Care and feeding of some Lisp machines (TI Explorer and Xerox Star)
engel at multicores.org
Mon Sep 14 15:04:09 CDT 2020
whow, seems it's my lucky day :) - thanks a lot for the offer!
> On 14 Sep 2020, at 18:00, Josh Dersch <derschjo at gmail.com> wrote:
> I've restored a Star/1108 (and wrote a Star emulator) and am in the middle of an Explorer restoration, I'm happy to help out where I can.
> I'd recommend picking up an MFM Emulator (https://www.pdp8.net/mfm/mfm.shtml) along with the SA1000 adapter for same, for use with the Star. The original disks are getting more difficult to keep running, and it's also a lot more convenient for switching between different operating systems, etc.
That's definitely a good idea - and good to know that the disks are simply MFM and not some more exotic standard such as SMD.
> Remove the sound-deadening foam from the panels of the system, it's getting crumbly and isn't going to do you any favors to leave it in place.
Yeah, I remember the TI1500s also had that problem and the stuff is pretty nasty.
> I've found the power supplies to be fairly reliable.
Good to know - analog power electronics is not really my area of expertise, but we have a competent electronics department nearby that can probably help just in case.
> One issue is weak picture tubes in the displays -- the monitors are powered on with the system and have no separate off switch, so they tended to get a lot of hours put on them. We had good luck with a tube rejuvenator on the one we restored at LCM.
Yes, the picture seems to be burnt in quite a bit in all of the displays (we have three systems here). Not sure if I can find a tube rejuvenator here in Trondheim, but you never know.
> The Star mouse pad can be recreated with a laser printer (I've used this: http://www.digibarn.com/collections/devices/xerox-mousepad/index.html, and there's a postscript file floating around out there...). Or any surface with a fine pattern on it seems to work pretty well; I was able to make it work on a speckled countertop and the pant leg of my jeans at one point. It's a lot more forgiving than the Sun mice which need the fixed grid of the metal mouse pads.
> For the Explorer, there are a number of r fa line filter caps in the system, on the power supply board as well as on a separate board near the rear of the chassis. I suggest replacing these immediately as they like to let out smoke.
Good to know, I'll take care. The roller-type fans (hope that's the correct English term) also made some problems at least on the large TI1500 system back in the '90s, the Explorer should use the identical chassis.
> The optical cable is extremely rare and despite some valiant efforts we haven't found an equivalent, or new-old-stock replacements. A friend of mine is working on retrofitting modern optics, and has made some great progress.
It seems that the TI931 terminals also had a fiber optics option and used the same cable (http://bitsavers.org/pdf/ti/terminal/crt/2229228-0001A_Model_931_Video_Display_Terminal_General_Description_Dec83.pdf) - but I've never seen a 931 using that option. So that won't make the cable easier to find.
> The mouse is indeed a standard Mouse Systems, I'm missing mine at the moment and haven't yet gotten to the point of adapting a mouse to replace it. I suspect it's equivalent to the M2 used on the Sun-2 and LMI Lambda systems.
Hm, on some photos I found the Explorer mouse looks identical to the Sun-2 mice (black case with white buttons). However, the Sun-2 (just checked the schematics on bitsavers) already used a serial interface (via a Z8530 SCC) to connect to the mouse, the Explorer (see my other post with the pinout I found) seems to use quadrature encoding. But maybe there have been different options for the Explorer.
> Media for the Explorer is another question that I'm hoping to answer soon. There are disk images from the Meroko emulator but my understanding is that they are incomplete. Bitsavers has QIC tape images but I have yet to try them. The interface on the Explorer is SCSI but I haven't had luck booting it from a SCSI2SD w/Meroko images loaded. The disk boxes contain an Emulex SCSI->MFM bridge, so use of Dave's MFM emulator might make sense here as well.
I think the firmware on the NUPI-2 NuBus disk/tape controllers (this has its own 68000 CPU) speaks a pre-standardisation SCSI protocol variant (something closer to the original SASI, I think). The TI1500 used Imprimis/CDC WREN 670 MB disks (probably WREN VFH 94181-702) by default IIRC - and those were the only disks (of that size) that worked reliably on the 1500. I assume the NUPI-2 in the Explorer is identical, so that might be the problem with SCSI2SD. We have a number of the disk enclosures, so I hope that some of the disks still work.
> If you have disks in your Explorer, let me know -- capturing an image of their contents would be extremely useful, and the original Maxtor drives are not long for this world.
We have four machines with drives, I'll definitely do an image of all of the disks first and will also check if I can find some install tapes.
> Hope that helps a bit, happy to answer any questions... or try anyway.
Thanks a lot - the machines will definitely keep me busy in the upcoming rainy weekends and evenings...
More information about the cctech