Other Parts Discussed in Thread: TFP410
We have been using the DVI-transmitter TFP410 for quite a long time, but in more or less all designs we have had EMI-problems.
I should say that the implementation is "by the book" - routing, LC-separation of supply parts, groundplane design etc.
We can run a monitor (or other receiver) with more that 10 meter cable in 1920x1080 @ 60Hz. No jitter, no bad pixels.
This would not be possible if we had bad signal integrity - crosstalk, impedance missmatch etc, or one part in a diff.pair shorted to ground.
When the TFP410 is running without TMDS-termination (no receiver connected) there are very low emission levels.
The problem exists only when the TMDS-output-interface is active. The emission is TXclock, Txclock * 3, 5, 7 ...)
The Tvdd supply is bypassed with several 0603-caps (100n, 10n) so transient current from the outputdrivers will not introduce much 3V3-ripple.
I have also made tests with added common-mode-chokes on all diff.pairs, but the problem with EMI was not improved much.
/Pär
