7 Aug
2011
7 Aug
'11
1:57 a.m.
From: "B M" <iamvirtual at gmail.com>
Sent: Saturday, August 06, 2011 6:25 PM
To: <cctech at classiccmp.org>
Subject: TMA-11/TS03
With the help of various knowledgeable people on this
list, it was
pointed out that
I did not have the connection to the tape drive controller in the
correct position.
To test the tape drive, I used the programmer's console on my PDP-11/05 to
issue a read command to a mounted tape. When I issued a read command by
depositing a command into the TMA-11 command register, the tape advanced
slightly and then stopped. The 'Bus Grant Late' error was set in the
TMA-11
status register. The status register did indicate the controller and tape
drive
were both ready. No data was read in.
The read command was issued while the processor is halted.
I am always careful to give a response, because I am sure there are more
knowledgeable people here. True, you can give a command to the TMA-11,
like move tape forward, but if you command to read data, to must also
actually read data! Otherwise a buffer overflow will occur and then the
"late"
message will appear. I think ...
But as the tape moves a little, it seems that probably nothing died by the
wrong connection. Try XXDP at it. ( I have to do that too ... one day).
Based on previous posts from others, I suspect I
don't have the processor
grant set up correctly.
I am currently reading several different manuals to get a description of
how
the grant lines should be hooked up. I have read many descriptions of the
signalling, but I have not located information on where to use the Grant
Continuity cards or making changes to the wire-wrapped bus. In my case,
I will need to make changes to the bus, since I do not have any of the
Grant Continuity cards ;-)
Am I on the right track to assume that the problem is likely to be
the processor grant wiring, or should I be looking elsewhere?
Again, I am not sure, but if you can control the system and access the
UNIBUS, I'd say there is no GRANT problem. You can easily check this
by sending a character to the XMIT register of the console.
Just deposit 066 (whatever) at 777566 and see if a character is
printed on the terminal. If so, you know the UNIBUS is OK.
Grant cards are small "knockle-busters". They go in position D and connect
4 grant wires. I would not wire-wrap those 8 pins, but find a G727A.
I am sure somebody local to you has plenty of them.
The NPR Grant module G7273 goes in position C and D. It has the 4
connections as the G727A plus the 'connection for NPR in slot C. Those
are also relatively easy to obtain.
If you "mess" with UNIBUS systems it is handy to have a few G727A and
G7273 lying around.
- Henk.
Many thanks for any pointers!
--barrym