Cheap mains conducted emissions tester?

[treez]

Thanks
Hello,
We are building our own LISN for mains conducted EMC testing….
Will we be OK to make up the 8uF Line to Earth capacitor with 275VAC polypropylene capacitors?.... as follows…….(we will use three 2.2uF’s in parallel and parallel that with another 1.5uF 275VAC polypropylene capacitor)

2u2 polypropylene capacitors:
http://www.farnell.com/datasheets/2202913.pdf?_ga=1.208797074.1679746183.1489787856

LISN schematic (shows the 8uF capacitor on page 2 as C1)
**broken link removed**


We recognise that we should really use Y2 rated capacitors for line to earth connection, but it would be too expensive and need too many of them in parallel for our rough and ready circuit.
We will add 1 megohm resistors in parallel to the 2u2F caps in order to give the “leakage” current in case the isolated line voltage gets induced up to a high voltage with respect to earth.

Another reason that we cant use Y2 capacitors is because the high leakage current would trip the RCDs in our building.

 

[treez]

Schematic for LISN in offline SMPS circuit simulation

Hello,
We are wishing to simulate an offline SMPS with a LISN upstream of it. What would be the best way to add the schematic of the LISN?…should we have the two 1k resistors in there, or should we put 50 ohm resistors across the 1k resistors?....because they are there when the spectrum analyser input is connected.
Schematic situation is attached.

 

[FvM]

I believe, the question is answered in the previous linked schematics.

LISN operation depends on 50 ohm termination of both lines.

 

[treez]

thanks, does the measure of conducted differential mode emissions take place by going across 0ne of the 1k resistors with the spectrum analyser?

 

[FvM]

As far as I'm aware of, there's no differential measurement intended. Full featured LISN has an option to switch the measurement output between L1 and N. The other line is switched to passive 50 ohm termination.

 

[treez]

As far as I'm aware of, there's no differential measurement intended.

Thanks, if its not a differential measurement, then it must be common mode?

- - - Updated - - -

Hello,
The attached LTspice simulation is a buck converter which has the same waveforms as the buck converter in the document RDR-506 by power.com……

RDR-506 report on 1.44w offline buck converter…
https://ac-dc.power.com/sites/default/files/PDFFiles/rdr506.pdf

Page 23 of this report shows the conducted EMC scan of this buck converter. However, the LTspice simulation of this buck converter , using its FFT function, does not give a graph anything like this. Do you know why not?

Pdf schem of LTspice simulation also attached.

However, Another thing is that surely the conducted emissions graph on page 23 of RDR-506 looks wrong? I mean, if you look at the schematic of the buck converter on page 5 of RDR-506, it has a really heavy PI filter consisting of a 1mH inductor and a 4.7uF capacitor either side of it………for a buck converter of just 1.44W rating, that is an extremely heavy filter….how can all the Megahertz frequencies on page 23 of the RDR-506 document be there? If you look at the input current waveform, it is very smooth, there is utterly no evidence of the 10MHz waveforms which the EMC scan of page 23 suggests. Do you know what’s going on here?

 

[FvM]

However, Another thing is that surely the conducted emissions graph on page 23 of RDR-506 looks wrong? I mean, if you look at the schematic of the buck converter on page 5 of RDR-506, it has a really heavy PI filter consisting of a 1mH inductor and a 4.7uF capacitor either side of it………for a buck converter of just 1.44W rating, that is an extremely heavy filter….how can all the Megahertz frequencies on page 23 of the RDR-506 document be there?

To give a sweeping answer: Understanding why the switcher will spread not negligible conducted noise although it's equipped with a seemingly ideal filter is the very start of real EMC engineering...

 

[treez]

To give a sweeping answer: Understanding why the switcher will spread not negligible conducted noise although it's equipped with a seemingly ideal filter is the very start of real EMC engineering...

Thanks, i believe you are speaking of the parasitics of the filter?......and the MHz frequencies are simply going through the interwinding capacitance of the inductor?

The PI filter inductor datasheet does not state its SRF so we cannot work out its interwinding capacitance, but you could guess about 10pF or so.

SBC1-102-211 (1mH inductor datasheet):
**broken link removed**

6MHz “sees” 10pF as a impedance of some 2650 ohms, so it is difficult to see how so much of the 6MHz has gotten through this pi filter.

Would it be too cynical of me to say that the levels of energy at the various frequencies of the EMC test are set far too low? I mean, seriously, there shouldn’t be any need for a pi filter anywhere near as heavy as the one in RDR-506.
Is this a case of big companies making the regulations just miles too strict , in order to stop/impeded smaller competitor companies from coming into the market?

- - - Updated - - -

Also, why no common mode choke for the buck converter on page 5 of RDR-506?
The buck has a source voltage that is rapidly banging up and down from 0V to 330V continuously...surely thats a remedy for common mode noise?
.....................................................................
The pass level for conductive emc is around 45dbuV. That means a voltage of just 117uV across the resistor in the LISN. Thats is surely utterly ridiculous? These sensitivity levels are surely absurd?

[RDR-506] Page 23 shows pass lines for conduted EMC...
http://ac-dc.power.com/sites/default/files/PDFFiles/rdr506.pdf
......................................................................................
In the schematic for the LISN in post #22 above, the spectrum analyser's 50 ohm input reads the voltage across the 1k resistor of the LISN. Does the spectrum analyser amplify or scale this voltage in any way? If not, then how on earth does it read the voltage here to a sensitivity of 177uV? ...because 45dBuV is the pass level and that is just 117uV.

 

[treez]

Do you agree that the DSA815 spectrum analyser, in conjunction with the HM6050 LISN alone, is not enough kit to be able to do even rough conducted EMC measurements on our 150W offline Power Suppllies for LEDs?

It is pretty amazing that in order for a small company like ours to be able to do Conducted EMC emissions testing (QP and AV) for our 150W Offline LED products, the equipment cost alone would be a staggering £4022. And even this is just for a low quality measurement setup without the additional cost of the earthed metal clad room in which the equipment should really reside, and be tightly bonded to. Not to mention, it doesn’t include the ~£800 yearly cost of the calibration for this kit


DSA815 spectrum analyser = £1300
Including the following necessary options for DSA815 spec analyser..
-Advanced Measurment kit (AMK-DSA800) = £423
-VSWR-DSA800 = £400
-EMI-DSA800 = £454
S1210 EMI pre-compliance software = £450

HAMEG HM6050 LISN = £1000

DSA815 spectrum analyser for Conducted EMC measurement…
https://www.rigolna.com/products/spectrum-analyzers/dsa800/

HM6050 LISN:
https://uk.farnell.com/rohde-schwarz/r-s-hm6050-2d/stabiliser-line-z-german-version/dp/2057576

Its quite easy to see how the small companies cannot get a foot in, and go bust trying.

 

[Kajunbee]

Why does the lisn have to be placed on floor or below the product being tested. Also for the testing you are doing do you need anechoic chamber.

 

[treez]

Thanks but anechoic chamber is only needed for radiated emissions test....unless of course you have a near field test kit, for which no anechoic chamber is needed.

LISN needs to have a low impedance connection to earth, so it tends to be on the metal earthed sheet on the floor where it can be directly bolted to it....or at least connected to earthed sheet by a thick braid with low frequency to RF signals.

Truly , one really needs an earthing rod driven into the ground and the LISN bolted to that as close as possible to the point where it goes into the ground...and you need to be near sea level so you have a high water table and low earth impedance, i think above 10,000 feet, earth connection is too high impedance for good emc lab setup....because the earth impedance is too high due to its dryness......if you live in the desert , and all you have is dry sand under and around you....you cannot get a good earthing rod connection to earth, and so again, emc lab setup cannot be done.

 

[Kajunbee]

Yes, I was wondering about that also, how the water table effects grounding. And its not just as simple as wetting the earth around the area. I'm guessing your device is class b. Isn't the compliance standard different for devices used in both home and industrial. Seems like I read that somewhere but I couldn't find the site again.

 

[treez]

Thanks, indeed you'll notice that EMC test houses are at sea level most often...you wont find any up mountains or hills.....and certainly not at the top floor of a tall block of flats.

 

[Kajunbee]

As far as I know plastic pipe is not a good conductor so the option to connect to a water pipe will become increasingly less of and option also. Doesn't do much good to connect to a metal pipe in the building if the pipe in the ground is plastic I would think.

 

[treez]

Well yes, of course, the setting up of an EMC test lab, and the secrets of getting a good ground connection in it, are very closely gaurded industrial secrets.
But basically there will be a very long big thick metal grounding rod going very deeply into the earth somewhere....at least far enough down to get into the water table.

 

[FvM]

All you need for a reasonable precompliance measurement of conducted emissions is a table covered with a metal sheet and the LISN grounded to the metal.

 

[treez]

Thanks, but as you know, the LISN must have an earth connection, so it can take part in measuring common mode noise....but what if the builders forgot to put in an earthing rod into the building, and the earth wiring of the building, is not actually properly connected to earth?
As you know, a connection to earth can only truly be made at or near sea level when a large metal grounding rod is used which is drilled deeply into wet earth.

The thing about earth is that as you know, it is literal... you really do have to be well electrically connected (low impednace connection) to the huge vast mass of mud that sits beneath us all.

If the building we are in does not have a properly installed earthing rod, then those emc measurements will not be accurate, since the common mode noise will not be properly measured.

 

[mtwieg]

No, an earth connection is not "literal." It's just another arbitrary reference point.

Think of it this way. If you did have a nice shielded room, that shield would normally connect to "earth" via some means outside of the shield. But inside the shield, that connection is totally meaningless, because the shield (which is directly connected to on of the LISN terminals, usually its own chassis) completely masks the outside world from the inside. Whether it's actually connected to an earthing rod or not has no effect on observations conducted within the shield (for high frequency signals, anyways). That's why we use a shield+LISN in the first place.

FvM is suggesting that in the absence of a nice shielded room, you can get a reasonable good approximation by using a large conducting screen below your test setup. So long as most of the electric field lines terminate on that screen, it's a reasonable substitute.

 

[treez]

Thanks, so supposing the wiring in the building has earth wiring which does not actually connect to real earth at all....then that would have no influence on common mode noise measurements by a LISN in that building?

 

[FvM]

The metal sheet and LISN ground will be normally connected to the regular protective earth.

If the power lines in your building are heavily polluted with electrical noise, a separate ground connection may reduce the background noise level. You should think about such measures if an "empty" measurement with unpowered DUT shows already a high level near the limit values.