Thanks for the reply, the reminder about the ICD holding the chip in reset was
what was needed.

User error.  I had neglected to break the connection between MCLR and 5V (via a
1K resistor).  Once I had MCLR connected only to the ICD, things went much
better.

Thanks for the help.  Though chatty at times, this list (meaning its members and
administrators) is invaluable.

Kevin


Barry Gershenfeld wrote:
>
> If it's wired properly the ICD can hold the processor in reset.  At
> that point RB6/7 and all other pins should be set to input.  Though
> I'm not sure how the ICD gets control other than through the debug
> code that it installs.   But try hitting the Erase button to get
> rid of your program, or make sure you have Erase All before Program
> checked when you try to update with your new code.  (And make
> sure you have your new code loaded :-)
>
> That would be nasty if you could lock yourself out like that...I don't
> think I'll try that right away (though I almost did yesterday--until
> we realized we could enable Port B pullups instead of driving
> them high).
>
> Barry

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.