DoN. Nichols wrote:


Hmm ... HP made a nice front-loading self-threading tape drive
(which we had with the Solbournes), which was SCSI interface. And I've
got one which is AT&T branded, and which was thus slightly incompatible
with the Sun or Solbourne drivers of the time. I should try it with
Solaris 10 to see whether it can work properly there.

There were two front-loading 9-track drives. i don't think
either of them were very good, but the Digi-Data was certainly
one of the worst things I've ever seen. It could have amazing
disasters during threading, and I suspect any tape in the drive
when it faults or loses power is trash, as it will simply slide
off the hub onto the bottom of the drive when tension is lost.


Well ... I have known of people who use the long filenames
possible in the BSD FFS (also present in all versions of SunOs and
Solaris which I have used) as the first stage of a database, using the
shell's wildcarding to pre-select records form very ugly filenames. :-)

No, this wasn't just file names, (although it was a part of
directory searching) it had all the possible lowest-level
indexed, multi-keyed and hashed record location features built
into the FS.
Yeas, as soon as the first under-$10K laser printer came out,
Versatec was DEAD! Horrible ugly-feeling paper, like handling a
dirty chalkboard.


:-)

I wonder what the technology which It used was? It looked sort
of like a raster scanned photo buildup. Was it wet or dry processed?

Very wet. There was a big bottle of toner, which had something
like laser toner suspended in some kind of solvent that smelled
a little like paint thinner, a little like gasoline. Just
writing about it, I can smell the stuff now! Yucck!

Anyway, the paper had a (clay?) dielectric on one side, it had a
backplate with about 1" bars that were charged to 800 V, and
wire electrodes at 400 V (2112 across ~10.5") to give 200
pixels/inch.
A massive stepper motor drove the paper feed. It could do 1200
LPM in print mode, the paper just SPEWED out of the thing. It
could have gone faster, but the stepper couldn't keep up. No
computer we had could pump raster data fast enough to keep up
with it, but we could still get one-page plots in 10-15 seconds.
Now, if I could just find a good laser printer which could print
on a roll of paper, I would be happy. :-)

Yes, that was the ONE advantage of the Versatec, but the rest of
it was a DIS-advantage.

Well, I have one app that NEEDS Linux, the EMC machine tool
control.


Maybe or maybe not. Solaris is supposed to be a real-time OS to
start with, and with a PCI version of the Servo-To-Go card it should be
possible to bring it up on any of the Ultra systems (which all have PCI
slots instead of Sbus.)
There IS no PCI version of the Servo-to-Go card. Strictly ISA,
and thus woefully obsolete. My hardware runs on an IEEE-1284
capable parallel port. Is the real time compatible with rtai?
I've been sort of considering trying that with
my Bridgeport. (The linux seems to not be stable with my current
hardware, and I've never finished the conversion of the machine -- which
started life as a Bridgeport Series-I BOSS-3 running on an old quad-wide
LSI-11. The system had a massive case of electronic Altzheimer's, and
would forget what it was doing within fifteen seconds of being reset.
Usually, that was not enough time to load a tiny progrma to test it. :-)

Bridgeport never got the hang of cooling electronics, all their
gear cooked chips. They also built their own, very bad, power
supplies. I know garage hardware hackers who have no formal
training in electronics who build MUCH better gear than the old
BOSS systems. Not a whole lot of LSI-11's around that still work.
I've been converting the controller to EMC based, and the
steppers to servo motors, with one serious hangup. The servos are a lot
longer than the original steppers.
That is a common issue.
The Y-axis stepper fit into a recess
in the knee, and the servo can't fit. I've got to build an alternate
motor mount/belt guard which slopes at a 45 degree angle down to the
right, so the motor winds up beside the knee instead of trying to stab
through the knee's jackscrew. :-) And to do this properly, I really need
welding capability -- which I do not yet have. Or I need to get the
motor mount assembly from a later Bridgeport -- from after *they* moved
to servos.

I need some of the same parts. I have a 1938 round-ram
Bridgeport that I converted to CNC, but it still has the short
knee from those days. I pcked up a later BOSS knee and saddle
that have been stripped, and have been accumulating bits for it,
to eventually put it on the old machine. I found a new Y
ballscrew on eBay, but haven't come up with an X screw, yet.
And, I need all the bearing mounts/belt housings, etc.
Not a big rush, as it still has a lot of scraping to do before
it is ready to go on.

Jon