Author Topic: Conducted EMC scan shows Live and Neutral scans in curious situation  (Read 2177 times)

0 Members and 1 Guest are viewing this topic.

Offline ocsetTopic starter

  • Super Contributor
  • ***
  • Posts: 1516
  • Country: 00
Hello,
We have a 40W offline LED driver which is linear regulator based.
It failed EMC when it had no filter as per the attached scan.
Then we added the 100nF  AC input capacitor (between live and neutral)  and it passed, as per the attached scan.
Live and neutral scans are shown in each case. It  almost looks as if adding the filter capacitor has meant some of the emissions from the Live has been able to “jump” over to the neutral, resulting in a marginal pass. Do you know what’s happening here?
 

Offline Phoenix

  • Frequent Contributor
  • **
  • Posts: 420
  • Country: au
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #1 on: September 21, 2017, 06:24:51 am »
It looks like whatever was taking only the L1 as a path is now able to be shared across L1 and N. The 100nF capacitor is quite large and low impedance. One way you can think about it is that as far as the EMI is concerned the active and neutral now look like they are shorted at both ends.

I think a CM choke on the line side of the capacitor should be good to bring it down substantially, keeping the capacitor.

Why it's so different between L1 and N to start with, I can't tell!
« Last Edit: September 21, 2017, 06:33:02 am by Phoenix »
 
The following users thanked this post: ocset

Offline jbb

  • Super Contributor
  • ***
  • Posts: 1128
  • Country: nz
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #2 on: September 21, 2017, 08:36:37 am »
Sounds like you had a differential mode problem on th phase, and the 100n cap turned some of it into differential common mode. You could try a basic inductor on the phase before the 100n cap. It might be cheaper than a CM choke.

EDIT:  differential common
« Last Edit: September 21, 2017, 10:21:53 am by jbb »
 
The following users thanked this post: ocset

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #3 on: September 21, 2017, 09:09:21 am »
You sure that is a linear reg?
Something is very, very wrong if it is, as it looks to be honking at ~60Khz and triplen harmonics of that frequency.

Concur that this is a Differential mode issue being converted to CM by the cap, LC lowpass would probably solve it (Inductor on the PSU side of the 100nF), but you really want to get to the bottom of where that 60KHz is coming from in a linear supply.

Regards, Dan.

 
The following users thanked this post: ocset

Offline ocsetTopic starter

  • Super Contributor
  • ***
  • Posts: 1516
  • Country: 00
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #4 on: September 21, 2017, 06:52:22 pm »
Thanks, it is a linear regulator, banks of LEDs get kind of switched in and out as the mains half cycles rise and fall.
It does surprise  me that  so much common mode noise has arisen, but then the pcb does lie on a earthed heatsink.
 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13695
  • Country: gb
    • Mike's Electric Stuff
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #5 on: September 21, 2017, 07:03:01 pm »
Thanks, it is a linear regulator, banks of LEDs get kind of switched in and out as the mains half cycles rise and fall.
It does surprise  me that  so much common mode noise has arisen, but then the pcb does lie on a earthed heatsink.
Could it be that the switching is faster than it needs to be ?
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 
The following users thanked this post: ocset

Offline ocsetTopic starter

  • Super Contributor
  • ***
  • Posts: 1516
  • Country: 00
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #6 on: September 23, 2017, 08:30:29 am »
Quote
You sure that is a linear reg?
Something is very, very wrong if it is, as it looks to be honking at ~60Khz and triplen harmonics of that frequency.
Thanks, yes it is only a linear regulator…..maybe the offtheshelf control chip has a microcontroller inside it which works at high frequency, I don’t kknow. There’s certainly no Switching power supply there…..(other than the mains rectifier bridge which is sometimes construed as a switching device)

Quote
Concur that this is a Differential mode issue being converted to CM by the cap,
Thanks, this is groundbreaking, because if  differential mode nose can be converted to common mode noise via a capacitor, then the reverse must also be true…and if it is , then that’s great , because we’ rather use diff mode filters than common mode filters,  because  diff mode filters are easier to do as SMD only.
Quote
Sounds like you had a differential mode problem on th phase,
Thanks, surely a differential mode problem, by definition, would manifest itself equally on phase and neutral (I’m assuming by “phase” we both mean “live”)
There seems to be a multiple concurrence that as far as common mode noise is concerned, capacitors have two very distinct functions…
1….They sometimes have a filtering effect, as in the capacitor of an LC filter
2…They sometimes have no filtering effect, but act so as to “channel” the  noise to different places.
Is this true?
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #7 on: September 23, 2017, 08:40:09 am »
Appears to be a superposition of diff and CM.  Otherwise known as normal mode.  One wire acting alone, as it were.

Seems the thing to do would be, add a choke in series with the offending wire.  See what what does.

QP is also really high above average.  Seems to imply a very peaky signal.  Perhaps snubbers to reduce dV/dt are in order?

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline ocsetTopic starter

  • Super Contributor
  • ***
  • Posts: 1516
  • Country: 00
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #8 on: September 23, 2017, 08:43:12 am »
Ref the discussion about DM noise being converted to CM noise (and vice versa), the following thread tells that this is not possible...

During this thread about power supply common mode noise problems….
http://www.edaboard.com/showthread.php?t=368189
…there is discussion, especially by the expert Mtwieg, about common mode emissions, and even discussion (post #34) of  differential  mode noise being transformed onto common mode noise.
Post#7 of the above thread sets the scene, and states the theories about common mode emissions  that later in the thread get totally trashed.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Conducted EMC scan shows Live and Neutral scans in curious situation
« Reply #9 on: September 23, 2017, 08:58:18 am »
Ref the discussion about DM noise being converted to CM noise (and vice versa), the following thread tells that this is not possible...

Well, no...

And anyway, opinions are meaningless, and particularly-ill-informed ones abound in this field.

How fortunate that we should work within a field where every single possible thing we might do, can be fully quantified, measured, analyzed and solved for.

It's so much more the tragedy, then, that so many willfully avoid analysis, preferring baseless opinions instead.

In short: learn the equivalent circuit of the system in question.  Then it will be immediately apparent where capacitors and inductors will be useful, and where they won't.  You will have no need of peoples' silly opinions.
 Or, alternately, if you don't have a working understanding of the system's equivalent circuit, the same principle applies in reverse, i.e., testing a capacitor or inductor in a given location gives you positive or negative information regarding the amplitude and impedance at that node, and by testing enough locations, you can reconstruct the equivalent circuit that way.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: ocset


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf