Re: SIMH to stage MicroRSTS/E for actual MicroPDP11/53

[Rob Brown]

On 2012-09-19, phiber <phi...@phiber.com> wrote:
> I also posted this on the SIMH list, but just in case anyone is paying
> attention here...

Also see alt.sys.pdp11. I have set the followups to there.

> Since my MicroPDP11/53 has no tape drive (and I don't have a floppy
> install set for MicroRSTS), I had the idea to install the complete OS
> into a SIMH pdp11 instance, and then use Warren Toomey's "vtserver" to
> transfer the resulting disk image directly to my PDP over serial
> console. Of course the trick here is to specify a SIMH pdp11 config
> that most closely resembles my hardware, as well as choosing the
> proper hardware drivers during the RSTS/E sysgen phase. Everything
> went flawlessly, and vtserver transferred my freshly installed
> MicroRSTS/E v8.0-6 RD31 disk image to my real RD31 drive. However,
> RSTS refuses to boot on the real hardware, claiming that it can't find
> a KW11-L line clock, a fatal error which drops to ODT. According to
> my KDJ11-D motherboard manual, there is in fact "CLK" clock circuitry
> at the very same memory-mapped i/o location as a KW11-L, so I'm
> puzzled as to the cause of this error. I figure there are only two
> possibilities: the 11/53's CLK isn't KW11-L compatible (which I
> doubt), or perhaps something was corrupted during the disk image
> transfer over serial, resulting in a misleading but fatal error
> message. I'd be curious to hear anyone's opinions before I reattempt,
> as transferring even a 20MB disk image over 38.4kbps serial is quite
> slow (write access to the RD31 itself is the bottleneck :)

Use console ODT or whatever host software you have running on your 11
to probe the actual addresses:

777546 KW11-L Line time clock

If the the KW11-L is to work, we expect get a response at 777546, but
not at (for example) 777544. If the clock is running, you ought to be
able to clear bit 7 at 777546, and it will get turned on again at the
next tick.

If you don't get a response at either address or at both addresses,
then try prefixing the address with 17, thus 17777546. Also probe
777560 or 17777560, which must be there since it is one of the console
terminal registers.

772540 )
772542 ) KW11-P Programmable real-time clock
772544 )

The KW11-P is a different clock that is only purchased by those who
think they needs its special capabilities. So likely you will get no
response at these addresses. I don't know if having a KW11-P installed
has any affect on the KW11-L or not.

Have fun!


--

Rob Brown mylas...@gmcl.com

[shrb...@gmail.com]

This sure rings bells from 25+ years ago when I started with Digital Canada Customer support. Our in cubicle PDP-11s were 11/53s ... and as installed, one of them had exactly this problem.

I think I have an 11/53 manual somewhere ... but emphasis on somewhere ... pick one out of 20 boxes! And they are buried under 12 years of stuff after my move back from Colorado, so, sorry I'm not digging at this time ... maybe when I'm cleaning up my workshop in the next year or two!!! :-)

I'm really stretching the brain here ... but I thought that the 11/53 had onboard boot roms and console and didn't need an MXV or equivalent ... that was for the half width cards. If you did want to use an external boot source which provided LTC, you had a jumper somewhere on the card to disable the onboard LTC.

As I say after 25 years, I may be talking totally through my hat ... but I sure remember the problem as if it was yesterday, and I'm sure the solution was as dumb as a jumper.

Stuart