- Details
- Hits: 56
I'm interested in the nuanced release structure of the early 1990s IBM Thinkpads. Whereas the enterprise and business range of 7xx models was fairly sequential and linear (the model numbers from 700 to 770 are chronologically ordered and each machine is a new iteration of the last), the consumer 3xx models have a convoluted, non-linear development and release timeline.
It's also interesting how these two different model ranges interact when it comes to sharing designs, technologies and parts. You can see by some of the tiny details that there were probably lots of very detailed product meetings, where management poured over how one machine might rob sales from another or the unwashed masses might regard a cheaper IBM Thinkpad-branded machine as a 'better deal' than another Thinkpad at twice the price.
This makes restoring and preserving machines quite difficult, fraught with traps and gotchas.
What led me down the path of the title of this article is that I encountered a somewhat unusual 750 (9545-006/8) machine with a monochrome STN panel. I suspect this machine was not popular as by the time this machine was released in September 1993, both DSTN and TFT colour screens were considered infinitely more cutting-edge and quickly became de-facto standard of portable machines. I'm generally not happy to own a machine that doesn't work - unless it really really has tested me to the limit of writing the whole model off as a lost cause - and so I set about trying to make this machine work.
The 750 was really when the 7xx series got goooooood. The 700 and 720 machines are blighted with a terrible hinge design which is destroying every last one of them. And whilst some of the design ideals of the 700 and 720 are lost in the 750, it's certainly an revelation. Inside the circuit boards have clearly moved on a generation from the somewhat 'mechanical' look and feel of the 700 and 720. We have lost the front-loading floppy, battery and hard drives hot-swap ports, but these have now been almost as conveniently moved underneath the lifting keyboard design - retaining the idea of not elbowing your fellow passenger on a bus, train or plane whilst switching out your battery or HDD - a concept which would remain for many more generations.
In this case of this particular 750, this is how the machine looked on ebay - bloody awful. So I suspected I would get it for what it's worth - not a lot. And I did - about £21 and shipping fees from across the Atlantic. Vintage laptop aficionados will instantly recognise the vinegar syndrome LCD - a problem where early LCDs use a formula of adhesive to affix the two layers of polarising film inside and outside the glass LCD screen, which breaks down and fails - particularly in warm and humid environments. Secondly, it had no guarantee of actually working.
I received the machine and quickly checked the machine powered on - which it did, booting via the external SVGA port to a few error codes. I decided then to see if the rare monochrome STN LCD panel was salvigable. There are very few resources which give -true- details on what panel was used in what machine. It's clear that through the evolution of the IBM thinkpad, they uses a wide range of manufacturers - Sharp, Toshiba, Hitachi, Epson and others - so often the only way to know is to look.
It's a Toshiba display which cannot be found anywhere. So a direct replacement was out of the question and I set about Plan B - to remove the spunky dried up adhesive and replace the two polarising films. This takes about 3 hours of applying IPA, letting it soak for 30 seconds and then scraping as much of the glue into the centre of the screen and then lifting it off. After doing this for hours, you've left with what looks like a simple pane of glass, which you can insert polarising films either side (usually at 45 degree angles) to get the mono display working again. And after all this, it was totally dead.
So, Plan C. I knew that the 360 was a cost-reduced, consumer-end alternative to the 750, sharing lots of the same design and technology, However, there was also no clues online as to which display this used either. And I didn't own one. So, Plan D - the 340-series was released the same year as both the 750 and 360, and whilst it shares almost none of the same technology of those machines, it did have a mono display which was the same size. And crucially I owned 3 of them! So I took apart one of my 340 machines and the display was the exact same dimensions, connector, inverter, but a completely different make and model:
So would it work?
YES! It did! These LCDs are actually much more readily available, except at a higher cost than a 340. Much as it pains me to turn a working 340 into a non-working, no-display 340, this project is to restore a rarer 750 to working condition.
However, it would be worth noting that the 340 has a completely different internal video connector (specifically not via an interposer as used by the modular 750 and 360-series machines) and therefore whilst the ribbon connector from the mainboard to the LCD and Inverter is almost exactly the same, the connector onto the mainboard is completely different and cannot be used.
I believe the inverter is interchangable.
I'm also happy to report that the Optiplex 340 panel shared all the same dimensions and fit quite easily as a replacement for the 750's Toshiba panel. The only difference was that the 340 LCD had a thick, bendy RF shield all around it which could not be accomodated in the 750 lid as it has plastic strengthening strips preventing the RF shield laying flat.
The inverters only had 1 small side-mounted electrolytic cap on them which didn't appear to have been leaking but if you can be bothered, it might be worth replacing with a new one.
With the 750 lid back to full working order, I decided to crack on with the mainboard. The first oddity is that it would show the usual memory count and 161/163 (cmos dead) errors, but would not allow me to hold F1 to enter easy setup. I tested the F1 key and it was definitely working.
So I tried booting from an early IBM utility disk (used to update the BIOS) and it worked. However if you didn't have an IBM utility disk in, it would not enter easy-setup and would not boot off any drive, just showing an error I9990303 error.
There have been quite a lot of people over the years asking about this error, which is certainly related to the system board, but nobody ever appears to have fixed it.
I thought I would update the BIOS to see if that might fix it, but when you try, even with a charged battery, you get an error saying that 'power management has been disabled' preventing the BIOS update to take place.
The closest we have to an answer is "Bank–2 flash ROM check sum error." as described here.
So the machine has been parked. Waiting for a busted up 750 to come on the market, with a working system board.
- Details
- Hits: 53
I originally intended to spend the morning finishing restoring a 755CD which has been 'on the shelf' for about 2 years with an on-board 00201 memory issue. It ended up taking exactly 12 hours even after last night I'd already dismantled and resolved the issue with the memory.
The 755CD is quite different from all the other 755 machines because it is a fatter chassis - to incorporate the CD-ROM drive and, I think more so the massive video and digital sound processing hardware that accompanied this 'multimedia' expansion. Due to the battery and HDD being standard across all 755 machines, the 755CD has a plastic 'spacer on a hinge above these two bits, to keep everything in place.
The 486 755CD and 486 CE were released in October 1994, the Pentium CX, 486 CV and 486 CDV were released in May/July 1995, the Pentium CD, CDV and CV in Sept/Oct 1995. A Pentium 755CD with the SVGA CX screen would be the ultimate 755+ model. The 755CD & CDV were the only ones with the thick chassis.
All the other 755 machines came with floppy drives by default. Other things i've learned...
The 755CD was obviously the most advanced 755 machine, but it wasn't the highest specification by any means. Only the 755CX (which lacks the CD, DSP and other video enhancements) was the only 755 machine offered with an SVGA screen. This really surprised me that the 755 (out in 1994) was still defaulting to 640x480.
After much deliberation I decided to upgrade this 755CD to an SVGA panel from a CX. I always try to keep these machines 'original' but if you put it back to VGA then I guess it is.
As I said at the start, the onboard memory was duff. All 755 machines came with either 4 or 8Mb onboard, but later models had the onboard RAM removable via a little daughterboard which is only accessible by completely dismantling every. single. component. of the machine. Luckily I had a couple of spares and a replacement sorted that. Here is what it looks like:
This is the spacer only found in the 755CD:
If you look at the shield over the system board(s), though, you'll see that the innards are at normal height, but for the upstand at the back increased - is that green/white bit the switch so it knows when the keyboard is lifted? That would need to be higher to touch still, like the volume wheel projecting higher.
And here is my 'restored' machine. It's not exactly perfect - mainly because the plastic on these machines doesn't get brittle but the coating does not fare well.
Here is the difference in thickness between the 755CD on the right and a 755C on the left:
The internals of the 755CD is made up of about 5 different boards sandwiched together with the CPU attached to a wide flat heatsink right at the bottom. Each layer above that has a variety of surface mount caps which were all a bit dodgy and I replaced with ceramics. About 20 in total, including a couple of through-hole caps on one of them.
The LCD as well had 3 surface mounted electrolytics - the type that are harder to spot initially because they're encased sideways inside a plastic box. I replaced them too because the screen was a bit flickery.
One aspect that slowed me down was that the audio board (with the rolling volume control) has about 5 surface mounted caps and one of them pulled the pad off. I managed to botch wire the missing pad but then decided to swap it with a spare audio which looks visibily the same which did not need the botch. Then after re-assembling I noticed that due to the whole chassis being taller, they'd made the whole stand which the volume wheel sits on about 5mm taller! So I then had to desolder both wheels and switch them.
Upgraded with a 16Mb IC DRAM card, giving 24Mb RAM and a clean installation of Windows 95. Drivers were a pain to get from a variety of breadcrumbs around dead links on the internet - but they are there.
I tried to use the archive.org ISO of the original restore CD (which would have brought back a dual boot OS2/Win3.1 installation on a 320Mb partition, but despite considerable resources it wouldn't work. You have the option to use the built-in CD-ROM but you are required to use a floppy boot disk - but obviously you can only have one in the ultrabay at once, and you can't hot swap. Usually this would be no problem as the 755CD does have an external FDD connector but it's unlike almost every other thinkpad I've seen. It's small and fat and I couldn't find that cable anywhere.
So the alternative was to boot using a massive Selecta Dock II, which has a built-in CD-ROM drive, which I have and I tried using but it would not detect the CD-ROM drive in the dock.
I also tried copying the files from the boot floppy to the HDD and then running those from a bootable HDD, with the built-in CD-ROM to work but no-go - the bootable floppy uses some hidden magic formula that only IBM know about so you definitely need both.
Still Windows 95 looks rather 'charming' on this machine and it would have been the obvious 'upgrade' for most owners to make. And all the drivers are there so. Yeah.
Oh and for some reason, only the right speaker works. Kinda annoying but we can't exactly allow perfection to be our enemy here.
Anyway, this is the only 755CD I've ever come across. I'm sure they aren't that rare but some of them did burn up on MIR's re-entry, so a few less.
The 755CD actually incorporates TWO of those little green & white switches to know when they keyboad is raised - the second one sits ontop of the first one (first one found in all the other 755CDs) keeping the first one pressed down and, my guess was replacing it at a higher level. Quite Heath Robinson, but it works.
From memory, the aluminium cage that surrounds most of the innard is different on the 755CD to all the others too. In other words, on all the other 755 machines the interposer for the LCD is just a 'dumb' looking riser, but on the 755CD it's a huge PCB, covered in big fat chips taking up the whole top of the chassis.