Help with CE-102 test -- failing with emissions too high

[core7916]

I'm seeking assistance and recommendations for CE-102 testing (MIL-STD 461). When I power on my unit, the emission levels significantly exceed the limit. Since the unit is already designed, I cannot make any internal modifications. Could you please suggest potential causes for this issue and possible solutions?
I am using a 28V DC supply.

 

[berkeman]

Can you say more about this unit? What is its normal function? What IO cables does it have in addition to the 28Vdc power cable? It looks like CE-102 is a conducted emissions test[1] -- are you failing the test on the 28Vdc cable itself, or on another cable connection? What frequencies are you failing?

If you could post any of the test setup pictures, that would be a help. I understand that the details of your unit may be proprietary, though. Please post as much info as you can. Thanks.

EDIT/ADD: Does the unit have a CE sticker on it? What conducted emissions test did it pass (as shown on its Certificate of Compliance)? How does the emissions limit of CE-102 compare to the emissions limit that the unit has already passed?

[1] https://keystonecompliance.com/MIL-STD-461/CE102/

 

[core7916]

Can you say more about this unit? What is its normal function? What IO cables does it have in addition to the 28Vdc power cable? It looks like CE-102 is a conducted emissions test[1] -- are you failing the test on the 28Vdc cable itself, or on another cable connection? What frequencies are you failing?

If you could post any of the test setup pictures, that would be a help. I understand that the details of your unit may be proprietary, though. Please post as much info as you can. Thanks.

EDIT/ADD: Does the unit have a CE sticker on it? What conducted emissions test did it pass (as shown on its Certificate of Compliance)? How does the emissions limit of CE-102 compare to the emissions limit that the unit has already passed?

[1] https://keystonecompliance.com/MIL-STD-461/CE102/

There are signals given as input to actuators along with 28V supply.
It is failing at all frequencies from 10kHz to 10MHz.

 

[berkeman]

I'm not familiar with the CE-102 test, but the conducted emissions testing I've done has been on the AC mains connection of the products that I've tested. That's usually what conducted emissions pertains to -- preventing interference with other devices connected to the same AC mains.

Do you have to test low-voltage and IO cables as well for CE-102? If so, I wonder why...

 

[Ryoko]

This is a kludge fix, but you might try placing a ferrite bead on your power cable near the entry point to your device.

 

[core7916]

This is a kludge fix, but you might try placing a ferrite bead on your power cable near the entry point to your device.

The unit consumes max 3A at 28VDC supply. Would you suggest any ferrite bead value?

 

[DaveE]

Step 1 is usually to identify if the emissions are common mode or difference mode. This will have a big influence on how you proceed. Have you done those tests yet? What is the grounding configuration?

 

[core7916]

Step 1 is usually to identify if the emissions are common mode or difference mode. This will have a big influence on how you proceed. Have you done those tests yet?

Can you please provide details on how can I check if it is common mode or difference mode emission?
I am new to this topic and still learning. It would be really helpful.

 

[Ryoko]

The unit consumes max 3A at 28VDC supply. Would you suggest any ferrite bead value?

The kludge fix I was thinking of was using something like this. They just clamp on your cable. Ideally, you shouldn't need to worry about the current rating because your power and return currents should be balanced. You can also put more than one turn thru the ferrite collar if there's room to increase its effectiveness. This brings up an issue a previous poster mentioned that you need to look at your grounding. If you're grounding thru the cable and thru the chassis of your device, you may have set up a ground loop situation. Ferrite likely isn't going to work well in that situation. (But it's easy to try.) Another gotcha is if you're using shielded cable and using the shield as a return.

 

[berkeman]

@core7916 -- The EMI test lab where you are getting the tests done will have an assortment of those ferrite cable clamps that you can try. I'm surprised that the lab did not already make that suggestion.

 

[DaveE]

Can you please provide details on how can I check if it is common mode or difference mode emission?
I am new to this topic and still learning. It would be really helpful.

Common mode noise is measured in the ground wire, or alternatively with all supply wires in the current probe. It tends to have a high source impedance.
Differential mode noise does not flow through the grounding connections. It tends to have a lower source impedance.
https://www.allaboutcircuits.com/industry-white-papers/emc-basics-common-mode-vs-differential-noise/
https://www.murata.com/~/media/webrenewal/products/emc/emifil/knowhow/26to30.ashx

 

[DaveE]

One common mistake that can make conducted emissions is to have wiring inside your EMC enclosure between the input connections and the emi filters. Then radiated emissions inside the enclosure can be picked up on those wires like an antenna and converted into conducted emissions, but they bypass the filters. A pretty good fix for this is to shield those wires so they look electrically like they are still on the outside up to your filter. This is why there are bulkhead type filter mounting options for larger systems.

Have you asked the EMC techs doing your tests what they think. They've seen a lot of systems that pass and a lot that fail. They don't always know the theoretical side, but they know a lot of tricks, mistakes, and fixes. Often they won't comment unless you ask; like "does this dress make me look fat" kind of thing.

 

[Ivan Nikiforov]

First of all, it is necessary to identify the source of interference - these are the cables connected to the device or the device itself. If interference is transmitted from cables, shielded cables with shield earthing and interference filters (EMC filters) can be used. If the interference is transmitted from the device itself, you can use the following options: 1) Check and improve the grounding system of the device; 2) Place the device inside the screen. At the same time, it may be necessary to shield only part of the device or apply a solid screen in the form of a metal grid.

 

[core7916]

For wiring I have used twisted shielded pair cables. When I am adding a common mode choke I am able to see reduction in emissions.
Are there any other methods which I can try?

 

[Ivan Nikiforov]

For wiring I have used twisted shielded pair cables. When I am adding a common mode choke I am able to see reduction in emissions.
Are there any other methods which I can try?

I suggest you consider using LC filters (band-pass, resonant).

 

[frank8571a]

You are using a 28 VDC Supply.

1. Have you verified that the power source is clean by preforming an ambient scan across the specified frequency range (10 kHz to 10 MHz)? This is accomplished by removing the Unit Under Test (UUT) and substituting a resistive load drawing the same current as the EUT. IAW MIL-STD-461G this scan should be a minimum of 6dB below the 28 VDC Limit.

2. Check the bonding resistance of the EUT to the Ground Reference Plane (GRP). Is the grounding IAW the specification Drawings? MIL-STD-461 requires this to have been verified prior to test and this can affect the EUT's performance.

3. Is the Power cable the correct length 2 to 2.5 meters from the LISN to the EUT and a 10 MHz upper CE102 limit frequency provides less than a tenth-wavelength, this ensures the LISN properly controls the power source impedance.

4. Are you using the correct LISNs? If so are they being correctly employed?

Assuming all the above checks out, Based on the recorded emission profile of the EUT have you attempted to identify the source of the emissions from the device? Are you dealing with a clock, a regulator switching frequency, etc. Knowing the source of the emission may help to determine a path to control it.

 

[tech99]

We do not know on what parameter the unit has failed. So we cannot suggest remedies.