RS485 communication causing Conducted emission failures

[ku637]

Hi,

I m working with a communication board
operating on 24V input voltage-->Step Down Converter (fsw=700kHz)--> two outputs 1)5V/0.4A 2)3.3V/0.4A

The board has got an Isolated RS485 section..a small wattage isolated power module(fsw=20kHz to 90kHz) creates an isolated 5V from the Step down converter 5V output & an isolator IC(ADUM1401A) is used to isolate the serial data from the processor before going into the RS485 Transceiver..

The board with the housing passes Class B limits with enough margins without the RS485 communication happening..
But when the communication happens, the 24V power line is facing multiple disturbances peaks from 150kHz to around 600kHz..

Any idea how to debug this kind of problem..??

How to identify if the noise is common mode or differential mode in origin..if that is identified will it help to solve the problem..?
Im using LISN EMCO 3816/2 in combination with an Agilent Spectrum Analyser.

Thanks for any help,

 

[udhay_cit]

Post your schematic diagram....
The attachment may help you to select a good filter. i Got this from a seminar...

Regards
Udhay

 

[snail888]

Hi,do you housing is metal or plastic?For disturbance frequency 150K to 600K,it seems differential mode for it's so low spectrum.

 

[pasaajan]

You can do 1 experiment on your seat. Try to measure the harmonics frequency of RS485 isolated supply with & without comms. If you are using Tektronix (CRO), put in Math mode, in Nyquist waveform can see interesting frequency related to CE test. 150kHz to 600kHz freq may be getting prominent with & without RS485 comms, this require little patience d approach.
If you are able to catch the problem, 1 quick fix is try a RC over RS485 secondary filter (snubber).