Lpc2000

Hi,

> It's not easy and there are only well known general rules. Use emi 
> filters on high speed lines where possible (e.g. murata EMIFIL or 
> similar), use ground shielding as much as possible. Avoid current 
> flowing through shielding / ground planes. Add external shield if 
> nothing else works. It's tough task. I had similar issues with smaller 
> MCU (x51), which was solved with EMI filters and grounding with 4-layer PCB.
> 
> M.
> 
> tsvetanusunov wrote:

> > we use double side board, the problem is not in the crystal
> > definitely as if PLL is not enabled this noise is not generated
> > IIRC PLL works on 150-180 Mhz and the radio iterference is caused
> > probably when PLL frequency pass through 108Mhz FM range
> >
> > Best regards
> > Tsvetan

Your best shot I think is to try and isolate *where* the VCO signal is radiating,
little chance it would radiate so strongly from the die !
Probably like Fvco/2 or thru some divider, it is leaking onto a Vdd line (or also Vss)
and radiating from there.
The best attack - if this signal should be a problem - is to bypass the proper track(s)
as cose as possible to LPC to a solid GND, with a low inductance return path to the chip.
Worst case you must find where the leakiest signlas are, and terminate them.
Remember, at 100-200 MHz, 100 nF is a _very_ unsuitable bypass capacitor.
Bypass caps should be as low as possible in value, I doubt that EMI filters will do much in
this specific scenario, chances are that by the time the sigs get to the EMI filter, they've
already been radiating galore anyway. They must be bypassed to GND, and their energy
returned to LPC2000.

Good luck,
Kris


[Non-text portions of this message have been removed]