Lpc2000

Hi,

Recently I have been testing some circuits designed around a LPC2106
processor. All of the processors are Revision C with the same date
stamp. Out of ten prototypes that have been made, seven fail for
around 3% of all start ups. One prototype was monitored for 3,500
starts for one failure.

So I tried similar testing with a development kit that had a LPC2106
revision B on board and found the same results. Since the errata
released by Philips has admitted a fault with revision B locking up, I
replaced it with a revsion C only to find that for 4% of starts that
particular development kit and processor will fail at start up.

At this point I don't have feedback Philips, so I am wondering if
anyone else has done similar testing with(out) start up failures?

Naturally, I have tried to fix the problem by making mods to my
circuit, so I can post more background information...

Thanks,

(BTW- I am new to the forum concept, so please excuse me if I make any
faux pas.)