
2/29/88
Rob Horning

I worked on the first pass ERS for Gummby?.  We took a large part of it
from Topgun.  The CPU and CMUX do not have any documentation and so I had
to go around and ask people questions about details.

I prepared for and went to Guadalajara to design review the Kahlua memory
board and the Tabasco memory controller.  No major problems were found.
Some of our suggestions are listed below:

  -  Synchronize the power on signal in Tabasco.  They may not really need to
     do this, but it can be very difficult to verify that you do not have a
     problem.

  -  During reads of IODC the memory board would drive the NIO and float the
     inputs to the AS245's.  We pointed out the dangers of doing this and we
     came up with an easy fix.

  -  We pointed out some potential problems with having some chips powered by
     +5 and others powered by +5S.

  -  We suggested that they run more test vectors where refreshes hold things
     off and are held off.  They were not doing a lot of this.

  -  They had a parts placement for the memory board and we pointed out some
     problems with it.

