Barry Gershenfeld wrote:
> I've been developing with an ICD2 for years and have made peace with
> the unit.  The newly released ICD3 seems like the logical upgrade
> path, but your talk about the RealICE makes me wonder if I might
> benefit from that. Ironically, because I learned early on that I had
> little or no use for a true ICE, I simply ignored the existence of
> anything labeled "ICE".  Maybe you can cite an example or two of some
> problem solving you did using a feature that the ICD2/3 lacks.

As I said, I haven't used a ICD3 and don't see any reason to.  I have in the
past occasionally had to use a ICD2 and always found it a pain.  Various
things would just randomly not work.  Sometimes it wouldn't connect.
Sometimes it would never stop at a breakpoint you know it got to.  Sometimes
it wouldn't stop the processor when you hit F5, even if you waited 10s of
seconds.  Sometimes it would get into a state where MPLAB hung and you had
to CTRL-ALT-break and kill the process.  To sum up the ICD2, I think "flaky"
is the best description.

RealICE is still a in-circuit debugger with the associated advantages and
disadvantages, but it so far seems to work as intended.  The most annoying
thing about it is the skidding, which is the case with all the in-circuit
debuggers.  I get around that by a assembly time switch that adds NOPs to my
GCALL and MCALL macros so that I can stop on a subroutine call without going
into the subroutine.

The simulator is the best debugging environment as long as your inputs are
simple enough to be simulated.  Several times I've written programs that
process a higher level description of input to the PIC and then generate a
stimulus output file.

When the input can't be reasonably simulated, the next best environment is
the ICE2000 or ICE4000.  Both are very nice and essentially give you a front
panel to the PIC.  There is no break point skidding, and the trace
capability can be useful for figuring out how something got messed up when
the first symptom is well downstream.  You don't need it often, but when you
do it's really nice to have.

The RealICE has some trace capability too if I remember right.  I haven't
needed it yet though.

> Though I might be inclined to buy the "biggest box they sell", my
> experience with obtaining a PM3 showed me that isn't always the case.

I do PIC development professionally, so trying to save a couple $100 on
tools I may buy 3 copies of is silly.  Perhaps the ICD3 is very nice, I
don't know.  I know the RealICE is the best of the in-circuit debuggers, and
it's not much more money than a ICD3 so there is no point in the ICD3 for my
purposes.  I think its intended as the RealICE for hobbyists.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist