IBM 5100 Portable Computer, APL/BASIC

User avatar
Bass

16 Nov 2018, 08:36

Image

Introduction
Although there is some interest in vintage keyboards in the mechanical keyboard community, there is generally not as much interest in complete vintage computer systems. However, I would argue that the IBM 5100 is among the most well known of all vintage computers around here. This is most likely a combination of these things:
  • The keyboard uses IBM's now famous beam spring switches
  • It's among the first portable computers ever made (if not the first).
  • It was referenced heavily in the popular anime / visual novel series Steins;Gate, which is itself based on the story of John Titor who claimed to be a time traveler who was in search of one to save his dystopian future.
All three of these things apply in my case, so as I learned more about the 5100, the more interested in it I became. But as I quickly learned there is a reason John Titor and all the main characters of Steins;Gate had such a hard time finding one. It turns out these are quite rare. The only other active member of this community I know who has one is snuci, who impressively found two back in 2015.

The Acquisition
I feel like the story of how rare and unusual treasures are acquired can be really interesting in of itself, and this 5100 is no exception. In my case, my interest in the 5100 was reinvigorated by this ebay listing that I am sure some of you have already seen:

https://www.ebay.com/itm/Rare-working-I ... fresh=true

This unit is definitely in better shape than mine, but unfortunately the asking price is also just below an astonishing $8000 USD, which makes it out of reach for just about all of us.

This listing ended up making me curious enough to learn much more about the 5100 than I had previously known, and so I ended up checking out various websites with information on vintage computers. Eventually, after carefully reading the comments section for one of these sites, I saw a post stating that this particular 5100 was for sale, with great preference to local sales. This was about three weeks ago, and the post was dated in September. The seller left their email address in the comment, so I decided to ask if it was still for sale anyway even though it was definitely a long shot. About two weeks then past by since I made the inquiry, so I figured it was long sold out.

We now arrive to one week ago, where the seller finally responded to my email! He told me that he still had it but was also feeling wishy washy about selling it, and also strongly preferred local sales, mainly because he was worried about shipping something so heavy (it weighs around 50 lb by the way!). We then had a lengthy back and forth that took place over the span of a few hours, where I explained my experience in shipping heavy vintage electronics and my passion for vintage IBM computer keyboards. The seller in turn explained that the 5100 was given to him as a reward for helping a company write some software back in the 1970's when he was still high school. He pretty much left the machine in his attic for the next few decades mostly undisturbed, though it has gotten quite dirty as a result, and probably some rust on the inside.

After the reminiscing was over, I asked the seller what price he was thinking about selling it, and to my surprise it was $1000 plus shipping. That's about inline with what your run of the mill beamspring in this condition would cost you, but still a very good deal for something as rare and special as an IBM 5100. Needless to say, it did not take long for me to accept the offer. The seller quickly got it ready for shipment at a UPS store, and so I patiently waited. He was located on the east coast while I am on the west coast, so I had to wait almost a whole week before it arrived to me today.

The arrival
Even though I went through all the effort to convince the seller that shipping it should be fine, I was still a bit anxious, especially at the moment the package arrived to me:

Image

I guess this picture does not do it justice, but I was very disappointed in how the carrier (UPS) handled the package. In spite of the many fragile stickers on the box, it was pretty badly damaged. There was also some room for the unit to wiggle around in the box as I moved it. Thankfully a mountain of bubble wrap awaited me when I opened the box, giving me some reassurance that there wouldn't be an physical damage to the exterior of the unit:

Image

After digging through all of the packaging, I was elated to see what I had asked for arrived to me pretty much intact:

Image

The seller even left me a friendly message:
Image

First Impressions
This thing is definitely even more impressive to see in person, in spite of how dirty it is. One thing I like about the 5100 with respect to its successors was the black and white color scheme of the keycaps, much like my beloved 3278. Furthermore this model supported both the BASIC and APL programming languages, and so there are APL characters on the keycaps as well.

Image
Image

The weight is also quite impressive at around 50 lb. In spite of how badly it was handled in transit, I could not see any visual damage to the case. This came right after seeing many people post some pictures of their keyboard cases getting ripped apart during transit on the mechkeys discord. Just goes to show how impeccable IBM's dedication was at the time in regards to selling products that were meant to last, in contrast to the mass market crap that we use today.

I still need to spend more time deconstructing the unit, but I managed to get one rather nice picture with the top off:
Image

Does it work?
Unfortunately the short answer is no, and that is part of the reason for the price. The seller told me that it worked before he left it in storage. He tested it about a month ago and it was able to boot it until it failed in self test:

Image
Image

The seller's theory is that oxidation has caused a bad connections in the card cage. snuci reached out to me after I initially posted the news about my IBM 5100 acquisition last week, and warned me that I will very likely need to to fiddle with this part of the machine if I have any hopes of getting it to work. I couldn't resist the temptation and did attempt to turn the unit on before writing this post, but things have gotten worse in transit: The screen doesn't turn on and the "Process Check" light stays on. Most likely too much wiggle room in the box caused some of the components to get loose. I can't blame the seller for this since I said leaving the packaging up to the shipper would be fine based on my previous Unsaver experience. Goes to show that no carrier is consistent.

Future plans
In any case, this is a story to be continued. My plan is to restore this machine to the best of my ability and ideally get it cleaned up and working, but I do not yet know if the latter is possible. My main priority right now will be to take the cage out and see how the cards look, then see how much improvement I can make after cleaning up the interior. I intend to take the keyboard module apart and examine that afterwards. This is all very much a work in progress.

Stay tuned!

mcmaxmcmc

16 Nov 2018, 09:07

Very nice! Some can only dream to get a beamspring keyboard, let alone a 5100! Man, you're like the gainsborough of IBM -- they just keep coming at you!
Good job man. Hopefully the darn computer can be fixed.

User avatar
green-squid

16 Nov 2018, 13:59

Badass machine, would kill to own one!

User avatar
hansichen

16 Nov 2018, 14:06

green-squid wrote: Badass machine, would kill to own one!
In that case: good for me that I don't own one :lol:


I hope you'll be able to restore the machine. Good to see when keyboard collectors take care of machines and don't destroy everything like many other people do :(

User avatar
snuci
Vintage computer guy

16 Nov 2018, 14:36

Very nice! I am always a PM away for any help/advise. If you haven't found the maintenance manual, I scanned mine and put it here.

User avatar
//gainsborough
ALPSの日常

16 Nov 2018, 16:01

[el psy congroo intensifies]

Great write-up, dude! Love reading the stories that lead up to really neat finds - so thanks for the fun read! Also congratulations on scoring this epic machine!

User avatar
Sangdrax

16 Nov 2018, 16:14

Ancient number one rule of shipping: The more fragile stickers you put on it, the more it will be abused. The ideal number of them is 0.
Ancient number two rule of shipping: Heavy boxes will be abused anyway.


Anyhow, give the old girl a scrub, a recap and clean the connectors and see if that doesn't fire it up. Probably just some corrosion and stuff shorting to ground. Maybe replace the carbon comp power resistors with metal film versions if that doesn't work. They can drift a lot over the decades.

Man, what a find.

User avatar
JP!

16 Nov 2018, 18:45

I'd love to have one of these one day. Great story by the way. I think it would be worth the effort to restore.

User avatar
Bass

18 Nov 2018, 03:29

Thanks for the kind words everyone! Especially to snuci who has provided a link to the maintenance manual as well as some useful tips.

I have made a bit of progress on the restoration over the past two days. The first thing I decided to do was examine the card cage and motherboard. Just as snuci warned me about, a nasty surprise was lurking underneath:

Image
Image

This thick, tar like block of rotten foam was quite a pain to clean off. I am not sure what its originally functionality was, but my guess is that it was used to help dampen the sound as it's quite loud now. The portion on the case was simple to remove as it was attached via double sided tape, but the cards were a different story:

Image

Fortunately only one of the cards was loose. I have heard cases of all of them falling out upon flipping the board over, which can be problematic as every card must be placed in the correct slot on the motherboard in order for the machine to function.

Removing the rotten foam off the cards proved to quite labor intensive. To prevent myself from mistakenly placing them in the wrong slot, I decided to clean each of them off one my one, which made the whole process quite time consuming. Some of the foam also crumbled and fell on top of the motherboard slots, which was especially tricky to remove given how it got stuck in such a tight space. After about four hours of work using (probably too much) 90% isopropyl alcohol, this was the result.

Image
Image

Much to my frustration, I still couldn't get the display to turn on after all of that effort. I was initially worried that my cleanup effort to remove the spilled foam from the motherboard was insufficient and that I was getting bad connections, but as it turns out one of the cables was also loose. I did not notice this until removing every single card from the board at once to check. Thankfully, this fixed the issue, and now I can finally boot the machine up to reach the error message that's consistent with what the seller was getting before shipping to me.

Image

Of course, there was a bit of flicker in the display since the connection of the cable might be damaged or still a bit loose, but I am still very happy to have made some progress! Now that the display works, it should be much easier to track down the cause of the issue. Based on the manual, this error message corresponds to the ROS cards which are responsible for running BASIC and APL on the machine.

Anyways, I will need to examine some of the cards and see if cleaning them more thoroughly can resolve the problem, but in the worst case scenario I'll need to find a replacement, which can be both difficult and costly. I am thinking that for now I might want to take a break from the actual repair and get back to cleaning some other parts of the machine, namely the the exterior of the case and keyboard which I am sure is of great interest to everyone on here.

User avatar
zslane

18 Nov 2018, 19:26

Such a fascinating project! I can't wait to read the next update. Well done, Bass!

User avatar
Bass

19 Nov 2018, 20:24

I have made some progress in getting the rest of the unit cleaned up, starting most importantly with the keyboard. I wish I had more close ups of the beamspring goodness, but I realize I forgot to take those pictures after putting everything back together :(

Image
Image

I was surprised to find that both the plates and the stems were pretty much rust free, especially given how this was stored in a potentially damp attic for several decades. Nearly every beamspring I have ever found has had at least a few rusty stems, but thankfully I shouldn't need to take this one apart since it doesn't. The keycaps on the other hand were horrendously dirty, and quite a pain in the neck to clean:

Image

Furthermore like all other beamspring keycaps, the front legends are pad printed which makes them very prone to rubbing off when being cleaned. To prevent this I had to carefully wash these particular keycaps by hand, which ended up taking quite a while as all of the alphas on the 5100 have front printed legends. Although I think they could still be a bit cleaner, I am satisfied with how they came out for now.

Image

Finally, I also cleaned up the front panel quite a bit and as a result the whole machine looks much more presentable now!

Image

There are a few things left that I plan on doing:

1) Use a stronger cleaning solution to remove stains from the case. I have tried using 90% IPA, 409, and soap and water but those have all proven to not work as well here.

2) Find replacement APL and BASIC ROS cards. I have attempted to clean the contacts multiple times to no avail, and many sources I have read have suggested that these are both very prone to failure anyway. For now I have managed to find one for BASIC on ebay, so I'll most likely have to settle with having the machine only run in BASIC if I am fortunate enough to get it working.

User avatar
digital_matthew

21 Nov 2018, 22:15

The horizontal lines on the display are probably being caused by bad capacitors on the CRT control board. This is turning out to be one awesome restoration! :)

User avatar
snuci
Vintage computer guy

21 Nov 2018, 22:28

digital_matthew wrote: The horizontal lines on the display are probably being caused by bad capacitors on the CRT control board.
He's got the brightness set too high :)

User avatar
Bass

21 Nov 2018, 23:03

Yep, I have managed to make the flicker better by reducing the brightness and reseating the power cable again, it was still a little loose.

Unfortunately, I still can't get past the dreaded ABCDEFGHI 12 07 error even after trying a replacement card. Not sure if the contacts on the motherboard itself are the problem rather than the cards, or if this card doesn't work either :( . I am starting to have a hunch that a different card (ROS control) could be the culprit too, the manual is a bit unclear on this. I am thinking I'll either need to find yet another replacement, or see if I can clean the contacts on the board better with a more heavy duty contact cleaner and some smaller q-tips.

User avatar
snuci
Vintage computer guy

21 Nov 2018, 23:59

Bass wrote: I am starting to have a hunch that a different card (ROS control) could be the culprit too
I was afraid of that. An expensive goose chase, unfortunately. It also might have been the ROS controller (slot E) but buying one of those replacement cards and trying it is no guarantee either.

User avatar
JP!

22 Nov 2018, 00:02

Sorry to hear about these setbacks. Are there individual components on the cards which could go bad or out of spec?

User avatar
snuci
Vintage computer guy

22 Nov 2018, 00:08

JP! wrote: Sorry to hear about these setbacks. Are there individual components on the cards which could go bad or out of spec?
Unfortunately, the ROS chips are "read only storage" or proprietary PROMs that are not available. They were only available as swappable cards. There are TTL chips on some of the cards but they all have IBM part numbers to replace the normal TTL chip numbers (74 series, etc). IBM did not make it easy to fix these. Hence why there are not a lot that actually run.

User avatar
JP!

22 Nov 2018, 00:12

snuci wrote:
JP! wrote: Sorry to hear about these setbacks. Are there individual components on the cards which could go bad or out of spec?
Unfortunately, the ROS chips are "read only storage" or proprietary PROMs that are not available. They were only available as swappable cards. There are TTL chips on some of the cards but they all have IBM part numbers to replace the normal TTL chip numbers (74 series, etc). IBM did not make it easy to fix these. Hence why there are not a lot that actually run.
In that case we need a black budget, a laboratory, and a working 5100 to reverse engineer!

User avatar
Sangdrax

22 Nov 2018, 19:58

JP! wrote: Sorry to hear about these setbacks. Are there individual components on the cards which could go bad or out of spec?
The carbon comp resistors will drift in value a lot. The capacitors will dry up and short. Some of those actually look like ancient oil and paper ones. Yeesh. Solder joints will crack. That said, storage and ram chips go bad with disconcerting regularity on really old stuff. But you need a chip reader to dump code to, to see if things are bad if you don't have a working one to trade parts on.

User avatar
ikamusume

11 Dec 2018, 06:47

Hello. I just registered in this forum since I've repaired that kind of ROS error.

Hope to give some tips for restoration...

__red__

11 Dec 2018, 16:54

snuci wrote: Unfortunately, the ROS chips are "read only storage" or proprietary PROMs that are not available. They were only available as swappable cards. There are TTL chips on some of the cards but they all have IBM part numbers to replace the normal TTL chip numbers (74 series, etc). IBM did not make it easy to fix these. Hence why there are not a lot that actually run.
If we could find a working card of this type it may be possible to read those PROMs using modern equipment and emulate them.

I would love to donate data like this to the Internet Archive.

User avatar
ikamusume

15 Dec 2018, 14:33

__red__ wrote: If we could find a working card of this type it may be possible to read those PROMs using modern equipment and emulate them.

I would love to donate data like this to the Internet Archive.
We can dump its ROS cards in DCP mode. Just enter some hexadecimal codes(have to modify it for 5100) in memory directly, and execute it. Then we can dump ROS to its RAM, and one of the internal utility allows copying RAM data to a tape cartridge.

But modern computers can't read 5100's data cartridge, so we must transfer this data to a 8" floppy using IBM 5110. There's a program called IBMDISK which was used to read 5110 floppies on DOS machine.

Copying data from a 8" to a 3.5" floppy or a CF card is the last step. We can use this ROS dump with Christian Corti's 5110 emulator.

__red__

15 Dec 2018, 17:17

Do you have all the equipment to do this?
(Side-note: I acquired the contents of an S100 engineer's lab and have a metric-f-ton of S100 computers/cards/8" floppies/docs/etc that I need to re-home)

Dave_K.

19 Dec 2018, 19:21

First, what an awesome restoration thread! This kind of project would inspire me to also learn APL. :)
__red__ wrote: Do you have all the equipment to do this?
(Side-note: I acquired the contents of an S100 engineer's lab and have a metric-f-ton of S100 computers/cards/8" floppies/docs/etc that I need to re-home)
I'm working on restoring two S100 computers, and would love to learn what inventory you have. Not sure if you are on the VCF (Vintage Computer Forum), but that would likely be a good place to drop some inventory you are looking to sell.

Post Reply

Return to “Gallery”