to 75dBuV then we could assume 2-3dB error in your setup; and, you may be able to track that down and reduce it.

Steve

PS The ARB/Sig Gen will load down below 10MHz.   The impedance at 600kHz should be about 25 ohms.   This shouldn't effect the substitution test.

[T3sl4co1l]

Nice!  Excellent example of how to do it!

Tim

[uski]

Hi,

Thanks guys ! It's great to discuss stuff like this.

Well done, looks like a good result.    The main problem with the LISNs would probably not have been an issue in this test because of the low DC current.   The test lab used a LISN that still meets the impedance spec when carrying 100 amps.   The LISN inductors in the little guys you made would probably start dropping inductance at a few hundred milliamps DC; but, if thats all you need then thats great.

This particular UUT uses 200mA at most. I'm curious to see how much the inductance drops with high currents, is there a way to test this ? For example, I could draw 5 amps of DC current within the LISN (I have a floating power supply and a floating active load), and inject the same test signal through a DC block, would that work ? Or do I need a coupler or something more complex ?

I would be interested to know if "verification through substitution" would have reduced or detected the error in your measurement (BTW who says you had error, it could be the test house that had the error!).

Well I did just that.
I used a HP 8648B signal generator (calibration due date 04-2012 so it's not terribly well calibrated but it shouldn't be too far off...), with a DC block on the spectrum analyzer, the generator, and the LISN in between (I tested with removing the DC block at 300kHz and results were the same).

I soldered a coax directly to the UUT side of the LISN (you need to be logged in to view the pictures) :


The setup looks like this (for the test open at the source side):


I took a few measurement points, setting the generator to 75dBuV:


It looks like there is quite a bit of attenuation at sub-1MHz frequencies, any idea of the possible cause ?
EDIT: I believe it's simply because the impedance starts to decrease below a few MHz. The Tekbox LISN does the same thing :

PS The ARB/Sig Gen will load down below 10MHz.   The impedance at 600kHz should be about 25 ohms.   This shouldn't effect the substitution test.

Could that explain the dip below 1MHz ?
EDIT: I think yes, it's because of the impedance dip.

uski

[rx8pilot]

This is now in my TOP 10 threads list. The timing is perfect since I am about 6 months away from certification testing. Its my first time and a little nervous about the learning curve.

This was super helpful.

[uski]

Hi,

This is now in my TOP 10 threads list. The timing is perfect since I am about 6 months away from certification testing. Its my first time and a little nervous about the learning curve. This was super helpful.

Glad it helped
You can also use the LISN to do conducted immunity testing (i.e. inject crap into the power supply of the UUT).
Don't forget radiated emissions and immunity testing too.
If it's your first time, plan ahead by designing in some filtering, better safe than sorry.
For automotive stuff, TI has good examples of power supply filtering such as, enter "ti cispr 25" in Google for examples. They generally use a CMMC (common-mode choke) for common-mode noise and a LC filter for differential-mode noise.
There is quite a bit of good information scattered on the internet, here's what I found helpful :
"Basics of noise countermeasures" series by Murata
There are quite a bit of papers (registration is free) on the Coimpliance Club, I liked the Banana Skins which gives real-life examples of what bad EMC performance can do.

[uski]

Guys, it's time for another real-life comparison !

I have added a filter on the cable to remove the nasty 650kHz noise. I also did a few other things to my design including raising the switcher frequency to 1.05MHz, to make it easier to filter out the noise. And I sent that to the certification lab again, hoping to fix the issue and finally pass the tests.

So... how does that filter performs ? Was it successful in filtering out the noise and making the design pass the certification ?

Here's how to proceed based on my experience with this. I advice to start by doing a full-scan with the peak detector. The peak detector gives higher noise results than the quasi-peak detector, so if you pass with the peak detector you should pass with the quasi-peak. For readings which are a bit close to the limit or for interesting points, you can do a smaller span around the peak frequency.

With the setup I have mentioned above (first post), I was getting these results, with a peak detector for 150kHz to 30MHz. I disabled the SA built-in input attenuator to get the lowest possible noise floor. This removes amplification inside the spectrum analyzer, so the noise doesn't get amplified.


Then I did a reading with the quasi-peak detector just around the problematic frequency (1.05MHz). The quasi-peak detector is what the certification lab is using :


And I just got the results back from the certification lab, so we can compare :


Pretty spot on ! It looks like my setup is definitely enough to get some confidence before sending a design to the EMC lab for certification, at least for conducted emissions testing.

About the small offset I think I saw previously, I believe it may be related to the UUT input voltage.
The certification lab uses a "car battery" but they didn't specify its voltage. Depending of its stage of charge it could be anywhere between 11V and 13V. In my tests, I used a power supply set at 12.5V to try to be close, the the input voltage level definitely changes the noise output by the switcher. So it's totally possible part of the difference I saw in term of amplitude comes from this. Also it's important to use the proper detector, I used peak to be quick, but for precise measurements quasi-peak (as specified by CISPR 25) and the correct RBW must be used.

[uski]

Oh and by the way ! Very important !

The setup above does not differentiate from common-mode and differential-mode noise.

To do that, you need to first measure with one LISN (the output of the other one is terminated by a 50ohms load as shown).
Keep that data and then reverse the LISNs, measuring from the other one.

As EMC said in another thread, which I highly recommend reading :

When used in pairs the test setup emulates intended environment; i.e. power cable to the DUT.   If the positive result is higher than the negative result then differential suppression, cap to ground.   If positive and negative results are the same then common mode suppression.   If negative result higher than positive result then DUT ground scheme is wrong.

Doing that, I saw that all the noise I was seeing was common-mode noise. So I could take the shortcut of just measuring the noise on the + LISN and I knew it was representative of the common-mode noise. But you can't do that approximation if you didn't check first if you didn't also have differential mode noise !
Also checking the noise on the - LISN is very good, it allowed me to notice something was wrong when I left the PC USB cable connected to the UUT, screwing up the ground scheme and the results...

It seems to be possible to build a CM/DM noise separator, I might attempt it later.
One is shown here or here (2.1.8).

For a very precise measurement I guess you also need that kind of separator, and calibrate it.

[Jay_Diddy_B]

Uski,

Well done, looks like a good result.    The main problem with the LISNs would probably not have been an issue in this test because of the low DC current.   The test lab used a LISN that still meets the impedance spec when carrying 100 amps.   The LISN inductors in the little guys you made would probably start dropping inductance at a few hundred milliamps DC; but, if thats all you need then thats great.   I would be interested to know if "verification through substitution" would have reduced or detected the error in your measurement (BTW who says you had error, it could be the test house that had the error!).   i.e.  1) remove DC power and ground the power supply end of the LISNs.  2) Attach an ARB/Sig Gen instead of the DUT at the other end of the LISNs.   Increase the ARB/Sig Gen until you get the same DSA832 deflection/amplitude.   I was wondering if the resulting ARB reading would be closer to 75dBuV rather than 72 dBuV.   If closer to 75dBuV then we could assume 2-3dB error in your setup; and, you may be able to track that down and reduce it.

Steve

PS The ARB/Sig Gen will load down below 10MHz.   The impedance at 600kHz should be about 25 ohms.   This shouldn't effect the substitution test.

I am the designer of the LISN that uski built.

Let me talk about the inductor.

I decided to use shielded inductors, instead of the more conventional air cored designs used in nearly every other LISN that I have seen. The air cored inductors need a large enclosure so that the magnetic fields do not couple the case. I also want to make the LISN very repeatable.

These are the key specifications of the inductor:



You can see that the inductor has a self-resonant frequency well above the frequency of interest.

Here is the graph showing inductance variation with DC bias current:



The inductance is down to 1uH at 5A. Because I used five inductors in series, the inductance is 10% high at low currents, the nominal value (5uH) at 5A and 1bout 10% low at 10A.


LTspice Model

I can use this LTspice model to explore the impact of the inductance variation:



Note that the inductor model includes these parasitic components:



The model results show:




The inductor variation has more impact at low frequencies at a typical switching frequency of 500 kHz the error is around 1 dB.



The measured results for my prototype are in this message:

https://www.eevblog.com/forum/projects/5uh-lisn-for-spectrum-analyzer-emcemi-work/msg404662/#msg404662


Regards,

Jay_Diddy_B

[uski]

The model results show:

Nice, this is almost exactly what I measured in my post above using the signal generator !

[T3sl4co1l]

Oh and by the way ! Very important !

The setup above does not differentiate from common-mode and differential-mode noise.

To do that, you need to first measure with one LISN (the output of the other one is terminated by a 50ohms load as shown).
Keep that data and then reverse the LISNs, measuring from the other one.

As long as the path lengths are matched (important for VHF+ range), you can use 0/180 degree couplers/splitters (the 3db "lossless" kind, not the 6dB resistor kind) to resolve two outputs into one analyzer.  These are available from e.g. Microcircuits, I think with sufficient bandwidth for the present purpose, too.

Of course, you could build your own LISN with two channels, and this function built in (the transformer won't be terribly hard to wind, but probably won't offer terribly accurate phase matching at higher frequencies, so take it with a grain of salt), if you like.

Also, make sure to terminate the unused LISN output, if you are doing single ended measurements!

As for common mode, trying to filter low frequencies is difficult -- for example, just adding ferrite beads is almost futile, as you're trying to establish a divider between a relatively low impedance (the LISN at 50 ohms or so) and a fairly high impedance (the cables to the device will typically be in the 150-300 ohm range, as a transmission line over the ground plane).  For this example, it also appears the device is pretty small, and the leads short, so it will be electrically short at those frequencies, i.e., acting as an even higher impedance capacitor.

Low frequencies are almost always either due to unfiltered (differential) ripple, poorly designed filtering (the ripple current couples into the ground line), or poor layout (currents flowing across the board create voltage drop between most of what "ground" is, and in turn, between the cable(s) leaving the board).  In all cases, better filtering, and responsible control of ground loop currents (keep everything tightly localized, add ground plane cuts only where absolutely necessary to avoid currents across a filtering or signal path), will greatly improve things.

Tim

[EMC]

Guys,

1) The above data, regards the inductance dropping with DC current, clearly shows my pessimism was unnecessary.   The inductors are fine right up to 1 amp; even then the drop off is negligible.

2) With regard to verifying to LISN impedance at higher DC currents.   Two ways ''test'' or ''simulate''.   I have not done this but believe ''test'' can be done using a DC coupling/decoupling circuit.   If I look up the chebychev tables for a 3 element high pass with series caps and parallel inductor; then combine the cap values, it would be a 400nF cap and 1mH inductor.   Combine both caps so that we have an L circuit, cap to block at the input and L to enable application of the test current, then same at the output but reversed.   Inductor value is too high so it is too hard to realise.   So all of the above LTSpice data and ''simulate'' would be the best way.

3) All of the above concern about impedance and measurement error is not required if you use the substitution method I described.   Test your device to get a a max hold trace on the DSA.   Then set up as I describe and adjust the sig gen to the same deflection.   Read your result from the sig gen then compare to the limit.   The LISN ímpedance, and other errors, are in both readings and therefore removed from the result!     (my only concern is the original 3dB difference with the test house)

4) Regards frequency selection.   I understand your shift to a higher frequency but it's still in the AM band.   Any chance of  going to 1.9MHz?   Apart from a little band at 5.9 to 6.2MHz there are no limits to worry about.   Just dont go for exactly 2MHz or the third will fall in that little 6MHz band.

Great discussion guys.

Steve

PS   The reason the above reading goes low at low frequencies is that there is no inductive reactance, XL = 2 Pi F L.   At DC and low frequencies XL is very low; as frequency goes up so does XL.   But, because of the DSA input Z of 50 ohms is in parallel with it, the total Z asymtotes to 50 ohms.   (bit less that 50 actually because there is a 1k parallel as well)

[uski]

2) With regard to verifying to LISN impedance at higher DC currents.   Two ways ''test'' or ''simulate''.   I have not done this but believe ''test'' can be done using a DC coupling/decoupling circuit.   If I look up the chebychev tables for a 3 element high pass with series caps and parallel inductor; then combine the cap values, it would be a 400nF cap and 1mH inductor.   Combine both caps so that we have an L circuit, cap to block at the input and L to enable application of the test current, then same at the output but reversed.   Inductor value is too high so it is too hard to realise.   So all of the above LTSpice data and ''simulate'' would be the best way.

Maybe Jay_Diddy_B, our LTspice expert, can give this a try if he has some time to spare on this ?

3) All of the above concern about impedance and measurement error is not required if you use the substitution method I described.   Test your device to get a a max hold trace on the DSA.   Then set up as I describe and adjust the sig gen to the same deflection.   Read your result from the sig gen then compare to the limit.   The LISN ímpedance, and other errors, are in both readings and therefore removed from the result!     (my only concern is the original 3dB difference with the test house)

Oh I finally understood what you mean by the substitution test ! Definitely worth a try, especially if there is a need to get a precise reading. But in that case you'd need to have calibrated equipment (at least the generator) otherwise you wouldn't know where the error would come from. In my case the margin was sufficiently high that I wasn't really bothered by 2-3 dB of uncertainty.

If you look at my second measurement, you see that the difference with the test house is not all that big, less than 1dB, when using the proper detector and setup.

EDIT: There may still be a problem with the substitution test: it would not simulate the bias current (see post below by T3sl4co1l). So even if you see the same reading, it doesn't mean the amplitude of the generator is right, because when tested under bias the LISN might have different characteristics.

4) Regards frequency selection.   I understand your shift to a higher frequency but it's still in the AM band.   Any chance of  going to 1.9MHz?   Apart from a little band at 5.9 to 6.2MHz there are no limits to worry about.   Just dont go for exactly 2MHz or the third will fall in that little 6MHz band.

Now that's an excellent point which I haven't been considering...
It is too late for me to change the frequency in that design revision (I can't delay the project and certification with another change and round of test); but I will definitely do it for the next version. I like to be a good radio citizen I'll aim for 1.9MHz. And increasing the frequency will probably make the noise even lower (higher freqs are easier to filter).

I'm surprised that the test house specifies EN 55022 class B limits which are higher than than EN 55025 limits you have mentionned. They specify CISPR 25 but use EN 55022 limits ?! Could that be a mistake on their side ? It looks like I'm still slightly below the EN 55025 limits anyway, but still, this is weird...

Very interesting discussion indeed. I'm learning a lot, thanks guys !

uski

[T3sl4co1l]

2) With regard to verifying to LISN impedance at higher DC currents.   Two ways ''test'' or ''simulate''.   I have not done this but believe ''test'' can be done using a DC coupling/decoupling circuit.   If I look up the chebychev tables for a 3 element high pass with series caps and parallel inductor; then combine the cap values, it would be a 400nF cap and 1mH inductor.   Combine both caps so that we have an L circuit, cap to block at the input and L to enable application of the test current, then same at the output but reversed.   Inductor value is too high so it is too hard to realise.   So all of the above LTSpice data and ''simulate'' would be the best way.

Maybe Jay_Diddy_B, our LTspice expert, can give this a try if he has some time to spare on this ?

Simulating inductance (let alone impedance) under bias is essentially a useless problem:
- Absolutely, positively no one publishes inductor models of that detail.*
- If you have construction data (sometimes**), and can find core data from that manufacturer in turn, you might be able to construct a model for losses and saturation.
- But good luck figuring out losses and inductance as a function of amplitude and bias; in general, you can expect both to vary, but even the core manufacturers don't tell you jack about this (and quite possibly, don't even know, themselves).
- And anyway, analyzing and quantifying all of this is *really* hard.  You get into "butterfly" B-H curves, minor hysteresis loops (which need not be proportional to the full B-H loop), frequency dependency (the size and shape of the loop varies with rate), and heck, might as well throw temperature into the mix as well.  At this level, you're better off making an approximate model from measured data, rather than trying to build any kind of physics-based model.

*You can assemble a basic DCR + L || Rloss || (Rc + Cp) model from standard datasheet values (DCR, L, Q @ F, SRF), with presumably at least gross agreement to a little above SRF.  Few manufacturers give L(F) or Z(F) curves, and even fewer give SPICE models.  Of note, Kemet and TDK have some data available for their capacitors and inductors, though these are usually prepared at a given frequency; and CoilCraft has extensive data (including a slightly more complicated, more accurate AC model) on probably more than half their product line.

I haven't seen a single model published, ever, that claims to address core saturation, for a given commercial part.  (There are simple SPICE models for this -- but, you're doing the parameter fitting yourself.)

**So far, I've asked Bourns and Murata for data on single occasions, and gotten useful enough answers.  YMMV.

So for the present case, I would be more than happy to accept a much more basic problem: does the effective inductance, of the main series inductor, remain within tolerance, at rated current?  And this can sometimes be found in the datasheet already, or calculated from core data.

Tim

[uski]

Hi

So for the present case, I would be more than happy to accept a much more basic problem: does the effective inductance, of the main series inductor, remain within tolerance, at rated current?  And this can sometimes be found in the datasheet already, or calculated from core data.

You're right... and I think Jay_Diddy_B answers that above !

uski

[EMC]

Well test may be the only way, I will work on that.

Just another unrelated comment, if you update the DSA815TG to firmware 1.12 at the bottom of the span menu you can select log x axis.   Much better for this device under test.

[Jay_Diddy_B]

Hi,

Let me try and show why the inductor is absolutely fine for this application.

Test 1 Inductor Saturation Test

I used an Inductor saturation tester, this is described in this thread:

https://www.eevblog.com/forum/projects/inductor-saturation-tester-alternative-route-to-dump-the-excess-energy/msg181720/#msg181720



The scope shows a linear ramp, indicating that the inductance is not changing significantly through 10A.


AC analysis of the LISN circuit



The 0.1uF can be considered to be a short at 800kHz, so the 3dB point on the response is given by:

F = 1/ (2 x PI x L / (Rsource in parallel with Spectrum Analyzer))

If Rsource = 50 Ohm

Spectrum analyzer = 50 Ohms

F = 1 / (2 x Pi x 5uH / 25 Ohms) = 795 kHz (3 dB point)

This is shown in the modelling results:



Stepping the Source Resistance

The source resistance is probably only 50 Ohms if the LISN is being tested with a Network analyzer or a signal generator. The impedance of the noise source is during use is probably much less than 50 Ohms. This is true, because if the noise source had a 50 Ohm impedance it would be very easy to filter.

In this model I am stepping the Source resistance from 1 ohm to 100 Ohm:



The results show how the frequency response changes with the source impedance. The effect is much stronger than a 10% change in inductor value.



The inductors chosen are fine in this application.

Regards,

Jay_Diddy_B

[EMC]

Jay_Diddy_B,

With regard to inductor choice.   If you had the data and knowledge you now have after design, build & test would you go with the WE 744 314 110 again?   e.g. would you move up in current range, i.e. WE 744 331 010 0.   I understand the major changes are:

RDC goes down 3.15m to 2.5m
I Sat goes up 13 to 52 amps
Core 'material from 'Superflex' to iron power 
Size increases to 1210 (12.1 x 11.4mm)
and I assume but don't know max freq comes down

Regardless of that option are you confident that WE 744 314 110 is the best choice?

(Implicit in that question you may think I am saying it is the wrong choice; that is not the case; you have proved it is fine no question.  I just wondered if it could be stretched a bit for higher current.)

Thank you,

Steve

[Jay_Diddy_B]

Hi,

The choice of the WE 744314110 (HCI series) was originally made because I got 500 pieces very cheap from eBay, but it turned out to work exceptionally well. My thoughts were that any deficiencies in the inductor, versus an air core, would be balanced by the really tightly controlled construction.

I have measured my LISN design with a Network Analyzer and it has been compared against commercially available LISNs all with good results.

If you need higher current, then the Wurth 7443310100 (HCC Series) should work. In fact, I used a similar inductor 744 332 1000 in my line voltage LISN. The line voltage LISN is shown in this thread:

https://www.eevblog.com/forum/projects/5uh-lisn-for-spectrum-analyzer-emcemi-work/msg641108/#msg641108


So I can't really say that the 744 314 110 is the 'best' choice. It works well for my applications below 10A.

Regards,

Jay_Diddy_B

[rx8pilot]

I need to buy a spectrum analyzer now.....

Not only do I want to do pre-compliance conducted emissions tests, I want to learn more about the process of the testing. My products need to pass for certification, but they also have to work in electronically delicate environments where conducted and radiated emissions will actually cause problems. I already had some real world problems with immunity where my circuit shut down when exposed to emissions from nearby walkie-talkies - <6 inches. Having a proper setup and some additional education will allow me to test my circuits and the various nearby systems that will interface with it to create an understanding of what is happening.

I figured it out by experimentation, but I need to go full in to be successful. Super thankful for this thread to kick-start the effort. Thank you for the posts.

[uski]

I need to buy a spectrum analyzer now.....

A DSA815 may probably do the trick, at least for conducted emissions.

For radiated, the tests I had to do (depends of your product and market) go up to 4GHz so a DSA815 is not enough. But I think it's best to move that to a separate thread and keep this thread for conducted emissions.

From my understanding CISPR 25 (for automotive) actually allows a spectrum analyzer to do the conducted emissions measurements.

The EMI option from Rigol is nice because it gives you the 9kHz and 120kHz EMI RBW, and an EMI filter, plus some extra features (including the PC software - but it's a bit buggy as I've described).

The HMS-X from R&S might also be an option. Their HMExplorer software is actually much better for EMI than the software from Rigol. You might want to get in touch with them and get a quotation, they may have some demo units for sale.

Not only do I want to do pre-compliance conducted emissions tests, I want to learn more about the process of the testing. My products need to pass for certification, but they also have to work in electronically delicate environments where conducted and radiated emissions will actually cause problems. I already had some real world problems with immunity where my circuit shut down when exposed to emissions from nearby walkie-talkies - <6 inches. Having a proper setup and some additional education will allow me to test my circuits and the various nearby systems that will interface with it to create an understanding of what is happening.

The way the tests are done are specified in the various standards. What you need first is to know which standards apply to your product.
Then you can read these standards and figure out the setup.

If you're serious I guess you should find a partner to help you (a company specialized in EMI testing and project management). I may have some to suggest if you want.

For radiated immunity, you might look into a TEM cell. I heard mixed opinions about this technique, maybe someone with more experience can discuss it (again it might be best to start another thread). It can also be used for radiated emissions but I know you need a big one to get good and repeatable results. TekBox sells one but it's really, really small. Maybe you can find a bigger used one on eBay.

My personal opinion (which can be challenged) is that for serious radiated emission testing, it's best to go into a chamber. For for conducted emissions it's easy enough to do it in house.

EDIT (safety note): I would like to add that for mains-powered equipment, the LISNs used may be dangerous if used improperly due to the high leakeage current they generate. They must be grounded using two separate ground connections and must be checked before each use. So people can do it in house but the equipment can be dangerous when used by untrained people. Also of course the LISNs used in this thread are for DC powered low-voltage applications only.

Near field probes can be used when trying to fix a radiated emission issue, once it's been identified in a chamber, and a TEM box might help (if you're very very careful when placing the product inside the TEM box) to do comparative measurements when attempting to fix the issue.

I figured it out by experimentation, but I need to go full in to be successful. Super thankful for this thread to kick-start the effort. Thank you for the posts.

Yes I really enjoy sharing with the community here we all learn from each other and it's awesome.

uski

[EMC]

Just a couple more thoughts on the Jay_Diddy_B LISN; particularly if a PCB revision is likely.

1) Add 1.6mm PCB edge mounted SMAs at each end for less error when VNA or SA + TG are attached.
2) Introduce two SPST switches.   One to switch C4 & C5 out, for ISO 7637 configuration see attached (the caps would interfere with the transients, good app note from Teseq also freely available on the Internet on this).   And, one to switch in a 50 ohm resistor to the BNC, just for convenience.

Also, another thought to easily test the impact of high DC currents.   It would take some testing to correlate; but, place a permanent magnet with a magnetic field proportional to the DC current of interest on each inductor.   The core would saturate identically as if that DC current were flowing through the inductors.

(Teseq app note http://www.tech-dream.com/Seminar/EMC_Test_Seminar%28Tim2%29.pdf)

[Jay_Diddy_B]

Just a couple more thoughts on the Jay_Diddy_B LISN; particularly if a PCB revision is likely.

1) Add 1.6mm PCB edge mounted SMAs at each end for less error when VNA or SA + TG are attached.
2) Introduce two SPST switches.   One to switch C4 & C5 out, for ISO 7637 configuration see attached (the caps would interfere with the transients, good app note from Teseq also freely available on the Internet on this).   And, one to switch in a 50 ohm resistor to the BNC, just for convenience.

Also, another thought to easily test the impact of high DC currents.   It would take some testing to correlate; but, place a permanent magnet with a magnetic field proportional to the DC current of interest on each inductor.   The core would saturate identically as if that DC current were flowing through the inductors.

(Teseq app note http://www.tech-dream.com/Seminar/EMC_Test_Seminar%28Tim2%29.pdf)

Why would you add the SMA connectors? I cannot find any commercially available LISN with RF connectors for power connections. Smaller LISNs use 4mm banana jacks, larger LISNs use screw terminals or Superior Electric connectors.

I use a Pomona Adapter 1269:



When the LISN is being used it is connected to the DUT with cables that are not controlled impedance.


There is no real need to terminate the unused BNC connector in 50 Ohms. The 10 dB attenuator built into my LISN provides a maximum of 20 dB return-loss.

Here is the simulation used to determine Z in:



And the results show that if left unterminated the impedance is 60 Ohms:



Since the LISN is being used for conducted emissions, not the full range of ISO7637 transients, the CISPR 25 configuration is fine, so there is no need to switch out the capacitor in parallel with the power supply input.

If I was doing immunity testing to the transients, I would want to use a much better coupling capacitor, one with a Y safety rating.

The inductor selection has been 'beaten to death'. Comparison between my LISN and commercially available LISNs, costing 10x to 50x more has shown that they work fine.

Regards,

Jay_Diddy_B

[EMC]

The SMAs have much less error than BNCs.   If I use SMAs I get impedance and phase for this LISN meeting CISPR 16; phase is a touch high at 150kHz.   Variation over more samples will surely produce a phase fail at 150kHz.   But, I have never seen a LISN so well behaved; it is good to 1GHz.   i.e. can be used for conducted emissions to 1GHz to give a "pre-compliance" indication of radiated performance.   ... more LISNs to build more tests to do more results to produce.

PS SMAs and switches don''t have to be populated by eveyrone, just a good PCB mod to do if another PCB cut was done.

PPS Yes there are no RF connectors on any commercially available LISNs.   And test houses hate that; they have to send them out or get special fixtures made.