On the left, the Morrow CP/M computer connected to the IBM terminal complained about all its boot floppy disks. It turned out one of the 5.25 inch floppy drives had failed — it no longer span its motor. Fortunately I had a spare drive of the appropriate kind used for the BBC Micro and was able to swap it.

In the center, the Apple II had for some time been displaying intermittent but increasingly common determination to ignore all the keys on one of its input lines, which unfortunately includes both enter and space. During during the visit of the IEEE club a couple of weeks ago that made it impossible to play Oregon Trail and the fault seemed permanent rather than intermittent. Could be a bad connection or a stuck key switch. Removing the cable connecting the keyboard to the computer and cleaning it seems to have fixed the problem, for now at least.

On the right, the Atari VCS suddenly deteriorated, refusing to work with several games and not always generating a picture. Testing it out, most cartridges didn’t work and the ones that did work had a fuzzy picture and were glitching. Also it had reverted to black and white. I took it apart. This is a later model, where the circuit board was shrunk to save money, leaving the woodgrain case mostly empty. (Atari eventually shrank the case too).

There is not very much to go wrong because the Atari has mind-blowingly minimal hardware capabilities, for example just 128 bytes (not kilobytes) of RAM. I cleaned bits of the board with alcohol and removed debris from the cartridge port with compressed air and tweezers. Then I took my can of Deoxit, which can work wonders, and sprayed it into the cartridge port. Then I removed all three chips from their sockets. I sprayed Deoxit into the sockets and around the legs of the chips and replaced them.

Shockingly, it worked. The color was back and the console worked property. Except, there’s a potentiometer (i.e. the round plastic thing in the bottom left) on the circuit board used to adjust the color. I’d turned it up and down – for a while it’s black and white and then colors appear, but as you turn it further the colors change totally. Not just more intense, but complete different colors. As most Atari games use a selection of psychedelic backgrounds for different levels it was hard to know whether the screens I was looking at in PacMan or Superman were supposed to be orange, blue or green. But then I remembered Frogger, and the green border where the frogs live. It turned out to be displaying as orange, but with a little adjustment was a nice bright green.

(The green is actually the same up and down the whole screen, but the camera captured the TV half way through refreshing the screen).

And as a bonus, the Pitfall cartridge I picked up at the Midwest Classic expo now works. It’s referenced in my book, so it’s nice to have one to demonstrate the complexity achieved through ingenious programming tricks later in the trajectory of the VCS.

The next week during lab hours I put it back together and Wyatt Kalmer called by to help fix one of the flaky joysticks it came with. The Atari VCS joysticks are iconic, literally in the sense that the shape is used to represent 1980s videogaming in general. (https://en.wikipedia.org/wiki/Atari_CX40_joystick ) But also bad: they dig into your hands, the fire button is in the wrong place, they are stiff, they are imprecise, and they tend to fail. In this case, the down movement was hit and miss. Opening one up for the first time, I was struck by the weird combination of chunky 1970s construction and extreme cheapness.

By modern standards the circuit board and case are solid and robust, but the design was clearly very cheap by the standards of it’s day. There’s a flimsy plastic collar that pushes down on switches to register your inputs, but the switches themselves are flimsy little metal domes that get pushed down to short out connections the board. Now I know why rival joystick suppliers always boasted about having “microswitches.” The Atari metal caps are held in place by nothing more than a piece of plastic that looks a lot like packing tape.

Over time the tape goes gunky and the switches get a little oxidized and no longer make good connections. So you remove a section of tape, clean around the little metal cap with alcohol and Q tips, and then apply new tape. Conceptually simple but very fiddly – especially as we briefly lost the little metal cap. But in the end, thanks to Wyatt’s patience, victory.

We tried to play Combat to celebrate. This brought a sad discovery. The repaired joystick worked fine, but unfortunately the joystick I thought was good is now having difficulty with the fire button. Apparently the fire button tends to wear down, which I hoped to fix by applying another layer of sticky tape to offset the erosion. Frustratingly, we found that the fire itself button worked fine but the problem came from the cable joining the joystick to the console. A long term solution will require either a new cable, fortunately spare parts are still being manufactured, and probably also a new circuit board for the first joystick. But we found a temporary solution: as the cable works only when pushed in one direction we applied some packing tape to hold it in a tight loop.

We then finished up the lab session with a nice game of Combat played for the first time in this lab with two original Atari joysticks. Both players can now experience an authentically uncomfortable experience. As they used to say on Prairie Home Companion: Duct tape, because in the end all solutions are temporary.

The post Atari VCS Maintenance appeared first on Retrocomputing History Lab.

Here’s what we had to say about this little-loved machine in A New History of Modern Computing (pages 222-223).

The XT set the template for office computing in the mid-1980s but IBM had not given up on the home market. About seven months after the XT, it announced the PCjr, a cut-down version of the PC in a smaller case with more equipment built onto the motherboard but less internal expandability. It had ROM cartridge ports for video games or programming languages and a wireless keyboard for use on the sofa. Before its launch, the PCjr was expected to dominate the home computing market, but it proved too limited for people who wanted a cheaper IBM PC and far too expensive to compete with the Commodore 64 for video gamers. The New York Times review dismissed its keyboard as “rather strange” with “all the grace of a rubber-kneed centipede.” “Two pages into a manuscript,” noted the review, “I was looking for a pad and pen with which to finish my story.”[i]

The PCjr has become one of business history’s most famously unsuccessful products, on a par with corporate disasters such as the Ford Edsel. It did, however, debut a new sixteen-color PC video mode intended to make video games more appealing. Software company Sierra On-Line worked with IBM to produce a game to show off the new graphics, delivering the first of the hugely popular King’s Quest “graphical adventures” designed by company cofounder Roberta Williams. Sierra’s adventures, which also included the humorous science fiction Space Quest and risqué Leisure Suit Larry titles, used a hybrid of video-game controls and typed commands.[ii] Unlike the Infocom games, which they quickly eclipsed, they relied primarily on pictures to describe the environments through which players traveled. The PCjr video mode is remembered as “Tandy graphics” because it was copied by Radio Shack for a popular line of low-cost PC compatible machines that replaced the aging TRS-80 series, succeeding where IBM had failed in bringing PC technology to the home computer market.

---

[i] Erik Sandberg-Diment, “The Little IBM Finally Arrives for a Test,” New York Times, December 27, 1983.

[ii] Sierra’s introduction of the graphical adventure, with an earlier game for the Apple II, is discussed in Laine Nooney, “Let’s Begin Again: Sierra On-Line and the ‘Origins’ of the Graphical Adventure Game,” American Journal of Play 10, no. 1 (2017): 71–98.

When I wrote those words I’d never actually seen one in person, still less had the opportunity to probe the rubber-kneed centipede with my own fingers. Last semester, however, Robert Polivka from Campus Technology Support dropped off a PCjr that had been lurking for decades somewhere on campus. None of the connectors are standard, so testing it meant purchasing some additional equipment. Last month I blew $35 on a power supply brick on eBay, and when I hooked it up to a monitor and saw a picture I spent another $50 on the famously awful keyboard. (IBM soon rushed out a less bad model to try to save the machine, but I wanted the authentic disaster).

Everything worked but the floppy drive, which span the disk but failed to read anything. Instead the computer gave up and went to the BASIC programming language burned onto a ROM chip, so that people who purchased the cheaper option with no floppy drive could still use the computer. They would have had to hook up a cassette recorder to load and save programs.

At launch the diskless version cost $770 (around $2,000 with inflation) and the system I had would have cost $1,269 which, with inflation, is almost $4,000. Not exactly an impulse buy, except for the wealthiest of families.Though IBM eventually slashed prices massively to clear out hundreds of thousands of unsold systems at a heavy loss. See https://en.wikipedia.org/wiki/IBM_PCjr.

Note the lack of labels on the little white keys, which explains why this is often called the “chicklet” keyboard. They key’s themselves didn’t feel as bad as I’d expected from the review, nothing as horrific as the Sinclair systems I grew up on. But between the tiny labels and the fact that many regular PC keys are missing and require key combinations to access it’s not a great experience. Like old school remotes they keyboard only works if pointed straight at the front of the computer, which is easier to do from a sofa than at a desk. The keyboard could be connected with a cable, but it’s an optional extra.

Fortunately I wasn’t restricted to loading and saving BASIC programs from tape. Pushing the disk drive’s read head up and down its shaft and applying some grease fixed that and it successfully booted from a DOS disk.

Considered as a PC the PCjr is indeed crippled. IBM released two versions, and I had the high-end one with a single floppy disk drive and 128KB of RAM. It defaulted to chunky 40 column text, legible on a TV but hard to work with.

Pulling out a floppy disk of mid-1980s PC games I found that several of them refused to load because of the memory limitation. PCMAN, a Pacman clone, ran at glacial speeds. In a cost saving feature, the main memory of the PCjr doubles as video ram, whereas in other PCs the video circuitry was a separate circuit board with its own memory. This burdens the processor, making the PCjr pretty much unique as a desktop PC compatible running slower than IBM’s original 1981 model.

Only one game on the disk ran well, Alley Cat which was released by IBM itself and had special support for the PCjr sound and graphics. For example, the background was a restful blue rather than the usual horrific bright pink.

While IBM itself initially offered no upgrades for the PCjr beyond an unbelievably slow internal modem card, a joystick and an add-on printer port other companies produced a variety of add-ons that plugged into the side of the computer or sat underneath it to provide amenities such as additional RAM, a second floppy drive, a regular keyboard, better compatibility with the full size PC, and so on. Even interfaces for hard disk drives. These can be purchased at high prices on eBay, and one company makes and sells a modern board that provides most of the missing capabilities and lets you use memory cards in place of a hard drive. https://texelec.com/product/jr-ide-for-the-pcjr-by-retrotronics/ But I have a non-crippled IBM PC Portable in the lab already, so keeping the PCjr authentically useless seemed a wiser and cheaper course.

What I wanted, of course, was King’s Quest, a game that along with Zork and Oregon Trail is one of the experiences most requested in the lab by visitors who grew up with computers in the 1980s. When demonstrating the system in class this will be the thing to show. While the game is readily available online (you can buy remastered versions of the entire trilogy for five bucks) I needed the original version, which only runs on the ill-fated PCjr. I eventually found a downloadable disk image online, suitable for writing back out to a physical disk, but my usual WinImage software for Windows 2000 refused to recognize it. A bit of Googling revealed that the disk is in a non-standard format that can’t be read by DOS. The computer has to be booted from it, partly so that none of the tiny memory is take up with DOS and partly to make software piracy harder to force people to buy their own disks. After find the appropriate ancient DOS software and settings to write this image format I was able to use the Gateway 486 to prepare a disk. It looked blank to DOS, but after a few attempts the PCjr booted successfully from it.

King’s Quest showed the computer in a whole new light. Seeing it running on the original hardware made me realize how the game had been designed to minimize the limitations of the jr and play to its strengths. The game booted with a short blast of trumpets, showing off the jr’s relatively good sound. Its graphics were colorful and worked well on the TV monitor it was hooked up to. Regular PC owners would have to wait until the late-1980s and the arrival of affordable EGA graphics cards to upgrade their machines for similar results. Indeed, the game was seen at the time as a technical triumph. https://www.atarimagazines.com/compute/issue57/kings_quest.html

Limited memory and sloth meant that moving from one screen to another took a couple of seconds to whirr the floppy drive and redraw the screen, but the game moves so slowly that you hardly notice. Grahame, the protagonist, moves around the screen at a relaxed amble and because he’s liable to die if he steps one pixel off the correct path players move him cautiously. Animating a few areas of the screen, like flags on the castle and a deadly wolf, do not overtax the system. The game stores its colorful screens as drawing instructions rather than bitmaps, letting it cram 80 graphical scenes into the relatively generous storage of the 360KB floppy disk. Sierra games running on much more powerful hardware continued to work in much the same way for the rest of the 1980s, embedding the limits of the PCjr into the experiences of a generation of gamers.

The PCjr sits next to an Apple IIe, released in 1983. That physical conjunction helped me understand why IBM, and much of the computer industry, could originally have seen the PCjr as a strong contender for the home market. On launch the PCjr cost hundreds of dollars less than a comparable Apple system, though Apple reacted by offering cheap bundles of the Apple IIe with disk and monitor. The PCjr is no slower than the Apple IIe, has better graphics and sound, is less quirky, is easier to use, and with the wireless keyboard can actually be used in the living room from the sofa. IBM crammed more than twice as much data as Apple onto the same floppy disk.

The PCjr could run some IBM PC software, unlike the Apple, but in practice this was a mixed blessing. For the Apple IIe having 12KB of RAM was a big step up from earlier Apple models, allowing it to run pretty much all Apple II software produced since 1977. For the PCjr a memory of the same size was a major limitation, making it unable to run the most popular IBM PC titles and leaving it with a much more restricted range of software than the Apple. Not to mention the much cheaper Commodore 64, which was better suited than either the Apple IIe or the PCjr to videogaming, the actual killer application of home computers. In as much as there was any compelling application at all for home computers, because it turned out that the home computer market was collapsing anyway and most gaming enthusiasts (in the US at least) were about to shift their loyalties to Nintendo and Sega.

Hence the PCjr’s enduring status as a famous failure. But seeing King’s Quest next to comparable Apple titles of the same era like Ultima III let me understand why IBM didn’t see the disaster coming, and indeed to wish that more programs had been written to showcase the strengths of the PCjr. If I ever update the book hopefully I’ll be able to add a sentence or two to reflect this experience.

The post IBM PCjr appeared first on Retrocomputing History Lab.

You can see it advertised here, https://archive.org/details/pc-computing-magazine-v6i9/page/n61/mode/2up as Gateway’s flagship model. The advertisement is remarkably text heavy.

In those days, Gateway and Dell were neck and neck as leading mail order providers of solid but well priced custom built PCs to businesses and upper middle class families. Computers and monitors arrived in boxes with bold cow-style spots, celebrating Gateway’s origins in Sioux City, Iowa where the company was originally based in a “dilapidated cattle brokerage” (https://en.wikipedia.org/wiki/Gateway,_Inc).  A computer exactly like this was still in use in the Real Estate center of Penn’s Wharton school when I designed a database application for them in summer 1996, though as newer models had come in over the next three years it had been passed down to Gertrude the receptionist. To support the new system I upgraded it to Windows 95 which it ran tolerably.

To purchase a PC in 1993 you had to know what all that jargon meant, otherwise you’d have no idea if it was the right configuration for you and if it was a good deal vs. options from hundreds of other brands. Buying a PC basically involved picking a long list of standard components which the company would screw together and set up for you. Standard configurations changed month to month, even from the same company. The owner had added the optional model (card on the left) for an extra $200, allowing for dial up Internet access and the sending of faxes.

Here’s what it looked like when it arrived, with a lot of accumulated dust which, among other things, was preventing the CD-ROM tray from closing properly.

After cleaning everything worked great, but the included single speed CD-ROM drive hooked up to the SoundBlaster 16 sound card proved too limiting for the 1993-5 Windows 3.1 multimedia experience that the machine would otherwise have excelled at. With the huge storage capacity of CD-ROM, videogame designers and multimedia producers were very excited about the potential of putting full motion video into everything. Alas, playing to play it back properly at full screen size needs at least a double speed CD-ROM, which was available in 1993 but not fitted by Gateway as standard until later in the year. Fortunately we had a slightly older 486 computer from 1992 upgraded with a  4X CD-ROM (19994) salvaged from UWM Surplus. I swapped them over and verified both systems worked. The hard drive was a time capsule with a selection of mid-1990s applications installed, including the various utilities like Trumpet Winsock needed to connect to the Internet before those capabilities were integrated into Windows 95.

Alas, just as I was ready to put the cover back on and move the system off my crowded little workbench it abruptly started to hang every time I tried to launch Windows. Hours of historical reenactment of the huge frustrations of multimedia PC ownership  by changing BIOS settings, removing things from CONFIG.SYS, running Norton Disk Doctor, etc. revealed that the problem was corruption of several directories, including many of the files in the Windows directory. I ordered a replacement hard drive, which worked, but tests suggested that the original drive was also fine and just the file information had been corrupted in the file allocation tables DOS uses to track what is where on a disk. Norton Disk Doctor helped make the directories readable again, but several vital files were unusable. Dusting off my Windows 3.1 floppy disks to reinstall Windows and downloading the appropriate graphics drivers eventually worked at fixing Windows without wiping the whole drive clean and losing the historical mix of period software.

I think what had gone wrong relates to the battery used to preserve system information while the computer was turned off. This failed decades ago, and those batteries (from Madison-based Rayovac) are no longer produced. The replacement I rigged up with an AA battery holder ordered from Amazon, turned out not to be working because the new battery holder was faulty. That’s a sad comparison between modern equipment and the quality and durability of the thirty year old Gateway, but fortunately it was a four pack, and the next one did work. But in the meantime the computer was setting its configuration somewhat randomly every time it was turned on, and on one occasion it picked settings that pushed the hard drive controller too fast and corrupted part of the disk. Fate may yet prove me wrong, but other than that the drive itself and every part of the computer still seem to be working fine, validating Gateway 2000’s reputation for picking quality components.

Here is it set up in its new home, paired with its original keyboard and some Gateway branded speakers I’d earlier found in UWM Surplus.

The Gateway “Crystalscan” 15 inch monitor it came with is smaller and fuzzier than the beautiful 17 inch model I was using previously, but gives a much better idea of what ordinary computer users were dealing with in the period.

Also note the extra keys on this special “Anykey” keyboard, a Gateway 2000 signature in the period. To the left there are function keys in two rows, as on the original PC, as well as in the usual place above the number keys. There are diagonal keys as well as the usual up, down, left, and right arrows. In the top right, an extra block of keys records macro sequences of key presses into a small memory built into the keyboard itself. A useful feature in the days of DOS, less so with the shift to Windows.

The post Gateway 2000 4DX2-66V appeared first on Retrocomputing History Lab.