If it were me, I'd start by setting up some tests
to determine what exactly
the protocol analyzer is doing. If it really is dropping 0 bytes, then I'd
Well, I see 2 ways forward, in that I have found 2 things that look odd...
One is to investigate the serial link between the VAX console processor and
the TU58.
The other is to see just why the data from the read amplifier looks odd and
why the TU58 drive runs tapes off the spool (meaning it is not seeing the BOT
marker).
Intuitively I feel the latter is more likely to be the real problem.
probably hack up a home grown capture using a couple
of serial ports.
If the controller and the VAX really are speaking the right protocol to
each other, then it's time to worry about the correctness of the data. On
the other hand, if the controller really is not sending all the bytes it should,
I'd check for bit rot in the EPROM.
The TU58 ROM checks against a dump I made of it years ago (OK, that dump
might be bad). It also checks against a dump of ROM in my other TU58 (which
has bad rollers I have not repaired yet). The power on self test of the TU58 does
a checksum of the ROM (OK, not a very good test) and that passes. I think bit
rot is unlikely given that.
Furthermore from a read of the source code it appears the TU58 builds the
response packet in RAM and sends it. I can think of faults that would cause
it to be truncated. But not a fault that causes it to leave out bytes from the
middle of it.
-tony