Hi,
i am having serious radiation issues with a TFP401-board. The disturbances are spread around a wide frequency range from 150MHz to 1GHz and change with screen content, some peaks go up to 50dBµV. Which is way too much.
My board setup is a HDMI-LVDS converter as described in the application report SLLA325A. All decoupling capacitors / ferrite beads are used as described in the datasheet.
I can trace my problems to the HDMI-Connection by disabling all outputs of the TFP401 by using PDO. Also, if HDMI is disconnected, there is no radiation (so my power supply is clean).
By using a near field probe, i can confirm the disturbances are radiated by the HDMI-cable or the connectors. I tried several different HDMI-Cables and sources, but the Problem persists. I also tried several grounding concepts, with only minor changes.
The connections between connector and TFP401 is a direct connection (no parts, no test points), traces are matched in length (with <0,1mm error). Due to a CAD-error, i calculated the differential impedance as 88Ohm. This low, but should be within the HDMI spec.
My questions are:
Can this small Impedance mismatch cause so much radiation?
What else can cause this?
I am pretty much stuck here, as far as I know, the issue can only be board-traces or the connector (Molex, 0471510001).
