noise reduction

Hello Denis,

Thanks for your question. Adding ferrite beads to VISO and GND will improve the typical output ripple. I've attached the recommended applications circuit for the ADuM5020 for reference. We do not characterize min/max ripple for these devices, however with this filtering typical ripple at 90% load is ~20mVp-p, as shown in figure 10 of the ADuM5000 datasheet. Will this improvement meet your specifications? 

For EMI performance, I would recommend using the ADuM5020 rather than the ADuM5000. The ADuM5020 is capable of meeting CISPR32 Class B emissions with the implementation shown below.

Best Regards,

Tommy

Hello, 

I am looking into using the adum5020 but it looks likes it is not footprint / pin compatible to the adum5000.  

This device is followed by a couple LDOs, which have decent PSRR but that drops off after 100kHz.  Even it the PSRR is -30dB down that is still better than nothing.  Just looking for a kitchen sink approach for noise suppression flexibility.  I will add in the ferrite beads footprints.

dennis 

Let me ask one more follow up question that is a little more specific to the application. 

For the case of using the ADUM5000 for 5V in/out, if the 5V source is a brick type SMPS device with an Fosc around 580kHz, it will be noisy.  This will add onto the noise already stated in the adum5000 datasheet.  Any suggestions for noise reduction here in between the two circuits?  I was going to go with a PI circuit (shunt C > series L > shunt C) that has a reasonably sized inductor.  I would think that designing the cutoff frequency to be at least 1/10 of the Fosc would be a decent start/goal.  It would give reasonably sized inductor value (not in the milli Henry region) and resistive loss would not be too bad.   

thanks

dennis

Hi Dennis,

Apologies for the delayed response. 

The design of the pi filter you describe sounds good. I would recommend that approach.

I think that combined with the LDOs post ADuM500 will give good noise reduction here.

Regards,

Tommy