RE: Signature Analysis/CAT Box question

On Sat, 9 Oct 1999, John Robertson wrote:

> I suspect you need to insert a NOP modified CPU to get the signatures, I
> believe that the CAT box does the NOP in this case, but I don't remember
> noticing that in the docs on my CAT, probably you are just doing a write to
> an area of memory...

        No, you don't need to do anything of the sort. A "NOP modified
CPU" is just a socket with the data lines tied to the NOP instruction,
right? You just need to write this to the Data Bus with the Fluke. That
CAT Box does the same thing -- It just writes NOPs to the Data Bus at the
appropriate times.
 
> Signature is normally the reading of a signal that is clocked between a
> Start and Stop point on either a rising or falling edge, the bits are
> counted and the display give a reading of some magic done on the count. The
> letters in the Sig are arbitrary letters that could show the count from 10
> to 15 on a 7 segment LED display. That is about it for my understanding of
> the process, and I probably have it wrong... sigh... but it works!

        Yup, that's the basic idea. The signature is just a CRC of the
data between the start and stop times. The CRC polynomial can be
extracted from the CAT Box schematics (look for the LS166 (I think) shift
registers with feedback XOR gates.) You can get the CRC polynomial from
that. You just need to program the Fluke to implement that function on
the probe data....Or just use a 5004.

Joe
Received on Sun Oct 10 11:16:28 1999