Lpc2000

This board does not use RS232 control lines to reset the chip and set
P0.14  Using LPC21ISP.EXE and hacking the source a bit to give me a
second to reset the part AFTER the port is opened, but BEFORE it sends
the first '?' gets me much further.  This utility manages to
synchronize, set the oscillator value, unlock, read the bootcode
version, read the part ID, and then issue the first write command..
BUT.. Then it fails because the echo-back of the UUEncoded write data
has a couple of bit errors in it.  When I comment out this failure
detection, I get a CRC error from the LPC at the end of the block,
meaning that it is probably seeing bit errors as well.  On the host
side, the errors are generally 1 bit per symbol.

With clock jitter, I would expect all kinds of things to give out before
a 9600 baud RS-232 connection.  And, as far as I can tell, the
oscillator circuit is OK.  Looking at it with a scope is a clean sine
wave.  We're going to put a spectrum analyzer on it, but I'm betting
this is a power-supply decoupling issue, either with the LPC or the
MAX232..

I've got a bag of caps, and I'm not afraid to use them..