LNK306 EMI issue

Hi gitalleri,

Good day.  Thank you for your interest of using Power Integrations part.

Regarding your question, can you attached your schematic and screenshot of your PCB layout (top and bottom) for me to better understand your problem.  Can you also include your scan so I can see how high it is and figure out where the noise is coming from.

From my perspective with your limited info, it is a big challege passing the EMI with just an L-C instead of C-L-C filter.  Another issue that you might encounter is surge if you don't have cap after the bridge but depends how high is your requirement.  

If you are using off the shelf inductor, can you try to interchange termination if we can see some improvement?  Thanks.

Dear PI-Mallora,

good day to you and thanks for your reply. The engineer team already made a test with adding a C-L-C filter instead L-C filter as I told you. They informed me that nothing change about EMI. You can find the schematic and gerber and EMI test results.

You can see the small areas available to plece all components and the placement of them.

Looking forward for your kindly reply and suggestion.

Thank you.

Best Regards.

Hi gitalleri,

Thanks for the files.  Oh your problem is radiated EMI (30MHz to 1GHz).  Is your conducted EMI (150kHz 30Mhz) good?  With radiated EMI problem you need to add some common mode choke input filter.  Try around 10-100uH input CMC so your efficiency won't suffer.  Is your measurement done in a semi-anechoic chamber?  Thanks.

Dear PI-Mallora,

the enclosd files are the final test result of a second release....exactly the release I sent to you by schematic and gerber snapshots.

The engineer team informed me about this information:

Conducted EMI are ok. Radiated EMI are out of the mask (not so much as you can see in the last file I sent to you. Please not consider the previous files I sent to you about EMI test results. Consider only these).

The measurement are done in a semi-anechoic chamber.

Electrical specifications are: Vin after D1 is 308V, Iin is 4.7mA. Vout is 12V and I out is 40mA almost. The total efficiency is 65% almost.

The temperature on input capacitor C1 is 63°C.

Is the efficiency an issue for the temperature rising? Is this efficiency an issue for radiated EMI? Besides if engineer team use LNK304 instead LNK306 could be better for efficiency and temperature rising/EMI/etc.? If you need more informations please ask to me.

Thank you so much for supporting this design in order to go on mass production as soon as possible.

Best Regards.

Hi gitalleri,

Thanks for the info.  Just add input Common Mode Choke with 10uH to 100uH inductance for your Radiated EMI fix.  Optimize the Common Mode Choke with lowest possible inductance for lower impact on your efficiency.  This added common mode choke will incur additional input loss that is why your efficiency will decrease.  You have better efficiency with LNK306 so stay with it so you won't need to spend additional time testing new part.  

Off the shelf Common Mode Choke are usually bigger so I suggest to use custom made choke.

You can try any ferrite core, just meet the required inductance or you can use below part for your common mode core.  Just wind 3 to 7 common mode turns. 

CORE PN:  RL-1124-1800

MANUFACTURER:  Renco Elect Inc

Dear PI-Mallora,

thank you for your suggestions. The engineer team tell me that there aren't any chance to insert a Common Mode Choke due to a very very small plastic case constraints.

One more question: do you think there might be some changes in the layout of this design that can improve EMI radiated? The layout was done following the notes and suggestions of the datasheets and application notes and taking into account the plastic casing space of the product.

Thank you.

Best regards.

Hi gitalleri,

Your layout is already good but I think you can still squeeze a 10nF Xcap or 100nF ceramic cap before L1 to complete the ∏ (PHI) filter.  See my attached revision if you can implement it.  You can verify first with your existing board manually adding the cap and test the EMI before implementing the actual layout revision.  

I think you also have switching signals at your load side?  If so, can you verify if it is not contributing to the radiated EMI failure? Can you test your board from 12V battery supply instead of 12V from the LNK306 and see if it is passing the limit?  Thank you.

Dear PI-Mallora,

                         the engineer team re-made the EMI test from a certified test laboratory. 

They told me that the EMI test are compliant with the standards. They made EMI test on time with LNK-306 and second time with LNK-304. In the enclosed document you can see the EMI test results with LNK-304 and EMI test result with LNK-306. Both IC are compliant with standards but LNK-304 has lower EMI result about 8dB than LNK-306 EMI result!! The layout is the same and the consumption of the device is 34mA almost in this configuration. 

Could you explain to me why changing the IC from LNK-306 to LNK-304 the EMI result are lower? Is it due to the lower consumption? Or other?

Thank you for your support...the finally IC that will be used is LNK-304 from PI!!

Thanks.

Best Regards

Hi gitalleri,

Its great to hear that you are already passed the EMI requirement.

Regarding your question, the EMI result variation between LNK-306 and LNK-304 is due to different frequency operation of the two controller.  For a same loading condition, LNK-304 have to switch faster compared to LNK-306.  Consumption or load doesn't have direct impact on EMI, it is just the switching frequency changes as the load changes causing different EMI profile.  Layout is a big factor for EMI so you have a good one that is why you are passing the requirement on either controller.  

Thanks again for using Power Integrations' controller.  More power to you all and hopefully we can provide all your AC-DC power conversion needs.  Just post here your needs and we are more that willing to help.  Thanks.