Part Number: DP83825I
I am facing radiated EMI Problems in the Higher MHz to GHZ Sections. I was able to measure with a near field probe in the EMC lab. There are big spikes around the PHY.
I would like to ask you for an analysis of the Schematic and Layout around the PHY.
If I shall s end you more information please let me know.
Kind regards,
Cédric Trachsler
Hi Cedric,
The clocking circuitry in the schematic looks ok. I see there are smaller peaks every 25MHz and larger ones every 50MHz in the near-field probe. This makes sense as the 25MHz would be your XI/XO while 50MHz is a combination of the XI/XO harmonics as well as the 50MHz RMII clock. I suspect that by reducing either of these circuit emissions, you would see improvement at 50MHz harmonics.
Looking at the layout, the XI/XO circuitry looks good but could be optimized further. Some improvements you could make here would be to add GND fill and more stitching vias in the areas circled in blue. Additionally you can mount the load capacitor GND pads directly onto the large GND plane rather than connecting these through a trace (Circled green)
The 50MHz clock trace is difficult to follow from the PDF but this is the path I believe it is taking:
In this case, adding GND fill and increasing the number of stitching vias around the trace could help reduce emissions. Additionally if there is any way to reduce the length of the clock trace(s) it would help to reduce emissions. The ideal routing would have the RMII traces entirely on the top layer rather than routing them through two sets of vias.
Best,
Shane
