DEC M9312 Configs for a 11/40

william degnan billdegnan at gmail.com
Mon Nov 28 09:11:23 CST 2016


>
>
>>
>> Thanks.  Where is the M7762 in this lineup?


I removed it, put in a NPG card in its place, while attempting the TU58
emulation set up.



> You might try shortening the bus to eliminate extra variables (the second
> MM11 backplane, and the 2nd DD11 backplane can be removed) but you're
> probably fine.



Yes I have done this, no difference in performance.  This is actually good
because I get consistent results.  I have also eliminated the core RAM and
used MOS, no effect on the RL11 or TU58 issue.


> Based on the above, it looks like you've gone over all the grants and have
> the proper continuity cards and/or NPR jumpers in place.  I would also
> suggest cleaning the edge connectors on any that look dirty or oxidized, it
> can and does make a difference (in my experience).


ok.  Generally everything is very clean and I have used Deoxit when it
looked like it needed it.


>
>
> If anyone knows of or would be willing to make an XXDP TAP image
>>>> compatible
>>>> with a 16K PDP 11 I can convert to download via PDPGUI.
>>>>
>>>> PDP11GUI doesn't use TAP files.
>>
>>

This is how I load BASIC, using the BASIC papertape file converted into a
CONSOLE script that writes each memory address one at a time

L 0000
D 11
D 11
D 12
etc.
L 200
S
(for pseuo example)


> If you really want to make a bootable XXDP TU58 tape image, the
> instructions given by AK6DN here:
> http://www.vcfed.org/forum/showthread.php?42246-How-to-make-
> a-bootable-XXDP-file-for-TU58-emulator
>
>

this is what I have been doing.

<snip>
>>
>>
>> I use both the M9312 DL ROM and manually entering in the bootstrap without
>> the M9312 installed (replace with terminator card).   With the ROM method
>> the system repeats an attempt over and over.
>>
>> system is looping between instructions 173022 -->173156 (error reset the
>> world), 173160 (retry)..back to 173022
>>
>> When I first power on the values in the control status register (774400)
>> are good. 1-6 and 8-13 are cleared, bit 7 is set.
>> After I attempt to boot (DL at the @ prompt), you can see the light
>> pattern
>> repeat over and over, looping through 173022 -->173160.
>> I HALT the system and recheck the control status register. This time
>> 15 - error (composite error (ERR))
>> 10 - EO (operation incomplete (OPI))
>> 7 - stays lit
>> function code F2 = 1, F1 =1, F0 =0. this code indicates "HFT" (Header Not
>> Found)
>>
>> To me (and this is just me) the bootstrap is not waking up the RL11
>> controller, and can go no further.
>>
>
> According to the user's guide: http://bitsavers.trailing-edge
> .com/pdf/dec/disc/rl01_rl02/EK-RL012-UG-004_RL01_RL02_
> Disk_Subsystem_Users_Guide_Oct80.pdf
>
> If the CSR bits are as you report, the drive got the Read command (that's
> what F0-F2 indicate -- the last selected function, in this case "Read
> Data").  Bit 15 indicates a composite error, and bit 10 indicates OPI, as
> you note.  I don't see anything in there about HFT unless you left out some
> bits.
>
> If indeed bit 12 is set (indicating HFT) then the "Header Not Found"
> indicates that the drive was unable to find the header for the sector on
> the disk, which makes it look like the drive is trying to read but for
> whatever reason isn't able to find the sector.  This could be a damaged
> disk, misaligned heads, or any number of logic failures.
>
> I would start with the RL02/RL11 diagnostics via either XXDP or PDP11GUI
> and go from there.
>
>
this just bombs, does not run.  I agree though.  basically the computer
says "I can't reach the controller"  I should add that the RL ROM (run DL0
from CONSOLE prompt) keeps trying over and over again, you can see it.  If
you run the bootstrap from the front panel it just dies.

What I can do next is make more note of the registers, see if I can get
more clues.

It just seems to me that there is a CPU or UNIBUS issue when I see
performance like this, again my hunch. It's not the controller.  If the
disk was bad, it'd at least try something, you'd see something.

b


More information about the cctech mailing list