General TC11 DECtape diagnostic/formatter questions
derschjo at gmail.com
Tue Nov 8 15:24:16 CST 2016
On Tue, Nov 8, 2016 at 12:54 PM, Paul Koning <paulkoning at comcast.net> wrote:
> > On Nov 8, 2016, at 3:37 PM, Josh Dersch <derschjo at gmail.com> wrote:
> > Hi all --
> > ...
> > 2) I'm looking for means to format DECtapes on the TC11. I have a few
> > marginal tapes and I'd like to see if reformatting them brings them back
> > life. The maintenance manual only indicates "a special program supplied
> > with the TC11 system," and I haven't managed to find it. I *have* found
> > this:
> > http://mirrors.pdp-11.ru/inf.informatik.uni-stuttgart.de/pdp11/dtf.mac
> > which I've assembled and run on RT-11 and it goes through the motions of
> > writing out the timing and mark tracks, but when it goes through the
> > pass to write out the block numbers it fails immediately, with either
> > status 001207 (indicating a "Data Missed" error) or 020033 (Mark Track
> > Error). I haven't yet hooked up a scope to see if the T&M tracks are
> > *actually* being written, but given my experience with the diagnostics in
> > (1) above, I'm not averse to thinking there may be more than meets the
> > with this issue.
> > So in a nutshell: Anyone used a TC11 on a later PDP-11 (like the 11/44)?
> > Anyone have any thoughts on the diagnostics and formatter issues?
> In college I used them on 11/20 and 11/45 processors with no troubles.
> And I think at DEC we had them on an 11/70. In any case, there's no reason
> to expect trouble based on the CPU type.
Beyond the XXDP diags not running properly ;). Some further
experimentation reveals that if I set the system memory to 64KW or less,
the ZTCA, etc. diagnostics run as expected. Forcing the XXDP disk to boot
the SM (vs. the XM) monitor makes things work even on a machine with >
64KW. So that looks to be solved, fingers crossed.
> The standard DEC supplied formatting program was originally supplied as a
> paper tape diagnostic. Way back in 1974 I rewrote it slightly so it would
> do the whole job in one pass, without asking you to reset switches, and it
> would also write proper empty directories (DOS style). But the standard
> program should work fine, and the one you mentioned appears similar.
> It's sufficient to have both the WRTM and WALL enable switches set for the
> whole operation. The official approach is to set WRTM (only) during the
> first pass, WALL (only) during the second, and no switches after that. But
> having unnecessary switches set to enable is harmless.
Good to know, thanks!
> If you had a switch set wrong you'd get an invalid operation error.
> You're seeing some different code, which suggests either the mark or timing
> tracks weren't written properly, or that there's some issue with the read
> circuitry. Yes, I would say it's time to hook up the scope and start
> tracing some signals.
Yep. Hopefully with the newfound ability to run the real diagnostics, this
will be a bit easier. I'm guessing the read circuits are OK (since I am
able to use pre-formatted tapes without issue).
> If you have a tape believed to be good that you're willing to erase, you
> might try reformatting that one. That would help rule out issues caused by
> bad media. While it is very rare for DECtape media to fail to the point
> that formatting doesn't work, it *is* possible.
The tape I'm trying to format had a handful of bad blocks, none near either
end of the tape (where the error in formatting occurs). So I think the
media can be eliminated from the set of possible problems here.
More information about the cctalk