The IBM PS/2 line, released in April 1987, was IBM’s attempt to reinvigorate its aging personal computer line and fight off cloning. Although the line sold fairly well, it failed to hold off cloning and IBM never regained the market dominance it enjoyed in the first half of the decade.
The IBM PS/2 did offer numerous enhancements over the PC line it replaced, but IBM’s customers came to resent the high price and the perception that IBM was trying to lock out third party peripherals. IBM’s decline was slow, but the PS/2 was the beginning of the end for IBM’s personal computer business.
The IBM PS/2 product line
Initially the line consisted of the IBM PS/2 Model 30, 50, 60, and 80. The Model 30 featured Intel’s 8086 CPU and replaced the PC/XT. The Models 50 and 60 featured Intel 80286 CPUs and replaced the PC/AT. The Model 80 featured Intel’s then-state-of-the-art 80386 CPU and served as IBM’s answer to the brash Compaq Deskpro 386. All of the models were faster than their predecessors, but in early reviews, the PS/2 Model 80 didn’t match the Deskpro 386’s CPU performance. But its faster hard drive compensated somewhat.
Soon afterward, IBM released a Model 25, which was a smaller 8086 with a built-in monitor, and a Model 70, which was a 386-based desktop computer. Some believe the oddball 7496 Executive Workstation was a proof of concept for the Model 25.
Over time, IBM released other in-between model numbers with new CPU technology. Late in its life, IBM used its own 386SLC and 486SLC2 CPUs in desktop PS/2s with model numbers like 55, 56, and 57. IBM used a lot more of its own parts in the PS/2 line than it had in the PC line, which led to higher profit margins.
The PS/2 Model 30 sold for $2,295. It included an 8 MHz 8086 CPU, 640K of RAM, a 720 KB floppy drive, and a 20 MB hard drive. The monitor was extra. A version without a hard drive sold for $1,695.
The PS/2 Model 50 sold for $3,595. It included a 10 MHz 80286 CPU, 1 MB of RAM, a 1.44 MB floppy drive, and a 20 MB hard drive. The monitor was extra.
The PS/2 Model 60 sold for $5,295. It included a 10 MHz 80286 CPU, 1 MB of RAM, a 1.44 MB floppy drive, and a 44 MB hard drive in a gargantuan tower case.
The PS/2 Model 80 sold for $6,995. It included a 16 MHz 80386 CPU, 1 MB of RAM, a 1.44 MB floppy drive, and a 44 MB hard drive in a gargantuan tower case.
IBM hired the cast of M*A*S*H, minus Alan Alda, to star in one-minute commercials for the PS/2. The commercials ran in prime time, in heavy rotation.
IBM also did a lot of product placement in TV and movies. At the end of every episode, Doogie Howser M.D. typed in his journal on a PS/2. In the sitcom Cheers, there was a PS/2 in the back office and it sometimes even played a role in some episodes’ plots. There were others.
Although the list price on the machines was high, IBM discounted relatively heavily. In February 1988, dealers were selling most PS/2s at a 37.5% discount. According to the April 9, 1988 issue of Infoworld, IBM sold 2 million PS/2s in the first year, but the PS/2 didn’t accomplish its goal of slowing down clones. Instead, many large corporations started looking at competitors such as Compaq. According to the June 29, 1987 issue of Infoworld, Compaq’s sales jumped 47 percent the same quarter the PS/2 came out. The abrupt change to 3.5-inch floppies was one factor. Passing floppies around was still the major form of data transfer and pre-1987 PCs all had 5.25-inch drives. Adding the optional 5.25″ drive to a PS/2 was an additional expense and took up extra desk space. Compaq and other PCs could accommodate both types of drive within the case.
IBM marketing vs Apple
With the PS/2, IBM tried to shift the market from one disk format to a newer disk format. Apple did similar things all the time after Steve Jobs returned to the company, like shifting all the peripheral connectors to USB and eliminating floppy drives altogether on Macs from the translucent era, and removing the headphone jacks from phones in more recent years. The market tolerates this from Apple, but hated IBM for it.
One possible explanation is Simon Sinek’s Golden Circle theory. Apple marketing starts with a grand vision as its why statement, then progresses to how they build things and what they build. Accepting that somewhere between 16 and 33 percent of the market will never buy from them is part of the strategy. Apple thrives on capturing somewhere between 18 and 84 percent of any market it enters. This formula lets them build a devoted, zealous fan base willing to pay high prices.
IBM marketing, like most companies, goes the other direction, starting with what they build, then how, then why. IBM wasn’t content at all with the idea of having 18 percent of the computer market; the 1987 IBM behaved like 90 percent of the market was its birthright and did little to try to earn its customers’ trust. That’s why IBM never had the kind of cult following Apple enjoys.
IBM’s message was more like this: We’re IBM. We’re the best. Buy our stuff. Once 18 percent of the market realized they could continue to buy cheaper clones, it created a tidal wave IBM couldn’t recover from.
The IBM PS/2 operating system
Initially when IBM launched the PS/2 line, all PS/2s ran IBM PC DOS 3.3, which was IBM’s version of Microsoft’s MS-DOS operating system.
The similarity between the PS/2 name and OS/2 operating system is no coincidence, however. OS/2 was the future, as far as IBM was concerned, and IBM didn’t mind if people thought you had to buy IBM to run OS/2.
OS/2 came out in December 1987, but only ran on 286 processors or faster, so you needed at least a Model 50 to run it. The first truly usable version of OS/2 was 1992’s OS/2 version 2.0, which came out in 1990, and needed a 386. By 1990, Windows was starting to look unstoppable.
Of course, IBM PS/2s ran Microsoft Windows just fine.
IBM PS/2 features
IBM had to strike a balance with the PS/2 line. They wanted to improve upon the success of the previous PC line without breaking compatibility with it. In that regard, the PS/2 line was a success.
For the most part, PS/2s were much more tightly integrated than the earlier PCs. Serial and parallel ports and the disk controller were on the motherboard now, instead of in slots. Usually, video was on the board too. Because of this, desktop PS/2s tended to be much smaller than a PC/XT or AT. The towers were still behemoths, but they had a lot more room for network cards and hard drive host adapters.
The PS/2 line introduced MCGA and VGA video standards, bringing 256-color video to what had been a tired product line. It also replaced 5.25-inch disk drives with 3.5-inch drives like the Macintosh, Atari ST, and Amiga used. This made the PS/2 look less old-fashioned, but as I noted before, it made transferring data more difficult.
The other big innovation in the PS/2 line was a new expansion bus called Microchannel, also known as MCA. Reserved for the 286 and 386-based machines, it was a faster 16/32-bit bus that ran faster than the previous ISA bus and also had something resembling plug and play capability. Rather than mess with DIP switches, you loaded an option disk to configure your expansion cards.
Microchannel’s mixed blessing
Microchannel was superior and it solved real problems. A simple task like adding a network card could consume a weekend with ISA if you had a fully loaded system. Microchannel reduced it to less than an hour.
I dealt with a lot of Microchannel at the beginning of my career and it was much nicer to work on. Imposing less system overhead was nice too, but eliminating the tedium of all those DIP switches was the best thing about it. By far.
The problem with Microchannel was that IBM designed it as a closed system. IBM was happy to license it, but IBM wanted a royalty on ISA systems too. Non-IBM Microchannel systems existed, but they were rare indeed. ALR, Mitac, NCR, Olivetti, and Tandy all made Microchannel systems, but didn’t go all-in. The first 486 system to hit the market used Microchannel, but soon found itself in the minority.
The rest of the industry designed alternatives: first EISA, then VESA, then PCI. And since a usable version of Windows didn’t appear until 1990, the industry had time to catch up with alternatives. Had a usable, high-system-demand operating system appeared sooner, Microchannel might have done better.
Or maybe not. It’s usually the cheap, good-enough solution that wins. Otherwise, I’d be writing this on an Amiga right now.
But I never really addressed that cloning thing. Rather than try to clone the PS/2, the rest of the industry just kept cloning the earlier XT and AT, and bolting extensions onto them. Basically, the cloners did what they would have done if IBM had ceased to exist in 1987.
IBM’s retreat from Microchannel
The public saw Microchannel as an anti-consumer, monopolistic money grab. IBM retreated from it somewhat with the release of the PS/2 Model 30-286, which was a 286-based PS/2 with ISA. In 1990, IBM released its PS/1 line, which was ISA, and in late 1994, it replaced the PS/1 with the Aptiva line. For businesses, IBM for a time offered its PS/Valuepoint line. It, too, was ISA-based.
Microchannel dealt IBM’s reputation a blow. Even in 1994-95, when I sold computers at retail, selling IBM products was an uphill battle. There was nothing proprietary about the PS/1 and Aptiva lines by then, but some consumers wouldn’t even look at IBM. I could sell Thinkpads, but the IBM desktop computers were sales-resistant. OS/2, even though it was working really well by 1994, was an even tougher sell.
My impressions of the PS/2
As a guy trying to sell computers at retail in 1994 and 1995, the PS/2 made it awfully easy for me to sell Compaq, Dell, and HP computers.
As a technician working in the bluest of blue shops in 1995 and 1996, I appreciated the PS/2s, even if decommissioning them was a big part of my job. They were really sturdy, and they were easy to work on. Some PS/2s didn’t require any tools at all to disassemble them for upgrades or service. In that regard, working on them was more like working on late 1980s and early 1990s Macs than earlier PCs.
Most PS/2s were pretty underpowered by mid 1990s standards, and I had a really hard time getting late versions of OS/2 to run well on them. Windows 95 didn’t run all that well on them either. But they were obsolete, not broken.
I was impressed with how well these machines had held up, as some of them dated to 1987 and 1988. Then again, as expensive as they were, they had better have been. Y2K was the end of the line for them. We could have made them Y2K compliant, sure, but Y2K provided a convenient excuse to replace them with newer stuff.
The PS/2’s legacy
IBM pulled out of the personal computer market in 2001. I think IBM would have pulled out at some point anyway, but IBM’s reputation for being greedy accelerated it. IBM intended to use the PS/2 to stop clones but it had the opposite effect. Since it was a radical departure from the PC, it gave people a reason to consider clones. Prior to the PS/2, clones had a bit of a stigma. You bought a clone if you couldn’t afford a real IBM. Or if you were too much of a tightwad to pay IBM’s prices. After the PS/2, you could dismiss that by saying you bought a clone because you needed versatility that you couldn’t get in a PS/2.
I think the PS/2 did more to end the gray market than it did to end clones. IBM didn’t like the gray market either, but it made some money off gray market PCs. It didn’t make anything off clones.
Still, the PS/2 line has an enduring legacy. VGA existed before the PS/2, but it was the PS/2 that got people talking about it and wanting it. Thirty years later, some computers still have that 15-pin VGA port. And the PS/2 did accelerate the adoption of 3.5-inch disk drives. Once the PS/2 got them, it wasn’t long before everyone else started replacing the second 5.25-inch drive with a 3.5-inch drive.
Likewise, the PS/2 keyboard and mouse ports originated with the PS/2, hence the name. Not all computers still have those ports, but many do. And the 102-key PS/2 keyboard layout survives with very minor additions to this day.
It’s a significant legacy. It’s just not the legacy IBM wanted.