EMC troubles: identifying CM vs DM vs GND-referenced noise

[uer166]

Currently debugging some conducted EMI for class B compliance. This is a 240V split phase system with a internal PSU and a whole bunch of complicated parts. I've been trying to identify whether the noise is DM or CM by placing chokes in the DUT-side of the mains in various configurations, but I don't quite understand the results. What I see is

• Choke in the PE (GND) provides some attenuation
• Common mode choke on L1 and L2 provides the exact same attenuation as the one in PE only
• Choke in L1 provides most attenuation (this is the line that's being measured in the traces)
• Choke in L2 provides no attenuation whatsoever for L1 measurements
• The waveform shape looks like an LC filter to me, but I have been unable to identify where the resonant tank is. In one system it's at 1MHz, in another it's at 300kHz (see attached)

I'm waiting on a LISN-mate to hopefully get a better understanding, but does the above make any sense? E.g: if it's common mode only noise, then why does a CMC not provide as much attenuation as a choke in L1? Yet, if it's differential mode, why does a choke in L2 provide no attenuation whatsoever?

[Hawaka]

Hi,

If you have access to 2 LISN, I would recommend to get a Schwarzbeck CMDM 8700. This device let you separate the CM and DM noise. A very nice tool, but not everybody has 2 LISN laying around to use it.

Regarding what you are measuring, here is what I think:
- Choke on PE should be avoided
- If the CM choke is not providing enough saturation, it might be saturating, or simply does not provide enouh attenuation at 300kHz.

From what I see it could be a DM noise coming from L1 but not with L2. Maybe a single phase load ?

[jonpaul]

Hello: Many techniques exist to isolate DM vs CM noise but hardly ever needed.  Most PSU noise is CM not DM.

Any good EMI filter, properly specified for the Vline, I and ragion, has a CM choke and  VDE X and Y caps, it will  will attenuate both CM and by way of choke leakage L, the DM noise as well.

If you look at a CM choke, they should specific both CM L and DM L.

Bon chance

Jon

[uer166]

From what I see it could be a DM noise coming from L1 but not with L2. Maybe a single phase load ?

You were right on the money here, there was a Y-cap between L1 and GND at the end of a really long (10m) output cable, that was resonating with the cable inductance. The fact that it was a single phase DM load made it look weird in my eyes, but the scans make sense now to me. Without the Y-cap it totally disappears are expected.

[uer166]

Hello: Many techniques exist to isolate DM vs CM noise but hardly ever needed.  Most PSU noise is CM not DM.

The problem is when you have a large overly complicated system where you're trying to figure out where it's coming from in the first place. Of course we have fully specified CM chokes (2 in series actually) on the PSU, and the PSU itself passes conducted on its' own, but when you put it in a large+dense box with a bunch of other crap weird things tend to happen.

[Cubrilo]

Sorry for a late post, I've found it very interesting

• Choke in the PE (GND) provides some attenuation
• Common mode choke on L1 and L2 provides the exact same attenuation as the one in PE only

Too bad you didn't wrote which line is for PE and which on L1&L2 (there are green and yellow).

[Jay_Diddy_B]

Uer166,

This thread describes the operation and design of CM/DM separator.

Link: https://www.eevblog.com/forum/projects/diy-dm-cm-seperator-for-emc-lisn-mate/msg3117396/#msg3117396

You will a dual LISN or two identical LISNs.

Regards,

Jay_Diddy_B

[uer166]

Uer166,

This thread describes the operation and design of CM/DM separator.

Link: https://www.eevblog.com/forum/projects/diy-dm-cm-seperator-for-emc-lisn-mate/msg3117396/#msg3117396

You will a dual LISN or two identical LISNs.

Regards,

Jay_Diddy_B

Neat! Though it's been 2 years since my post and I'm glad I'm no longer dealing with that system/company!