Are any of the DECUS 12 bit SIG newsletters online? I just found parts
of two of them in my attic (#39 - March 1980 and #40 - Summer-Fall-Winter
1980) and they have some interesting comments about TU-58s and 12-bit
machines.
It's interesting considering the recent activity regarding TU-58s and
emulators. According to (I think) Jim Van Zee, formerly of Laboratory
Data Systems of Seattle), being able to send a break is critical for
reliable operation of a TU-58 (and gives the exact reasons). The M707
and the M8650 cannot do this without modifications (but he does describe
the mod to the M707 and says that it can be done to the KL8E (M8650) but
not the KL8EJ (M8655) or KL8A).
The author goes on to describe a handler he wrote for OS/8 - a non-
system handler. It adds "DTU0" and "DTU1" of 682 OS/8 blocks. He
also
writes that it is impossible to write a system handler without an
external circuit or ROM code (a-la 8K TD8E + ROM) to calculate checksums.
His hardware solution is 4 chips grafted onto an M8650 (one of which
is an Intel 8748 microcontroller!)
Additionally, he mentions a virtual TU-58 server written for the VAX
by Jim Gladden, and describes a submission to DECUS of a non-TU58
serial line device handler for the VAX (ASCII files only) that *may*
be submission number 8-921 (my handwritten notes in the margin).
So... if anyone wants to hang a TU-58 (physical or virtual) off of a PDP-8,
we have somewhere to start.
-ethan