SDF Beget's ICM - Interim Computer Museum

[Clem Cole]

Please excuse the wide distribution, but I suspect this will have general interest in all of these communities due to the loss of the LCM+Labs.  

The good folks from SDF.org are trying to create the Interim Computer Museum:
https://icm.museum/join.html

 

As Lars pointed out in an earlier message to COFF there is a 1hr presentation on the plans for the ICM.   https://toobnix.org/w/ozjGgBQ28iYsLTNbrczPVo

FYI: The yearly (Bootstrap) subscription is $36

They need to money to try to keep some of these systems online and available.  The good news is that it looks like many of the assets, such as Miss Piggy, the Multics work, the Toads, and others, from the old LCM are going to be headed to a new home.

ᐧ

[Neal G.]

Clem, Thanks for the heads up. I think it is useful to pass along general information such as this from time to time. And yes, glad to hear SDF will be picking up and maintaining many of the assets from LCM.

- Neal G.

[terry-...@glaver.org]

On Thursday, August 1, 2024 at 2:10:00 PM UTC-4 cl...@ccc.com wrote:

They need to money to try to keep some of these systems online and available.  The good news is that it looks like many of the assets, such as Miss Piggy, the Multics work, the Toads, and others, from the old LCM are going to be headed to a new home.

Not to be a naysayer, but I'd like to  see some proof of intention that entire systems (including all their spare parts, software and documentation) are going anywhere but to scrap metal buyers. Some rich magpies will launch bidding wars over the shiny bits and the rest will be dispersed, lost, or destroyed - a crime against history.

I have a PiDP-11 and a PiDP-10. I restore real PDP-11's (and other systems) as a volunteer. Not front panels with emulators behind them, but the entire 40-year-old assortment of TTL chips on dozens of boards, strung together with literally thousands of feet of backplane wires. My PiDP-11 is infinitely faster and (theoretically, but that's a different diatribe) easier to maintain. But it is NOT a PDP-11/70. I appreciate it because I don't have the space or power to run a real 11/70 with RP disks and TU77 tapes at home. But I find repairing (and then using) the real thing infinitely more satisfying.

Thousands of people are never going to get that experience if the LCM's equipment is split up, no matter where the pieces end up (if all of the pieces actually survive, which seems doubtful).

ᐧ

[Clem Cole]

On Tue, Aug 13, 2024 at 12:25 AM terry-...@glaver.org <terry-...@glaver.org> wrote:

Not to be a naysayer, but I'd like to  see some proof of intention that entire systems (including all their spare parts, software and documentation) are going anywhere but to scrap metal buyers. 

All I can suggest is that you join the ICM mailing list https://icm.museum/join.html and listen to progress as a minimum. It is better if you join the effort and contribute some money to help them try to save these assets.  I do believe Stephen and the SDF folks are doing what they can. In my opinion, they think that losing the LCM+Ls assets would be a bad thing, and they are doing what they can do to keep that from happening.  Please go back and read Stephen's message on this list, which offered more color than my original note.   BTW: Miss Piggy is a real 11/70, and Stephen and team do have it at the new site and is actively working to bring it back online.

 

ᐧ

[terry-...@glaver.org]

I'm already supporting two museums (rail and computer) on a retiree's budget.

I wish you well with your project, but I'm not sure you realize the scope of what it takes to actually run a museum (which isn't just exhibits, keeping the power on, and publicity). There's lots of both nasty grunt work that a limited number of people can do, and specialized skills (not just restoration / emulation, but things like legal, endowment planning, etc.) which even fewer people can do.

We wouldn't be in this situation in the first place if Paul (who had both money and brains) had made some more detailed preparations for the continued existence of the LCM. True, he didn't expect to pass away when he did (nearly none of us do), but he could have made his wishes (if any) more expressly clear and saved all of this last-minute panic.

Believe me, I know Miss Piggy is an 11/70 (well, technically a decdatasystem 570). I just spent several days in, under and around one over the past week. With real peripherals. There is another 11/70 (in H960 packaging) there that's waiting for the 570's core box (there was a Field Improvement Program by DEC which replaced core boxes with MOS for contract customers, way-back-when) once I get it debugged. And I know of yet another 570 that should still be where I left it (since it is in a basement with a 24" door and I had to take it completely apart, including drilling out the rivets, to get it in there and then reassemble it, and the house is still owned by the same person, I'm 95% sure it is still there).

I also got a VAX-11/780 (with 100% real CPU, memory, and peripherals) booting for the first time in 20+ years on that same trip. So this is not my first rodeo (far from it, actually). While I know what skills I bring to a museum like this, I know there are other things that are at least as important that I have no talent for whatsoever.

Assuming you get some or all of the systems from the LCM, I implore you to not make the same mistakes that led to its demise.

[Johnny Billquist]

I'm not going to comment on the museum attempts, or miss Piggy as such
(much). Just wanted to add a bit on the 11/70 details...

On 2024-08-13 21:35, terry-...@glaver.org wrote:
> Believe me, I know Miss Piggy is an 11/70 (well, technically a
> decdatasystem 570). I just spent several days in, under and around one
> over the past week. With real peripherals. There is another 11/70 (in
> H960 packaging) there that's waiting for the 570's core box (there was a
> Field Improvement Program by DEC which replaced core boxes with MOS for
> contract customers, way-back-when) once I get it debugged. And I know of
> yet another 570 that should still be where I left it (since it is in a
> basement with a 24" door and I had to take it completely apart,
> including drilling out the rivets, to get it in there and then
> reassemble it, and the house is still owned by the same person, I'm 95%
> sure it is still there).

The "improvement" on an 11/70 to replace core with MOS wasn't as much a
upgrade by contract as a separate product you could buy, and "newer"
11/70s were usually sold with them.

The core memory box is called MJ11. The MOS memory box is called MK11.
The MK11 memory box use the same memory bus as the VAX-11/750, and the
256K memory boards can be used in both machines. The 1M memory boards
for the VAX can however not be directly used in an MK11.

Update still have three 11/70 machines. One of which is a DECdatasystem
570. They are all operational, or can somewhat easily get to that state
if we just had more space. We used to run one 11/70 24/7 on the internet
before we had to move out from the university.

I was told miss Piggy was/is somewhat unstable, and for some reason LCM
never did much with it. Sadly.

Oh, and the 11/74mP systems use a variant of the MK11, called the MK11A.
Basically the same memory box, but the controller is multiported.

Johnny

--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: b...@softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol

[terry-...@glaver.org]

On Tuesday, August 13, 2024 at 7:48:51 PM UTC-4 b...@softjar.se wrote:

I'm not going to comment on the museum attempts, or miss Piggy as such
(much). Just wanted to add a bit on the 11/70 details...

On 2024-08-13 21:35, terry-...@glaver.org wrote:
> Believe me, I know Miss Piggy is an 11/70 (well, technically a
> decdatasystem 570). I just spent several days in, under and around one
> over the past week. With real peripherals. There is another 11/70 (in
> H960 packaging) there that's waiting for the 570's core box (there was a
> Field Improvement Program by DEC which replaced core boxes with MOS for
> contract customers, way-back-when) once I get it debugged. And I know of
> yet another 570 that should still be where I left it (since it is in a
> basement with a 24" door and I had to take it completely apart,
> including drilling out the rivets, to get it in there and then
> reassemble it, and the house is still owned by the same person, I'm 95%
> sure it is still there).

The "improvement" on an 11/70 to replace core with MOS wasn't as much a
upgrade by contract as a separate product you could buy, and "newer"
11/70s were usually sold with them.

I suspect this varied greatly by country. DEC US and Digital Canada often had

rather different business practices despite sharing a border.

I've seen enough of these with inexplicably missing memory terminators on

the MJ11 and ribbon cables with "Mxxxx J1", etc. where xxxx is the part num-

ber for one of the MJ11 controller boards plugged into the MK11 controller.
They also tend to have the MK11 addressing switch/LED assembly mounted 

in odd locations. The fact that all of these systems I've encountered still have 

their MJ11 boxes would also seem to indicate DEC didn't want them back, 

which they would have on a paid upgrade. I forget the paperwork (LARS, may-

be?) where you had to list the installed items as well as the removed items.

The core memory box is called MJ11. The MOS memory box is called MK11.
The MK11 memory box use the same memory bus as the VAX-11/750, and the
256K memory boards can be used in both machines. The 1M memory boards
for the VAX can however not be directly used in an MK11.

There are a number of variants of both core and MOS controllers as memory

became more dense. Getting above 3.5MB in a MK11 involved some creative

work with lower-density boards, as otherwise it stomps on the I/O page, etc.

It shouldn't (those bus cycles should never get out on the memory bus in the

first place). The 11/70 is full of those little oddities, like the system size regis-

ter. The only point of that is to generate an immediate NXM trap instead of 

waiting for memory to not respond and then doing the NXM trap. Unless you

make a habit of routinely addressing memory locations that don't exist, it is

essentially a no-op and can just be set to the maximum memory size.

I was told miss Piggy was/is somewhat unstable, and for some reason LCM
never did much with it. Sadly.

The backplanes in those systems were hideously complex, with the rework

instructions to go from one rev to another calling for hundreds of feet of wire

wrap wire, some of it replacing single-strand yellow with black/white twisited

pair.

It was so bad that DECmailer just refused any backplanes lower than rev "L".

Not saying that's what Miss Piggy's problem was, just mentioning it as a pos-

sibility. There are inter-related board revisions, but the 11/70 was never under

precise revision control as the 780 was.

 

Oh, and the 11/74mP systems use a variant of the MK11, called the MK11A.
Basically the same memory box, but the controller is multiported.

Right. Because 9 ribbon cables per memory box (4 in, 4 out, one OCP) wasn't

fun enough on its own...

[Johnny Billquist]

Oh. I'm not arguing that MJ11 boxes got replaced by MK11 boxes. After
all, at the start, there were only the MJ11. The MK11 upgrade, however,
was pretty significant, as that memory is faster. So customers would
certainly upgrade. But DEC's only involvement in many cases would just
have been selling those memory boxes.

I'm sure some also did that under contract, but I'm pretty sure many
just did it on their own. There were/are also 3rd party memory boxes you
could replace the MJ11 with. Since each MJ11 only held at most 256K, you
needed a total of 8 boxes to get a full memory on the 11/70. Which is
two cabinets worth. Quite a large thing.

With the MK11 you could reduce that to just one box.

> The core memory box is called MJ11. The MOS memory box is called MK11.
> The MK11 memory box use the same memory bus as the VAX-11/750, and the
> 256K memory boards can be used in both machines. The 1M memory boards
> for the VAX can however not be directly used in an MK11.
>
>
> There are a number of variants of both core and MOS controllers as memory
> became more dense. Getting above 3.5MB in a MK11 involved some creative
> work with lower-density boards, as otherwise it stomps on the I/O page,etc.
> It shouldn't (those bus cycles should never get out on the memory bus in the
> first place). The 11/70 is full of those little oddities, like the
> system size regis-
> ter. The only point of that is to generate an immediate NXM trap instead of
> waiting for memory to not respond and then doing the NXM trap. Unless you
> make a habit of routinely addressing memory locations that don't exist,
> it is
> essentially a no-op and can just be set to the maximum memory size.

Ok. Now I'm going to go technical on you... ;-)
No, it's easy to go max on an MK11 box. Each board is 256K. The Unibus
map takes 256K, so the max memory you can have in an 11/70 is 3.75MB,
which is 15 memory boards. Just put 15 in, and you're there.
However, I suspect what you are thinking of is the fact that if you have
an even number of boards, the box will internally interleave between the
boards, so you get better performance. And 16 boards cause an issue, but
not entirely what you describe (more in a second). There are also the
64K boards, and you could add two of those in the last two slots, so you
at least get 128K more. Still not 3.75, but more than 3.5.

The problem with having the full 4M in the box isn't the I/O page. The
I/O page isn't on the memory bus anyway, and any addressing to the I/O
page goes there directly, and do not pass memory from the CPU anyway.
However, the memory box also have a CSR, so you can play with the ECC
and stuff. And that CSR is in the same address space as memory, and you
get a conflict between that and the memory itself, if you go to 4M. But
even getting to that CSR is complicated, because it is in the range
where the I/O page is, but as I mentioned, access to the I/O page do not
even go on the memory bus. So getting to the memory box CSR is
complicated. You basically have to enable the Unibus map, and then setup
a mapping there that leads to the address of the memory box CSR, but you
access it in the Unibus map range.

And yes, the 11/70 have all kind of fancy stuff. Like the memory size
register you mention, but also the serial number register which also
comes from a switchpack in the CPU.

> I was told miss Piggy was/is somewhat unstable, and for some reasonLCM
> never did much with it. Sadly.
>
>
> The backplanes in those systems were hideously complex, with the rework
> instructions to go from one rev to another calling for hundreds of feet
> of wire
> wrap wire, some of it replacing single-strand yellow with black/white
> twisited
> pair.
>
> It was so bad that DECmailer just refused any backplanes lower than rev"L".

Yeah. The 11/70 is complex. But nice.

> Not saying that's what Miss Piggy's problem was, just mentioning it as a
> pos-
> sibility. There are inter-related board revisions, but the 11/70 was
> never under
> precise revision control as the 780 was.

No idea what the problem is/was. I was just told that when people were
exploring details on how an 11/70 works, the machine occasionally would
go weird, and they would have to power cycle it or something.

But that don't sound like some revision issue, but more of just old,
cranky hardware that needed some fixing.

> Oh, and the 11/74mP systems use a variant of the MK11, called the
> MK11A.
> Basically the same memory box, but the controller is multiported.
>
>
> Right. Because 9 ribbon cables per memory box (4 in, 4 out, one OCP) wasn't
> fun enough on its own...

Yup. :-D
Although, in all fairness, if you got the 256K cards, then one MK11 was
all that was needed, which means you only had 5 ribbon cables...

[terry-...@glaver.org]

Replying up-thread as it relates to the original post, not my 11/70 digression

The full Christie's listings are now up. The first one is almost all computer equipment, the second one is mainly non-computer stuff:

https://onlineonly.christies.com/s/firsts-history-computing-paul-g-allen-collection/lots/3726

https://www.christies.com/en/auction/pushing-boundaries-ingenuity-from-the-paul-g-allen-collection-30730/

Fortunately, Miss Piggy and the TOADs aren't listed, so they may have been saved.

The KA10 is being sold incomplete, with the only media or documentation being two DECtapes of unknown provenance / condition.

The KI10 is likewise incomplete, but does include a dozen and a half or so 3-ring binders of material, a file folder full of random notes, and a short stack of stapled period-appropriate DEC manuals. No magnetic media, though. Also includes a period-incorrect LA36 console.

The auction's terms and conditions at:

https://www.christies.com/pdf/onlineonly/EcommerceConditionsOfSale-NewYork-23618.pdf

make it quite obvious (refer to Section E(2), subsection (h) "Vintage Computers and Machines") that these are being sold as possibly incomplete (definitely in the two examples I listed), perhaps with fabricated or replicated parts, and are sold as "antique and collectible only". In other words, no guarantee that even the subset of the system being sold is functional. So even if they were functional, complete and operable at the LCM, there is no guarantee that that is currently the case, even for systems that appear complete. And of course the expertise to reassemble, test/repair, and operate these systems has reportedly been gone from the LCM for a long time.

It may not be possible to say (yet?) what the ICM got / will be getting / hopes to get from the LCM, but if it is, some sort of "here is what we hope to start with" might serve to attract new donors / members. At a (non-computer) museum where I also volunteer, donors can make unrestricted financial contributions (used as the museum sees fit), for specific general types of use (equipment / archives / power / etc.) or toward the restoration of a specific piece of equipment.

On Thursday, August 1, 2024 at 2:10:00 PM UTC-4 cl...@ccc.com wrote:

[terry-...@glaver.org]

On Thursday, August 15, 2024 at 4:51:47 AM UTC-4 b...@softjar.se wrote:

No idea what the problem is/was. I was just told that when people were
exploring details on how an 11/70 works, the machine occasionally would
go weird, and they would have to power cycle it or something.

But that don't sound like some revision issue, but more of just old,
cranky hardware that needed some fixing.

I think your referring to Oscar and other people researching system behavior for the PiDP-11 kit. While much of that work was of the form "What do the front panel LEDs show in <some specified> state?" a bunch of it was apparently to determine what would happen when a real machine was presented with things not documented by the architecture. Recall that even though the VAX had a formal architecture specification before the first CPU was built, there wasn't an actual compliance exerciser until quite late in the product lifecycle. Before then, problems were found by either a new machine not booting VMS / running diagnostics, or by problem reports from the field. Supnik documents a number of these (passive release on the MicroVAX II, etc.).

I expect that a lot of the instability encountered on Miss Piggy was from trying to introduce undefined states, mostly to see what the front panel would show (the project was definitely not trying to modify simh to perfectly reproduce the behavior of undefined states).

Oscar's specific words (at https://obsolescenceguaranteed.blogspot.com/2016/01/starting-to-make-pidp-1170.html) were:

"Mike in Zürich wrote us a test program exercising all the corner cases for which you will not find the exact front panel behaviour documented, and provided us with 17 pages of tests and checks using that program. A big thank you to the Living Computer Museum for providing us access & support, patiently rebooting the machine remotely from the upper floor 9 times..."

[Johnny Billquist]

I don't know the exact details of the issues, but from my talking to
Oscar it certainly sounded like specific problems with that machine.

Johnny

> for which you will not find the /exact /front panel behaviour

> documented, and provided us with 17 pages of tests and checks using that
> program. A big thank you to the Living Computer Museum for providing us
> access & support, patiently rebooting the machine remotely from the
> upper floor 9 times..."
>

> --
> You received this message because you are subscribed to the Google
> Groups "[PiDP-11]" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to pidp-11+u...@googlegroups.com
> <mailto:pidp-11+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/pidp-11/b1d534cd-dc5c-485c-8996-ad8d4d699d7dn%40googlegroups.com <https://groups.google.com/d/msgid/pidp-11/b1d534cd-dc5c-485c-8996-ad8d4d699d7dn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[terry-...@glaver.org]

On Thursday, August 15, 2024 at 7:44:31 PM UTC-4 b...@softjar.se wrote:

I don't know the exact details of the issues, but from my talking to
Oscar it certainly sounded like specific problems with that machine.

It could very well be. On several occasions I had folks from RSTS Engineering fly down to New Jersey to try things on my 11/70 systems (spc11d and spc11e.spc.edu were both H960 11/70's with 3.5MB of MK11 memory) because they were experiencing problems on their development system(s) that they couldn't explain. One visit was because they thought the RSTS/E source tree might have been damaged by an employee who left on bad terms (turned out they had a hardware problem, or multiple problems). I kept one as a KB11-B and the other as a KB11-C, both with their corresponding FPAs. I also had a huge assortment of disks and tapes that could be reconfigured onto those systems. 11d usually had 3 RP07s and a TU77, while 11e usually had 4 RM05's, a TU80 and a TUK50. But various RK/RL/RM/RP disks could be attached as needed, so I could replicate the engineering environment as needed. I had previously gone through and made sure the backplanes were up-to-rev, all the fans were changed (listening to all of those brand-new ball-bearing TA450's spin up in perfect harmony was a sound to behold) and the regulators adjusted precisely.

That last one can be particularly tricky, if different 5V regulators are providing slightly different voltages to different places in the CPU. The logic _should_ have been tolerant of even a bunch of supplies were some were adjusted to the minimum allowable voltage, and others to the maximum. But those machines were 10+ years old when I got them, the supplies had drifted and the logic may have become more intolerant.

[oscarv]

Yes, Miss Piggy was quite unstable, I remember I had to go up two floors to get the simulated disks reset - about a dozen times :).

The tests were run by Joerg Hoppe (retrocmp.com) with me holding the camera and trying not to get *too* confused. Joerg has the tests online - Mike Hill's test script and a bunch of short Youtube movies:

https://www.retrocmp.com/stories/pdp-11-70-panel-research/319-pdp-11-70-panel-research-the-videos

If you write your own simulator, I think these videos are extremely useful. I learned a lot - but then, over time, forgot most of it again of course...

Miss Piggy did have a few front panel switches soldered in upside down, but that's cosmetic only. It is important to know when you study the videos though. IIRC, some of the switches on the left side of the panel are upside down, so 0 becomes 1, 1 becomes 0.

Kind regards,

Oscar.

[Dave Plummer]

I’ll be bidding on Miss Piggy. Don’t bid against me :-)

Thanks,

Dave

Sent from my phone - Please excuse brevity and format. 

On Aug 13, 2024, at 12:35 PM, terry-...@glaver.org <terry-...@glaver.org> wrote:



--

You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/843aa7ea-334d-4295-b6bf-f3dd144ee0fbn%40googlegroups.com.

[Dave Plummer]

--
You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/2a76fc16-96da-487b-b53e-b09d8af136d4n%40googlegroups.com.