28 Sep
2011
28 Sep
'11
1:42 p.m.
On Tuesday (09/27/2011 at 05:40PM -0400), allison wrote:
I will too eventually but it's a lot easier for me to debug this thing on
the bench with a laptop sitting next to it than a rack full of -11 stuff
;-) My -11 stuff is not in the same room with the electronics bench.
My PS is looking OK. Good, clean +5 and +12.
OK. That's helpful info. Since I don't have it sync'd with a host
just yet, mine is off. I'll expect it to come on then once comms
are established.
I'm likely
to be writing a Linux app to communicate with the TU58
very soon. I am restoring a TU58 and need a debug tool to talk to it
more conveniently than using one of my PDP-11.
I use a pdp-11. On my unit, the
red LED only makes a brief flicker as the unit is being
powered up and down. Scope reveals life-- 8085 CPU is executing stuff.
I see the UART being polled, etc. so I have been believing mine is closer
to life than not.
The usual problem if the PS is not failed is the drive rollers have
turned to goo. I have yet to see one that has not. A goo roller will
foul a tape.
Absolutely. And I have replaced the rollers. I found "new" ones on eBay.
Interestingly, the old rollers were rusted onto the shaft of the motor and
it was a royal pain in the ... to get them free. I ended up scraping
all of the goo from the roller hub and then heating the hub with a
high power soldering iron and then quickly (and carefully) hammering
a small center punch onto the motor shaft at the center of the hub.
Eventually it popped loose. The motor shaft is steel and rusted while
the roller hub is aluminum but the rusted shaft had a real good bite
into the hub it would seem.
Is the red LED supposed to come on and stay on?
My impression from the
tech manual is that the red LED is a fault indicator of some sort so
I was OK with it being mostly off. If it's supposed to be mostly on,
then I have more debug to do.
It stays on when the system in sync with the host or task complete. It's not only a power indicator.
Got it.
Ther is a manual for the Tu58 that gives the protocal
(mostly).
Yes. I have it.
I have replaced
the capstan rollers in this thing and am really just
beginning the debug/restoration. I'm probably also going to find that
the first tape I put in it is going to break when it starts moving...
of the dozen or so tapes I found with the unit, four of them already
had the drive belt broken inside the tape cartridge.
Could be a lot of effort to get 512, 512-byte blocks ;-) But hey,
it's fun.
From memory..
Not that hard, basically hit it with a break, wait for an ack packet,
respond with a read block N command packet and wait for data as four
checksummed packets and a status packet.
FYI when it is sending data the host MUST be able to accept and buffer
at least 128 or 512 bytes (MRSP allows for an wait on ack for each 128 bytes
[pdp11/44 overflow fix] where the RSP protocol assumes you can buffer the
whole mess at line rate. From the documentation that I have, it seems pretty
straightforward. There
is a wakeup protocol discussed which is to send BREAK, then
0x04, then
another 0x04 and you should see a response (which I've forgotten at the
moment) come back.
The protocal is a bit convoluted.
But-- it seems simple which is probably all they needed then.
Thanks for the info.
Chris
--
Chris Elmquist