Reducing EMC emissions

[Pha5e]

When doing EMC pre-compliance tests for radiated and conducted emissions, what is the best method to identify the source of emissions exceeding limits and to reduce them?

One recent example is from a motor drive and emissions in the low MHz are exceeding the conducted emissions limit. There is a line filter in the system.

Is the position of the filter critical or should a filter anywhere on the supply in a system prevent any noise getting back onto the supply?

 

[KlausST]

Hi,

what is the best method to identify the source of emissions exceeding limits

Look at the frequency and find out which part of your application produces this frequency. Mind that there are overtones.

Klaus

 

[Easy peasy]

most RFI comes from the IGBT turn on ( motor drive ) or the aux supply

the filter should be at the AC input else it can pick up higher freq, 50MHz + if fitted to far from the AC input ....

 

[Pha5e]

most RFI comes from the IGBT turn on ( motor drive ) or the aux supply

the filter should be at the AC input else it can pick up higher freq, 50MHz + if fitted to far from the AC input ....

Thanks, the conducted emissions frequencies above the limit are less than 10MHz. Would filter placement still be as critical?

 

[D.A.(Tony)Stewart]

You can measure emissions with a shorted 10:1 scope probe with the ground wire wrapped around a single wire compared with LN together. Then connected to a scope then SA.. There are more accurate methods, but this is easiest.

Show the signals and line filter differences from this. Not all line filters are the same. The conducted noise can also radiate.

The 2nd has DM chokes on the left and CM choke on the right.

 

[Pha5e]

You can measure emissions with a shorted 10:1 scope probe with the ground wire wrapped around a single wire compared with LN together. Then connected to a scope then SA.. There are more accurate methods, but this is easiest.

Show the signals and line filter differences from this. Not all line filters are the same. The conducted noise can also radiate.
View attachment 172927
View attachment 172928
The 2nd has DM chokes on the left and CM choke on the right.

Thanks for the advice regarding locating the noise with a scope probe. The emissions measurements were taken using a LISN.

In your filter networks above, could you explain how each separate stage helps with reducing noise? Is the spacing of the stages in the SMPS filter critical?

 

[D.A.(Tony)Stewart]

Thanks for the advice regarding locating the noise with a scope probe. The emissions measurements were taken using a LISN.

In your filter networks above, could you explain how each separate stage helps with reducing noise? Is the spacing of the stages in the SMPS filter critical?

location needs to be near outlet, then you need to define if the noise is common mode CM or differential. DM then the solution depends on L/C ratio and LC product compared with noise impedance and time interval of noise squared. for LC or attenuation ratio of filter

Also show circuit of noise generator with values and filter with the emission spectrum

Testing and specifying the problem is hard. The solution then is easy.

 

[Taihung]

This field is hard to say when fix only with filter can pass EMI.
EMI like water in river. If you prevent this frequency point, other will rising up.
So, if only change only 1 position, it is very hard to pass.
You have to focus to "core" which generate noise.
You should try many step & measurement to define noise generator(component) then have way to treat it.
Layout is very important, you have to check current loop, dv/dt, di/dt and confirm parameters of each compponent in circuit. If you don't clear about any component in circuit, don't try to pass this. It is lucky with small power range, but with higher power range don't have lucky.

In this case the value of component is not as simple display.

PFC + FLYBACK 40W design
This is not good tunning component value:

This is after tunning correctly:

 

[D.A.(Tony)Stewart]

When doing EMC pre-compliance tests for radiated and conducted emissions, what is the best method to identify the source of emissions exceeding limits and to reduce them?

One recent example is from a motor drive and emissions in the low MHz are exceeding the conducted emissions limit. There is a line filter in the system.

Is the position of the filter critical or should a filter anywhere on the supply in a system prevent any noise getting back onto the supply?

Show photos of your system, EMI problems and frequency of high emitters for dI/dt and dV/dt with ground paths if you want a better answer

 

[mtwieg]

Troubleshooting EMC issues requires a detailed description of the device, any shielding and earth connections, cable routing, etc. Either photos or detailed diagrams.

Ideally, all EMC would be eliminated at the source, but obviously that's not always practical, otherwise nobody would need EMI filters.

 

[D.A.(Tony)Stewart]

2nd harmonic 110kHz of asymmetric 55kHz PWM was dominant at source loop area from local resonances. Lots of ways to sniff this and tune it. Also CM filter needs to be high Z to suppress into X caps to attenuate that needs to be measured in Diff and CM modes

 

[biff44]

Thanks, the conducted emissions frequencies above the limit are less than 10MHz. Would filter placement still be as critical?

placement critical?
WHERE it is placed is not important.
The quality of the ground, or earth ground, that you are trying to use is critical (as in any parasitic series resistance or series inductance).
For that reason, some filtering is often done in balanced configuration....like you have a standard DC motor with a + and - lead. You put the filter caps across the leads, and the series inductances in series with the two leads, without using any sort of board ground, or case ground at all.

BTW, using a balanced inductor, like shown here, allows a much smaller ferrite device for a given DC current, since the + flux cancels the - flux in the magnetic fields, and the ferrite does not saturate as easily with a 10 amp dc load...for instance