Teardown! Inside PC Labs' IBM PC Model 5150

If you were to make a movie about the 40th anniversary of the IBM PC (August 12, 2021), my vote for a title would be: There Will Be Dust. That's because it was my job to spiff up the old IBM PCs in PCMag’s PC Labs, bringing them back to the light for the big day. Achoo!

Over the decades, three 1980s IBM PCs, each with a matching monitor, have presided over PC Labs, stowed on a high shelf. Museum pieces, they’d survived actual use, multiple office moves, and the screwdrivers of countless curious hardware analysts. They’d been around so long that no one on 2021’s team could recall when last they’d been taken down, dusted off, and marveled over. Indeed, no one was even certain which models, exactly, they were.

I've spent many a day assembling and tearing down PCs, but I’m not a product of the initial PC generation. My first computer, circa 1984, was a cheap Commodore VIC-20. On a black-and-white TV, I keyed BASIC programs into it from magazines, saving them to the fussy cassette tape drive. (All the while, envious of friends with the much more capable C64 and its external floppy.) I splurged for an Apple Mac LC desktop in 1991, then a Mac clone (anyone remember the UMAX SuperMacs?), followed by my first PC clone...and I never looked back.

All this is to say: I missed the first 10 years or so of growing pains with the IBM PC. Why not catch up for lost time in 2021? 

So, What Exactly Are You, Anyway?

I hauled PC Labs’ three IBM PCs down from storage, and about five minutes of Googling, poking, and prodding revealed that two of the PCs were IBM PC XTs, the second-generation model. (The little “XT” badge on the front should have been a dead giveaway.) Back on the shelf they went. We’ll save those two for the PC XT’s 40th anniversary in March 2023. But one was, indeed, an IBM PC Model 5150—the family name of the first IBM PC. (Hurrah!) But it was semi-trashed inside and out. (Boooo!)

PC Labs, in the distant past, may have had a wild party inside its relic Model 5150 for the IBM PC’s 20th or 25th anniversary. One of the floppy drives had been removed, and was sitting alongside, possibly not an original. Some wiseacre had inserted an HD-DVD optical drive in half of the double-height 5.25-inch drive bay. (Remember HD-DVD? You probably shouldn’t.) The other half had an ancient, filthy 5.25-inch Seagate hard drive sitting in it, sliding around loose. 

Taking off the lid, inside the 5150 was like the aftermath of a frat party. Brown goop had taken up residence on some of the ribbon cables...somehow, still gooey. A Coca-Cola spill? (Given the age, maybe New Coke.) Battery acid? Whatever it was, alcohol wipes took care of it. (Ewwww.)

Rattling loose in the case: an interface card for the Seagate hard drive. Literally loose—like loose change. Took that out. Also not a good sign. 

What we were left with was a well-used, much-abused IBM PC Model 5150. A company bar code from Ziff-Davis (PCMag’s parent company for many years) indicated that this was once IT property, some long-ago editor’s workhorse. Some prescient PCMagger had saved it from the scrap pile for posterity.  

Identification of a PC

A little scouting around the motherboard showed up this tidbit along one edge: “64KB-256KB CPU.” This indicated a later revision of the Model 5150. Alas, our IBM PC was not an “original” original, an OG 1981 PC. It had been assembled a few years after.

Best guess: 1984. A fabric tag inside attached to the monoaural speaker noted that the speaker had been manufactured—or the PC had been assembled—in September 1984, perhaps in Florida. The tag was unclear which.

The thing is, the “original” IBM PC that debuted in 1981 was dubbed the Model 5150, but the 5150 saw subsequent revisions. It was in production through much of the 1980s, tweaked and refined, and sold in parallel with later models, like the next major release, the IBM PC XT. Our lab survivor wasn’t an original, but it shared the same outside and much the same guts.

We didn’t have an IBM keyboard handy, and the DOS floppies were lost to the ages, so how much we could actually do with the PC was pretty limited.  Attaching it to the monitor (more on that in a bit) and firing it up just led to a series of angry, bleating beeps. But we wanted to photograph the system to complement our re-publication of PCMag’s original review of the IBM PC from 1981. You can check that out at the link, and browse the original magazine here on Google Books (not to mention a trove of our back issues). 

So in the course of photography, I did what any self-respecting hardware hound would do: Tear it down! Our patient staff photographer, Molly Flores, stood by snapping. We gave this mid-1980s PC the 2021 PCMag studio treatment. And it seemed a shame to waste all those lovely images, so take a tour inside with us.

Again, I was not a child of the original mid-'80s PC revolution, so my analysis here will be a tad starry-eyed, and will defer in places to former editor in chief Michael Miller's excellent recollections of that time, which he reported first-hand. So set me right and drop your reminiscences in the comments at the article bottom!

This Model 5150 Cleans Up Nice!

First, some glamour pictures of the old Model 5150. Molly made it look nicer than it had any right to look. I scrubbed down the chassis with some mild soap and towels, plus alcohol wipes for the stubborn crud, and canned air for the dust bunnies inside. She scrubbed the rest in Photoshop.

That’s IBM’s Model 5151 monitor on top of the chassis. It’s a resolutely monochrome CRT with a glorious native resolution of….80 characters across by 25 lines. (That’s characters, mind you, not pixels.) IBM's Monochrome Display Adapter (MDA), an early video card, would have powered this panel, though we were to have a surprise in that vein when we opened up the case. (More about that in a moment.)

The two knobs up front are for contrast and brightness…

...and there are no other controls on the monitor, not even an on/off switch. That is because the monitor plugs directly into the back of the PC chassis, over two hardwired cables…

...one for power and one for the video signal. (We’ll get to those cards, ports, and sockets in a moment.) Powering up the PC desktop also powers up the monitor.

So here we have the outer chassis of the Model 5150 proper, cleaner than it has been since 1986 or so…

Those two drives up front are the 5.25-inch floppies, with locking levers that engage with a satisfying ka-chunk. More about them later; we’ll yank them out. (Indeed, I had to put one back in place for these photos; the old museum piece had had it swapped out decades ago.)

And here’s a look around back at the rear panel as a whole. The ISA expansion-card slots have three cards in them, while the unused slots have the covers missing, lost to the ages.

And here’s where the main power cable (right) and the monitor power cable (left) go…

You can also see above the model identification badge, and the old "Made in USA" indication. How many desktops can you say that about nowadays?

Then there’s the hardwired ports for the IBM’s peripherals. The cassette port, apparently, was not widely used, given the inclusion of a floppy drive in most versions of the 5150 sold. (Any mention of PC-connected tape sends shudders down my spine: thinking back to my VIC-20, and that I started my career covering earlier kinds of PC storage, much of it tape drives from the likes of Exabyte, Iomega, and Tandberg in the 1990s.)

The main power switch is on the right panel, near the back. It engages with a nice, meaty thunk-click. You don’t get authoritative switches like that on desktops anymore, that’s for sure. Fighter-jet consoles, maybe. 

As for the rest of the PC's ports, they were on the backplanes of several expansion cards. The one at left with the RCA jacks was clearly a video card; the Model 5151 monitor engaged with the 9-pin connector. What was the one at right, though? 

I’m old enough to remember SCSI cards and parallel ports, and that looked like the latter. There was only one way to find out. (To find out I was wrong, that is.)

Let’s Go Under the Cover

One thing to make clear: Everything about the original IBM PC is heavy. IBM's giant mainframes used to be known by the nickname “Big Iron,” but this desktop PC qualifies for "Little Iron" status (though yes, the chassis is actually steel). It’s dense and heavy.

The cover, especially so. It slides forward to remove; don’t drop it on your foot. I was left with this glorious bare frame…

...which, apart from the enormous, elaborate circuitry of the twin floppy drives, doesn’t look all that foreign versus the guts of a desktop system of today.

Indeed, there’s the power supply, the top label a bit worse for wear, but with a familiar layout and even a main power connector to the motherboard and...Molex connectors! Those are still used today.

I wanted to get down to the motherboard, so some components would have to move. First to pull the expansion cards, held down with screws at the backplane.

Here’s the first one, which was ribbon-cabled to the floppy drive. It was a floppy drive controller card. Back then, lots of basic system connectivity had to be added via expansion card…

You can see the internal ribbon cable still attached; it daisy-chained to the two 5.25-inch floppies. This card was, further research revealed, a later version of the controller card. The external port was meant for external floppy drives; it was not a SCSI or parallel port.

Removing that card better exposed the monaural speaker, meanwhile…

It was on this speaker that the "1984" date tag was tied. 

Next up, we yanked the video card…

This gigantic card is as long as a modern high-end video card, and it has a daughterboard fastened over half of it. The provenance of this card was a bit of a mystery, and required some homework. This was not IBM's basic MDA card, and at first I thought it may have been an aftermarket upgrade. It's IBM issue, though, the IBM Enhanced Graphics Adapter (EGA), and the daughterboard is an expansion of the graphics memory, mounted face to face with the card. (The card had 64K of graphics memory standard.) A precursor to VGA, it would be used for color output at up to 640 by 350. Luxury!

It's definitely an unusual beast, this card, by modern standards. There's a sheet of clear, thin plastic on the back of the daughtercard to keep the extra thickness from touching an adjoining ISA card and presumably shorting out. And the red block at upper right ("Grayhill") is a set of DIP switches for configuring the card's operation according to which monitor you were using it with. (We'd be seeing more of that kind of thing in a moment.)

The last card was easily identifiable as a memory expansion card in ISA card format. But, wow…

That’s a lot of chips and solder points. As best I could tell, it was one of Intel's so-called "Above Board" ISA memory expansion cards. This is how you'd expand your main system memory, well before PC-standard DIMMs and SO-DIMMs were a glimmer in anyone's eye.

And it is quite the international medley of a board. If you look closely, many of the chips (the big spread of NEC memory modules) are made in Japan, but others are made in Malaysia, Mexico, and El Salvador. Others without a clear citizenship may well be American-made, too. It’s less an ISA card than a U.N. summit.

So that got us down to the motherboard. But to get the board out would require taking out one of the floppy drives, which blocked access to one of the screws holding it to the chassis. So, detach the ribbon cable, work off the power connector from the leftmost drive…

Then, bring out the screwdriver, two screws off, and the drive came right out of the chassis like any good 5.25-inch device of yore…

Recommended by Our Editors

The Original IBM PC: Lessons Learned

Project Chess: The Story Behind the Original IBM PC

Here's What Bill Gates Told PCMag About the IBM PC in 1982

Here’s a look around the now almost comically chunky mechanism…

Who doesn’t love pulleys and belts?

And here’s the familiar Molex connector that powers the drive; still used in PCs today for some peripherals…

So, that critical screw exposed, on to the mainboard! A few screws at the corners (regular-head as opposed to Philips-head, that was a surprise) were easy to remove. But the mainboard is also held in place by a handful of mildly evil, plastic squeeze-connection posts that pop through holes in the board. You have to squeeze them a bit with fingers or pliers—gently—and work them back through the board's mounting holes. Problem is, you release a few and they pop back in as you work on the rest. You need six or seven hands to squeeze them all at once!

Down to the Motherboard

A patient 10 minutes or so released them all, though, and the motherboard slid right out one edge of the chassis…

What a glorious sight! And not an unfamiliar one. This doesn’t look a million miles from a modern motherboard, despite some clear differences. For one, this was the whole of the I/O port panel…

The DIN-type connectors for the keyboard and cassette add-on—that’s all there is. 

Now, these early PC motherboards didn’t have a BIOS in the way we know it. You configured your hardware with DIP switches, used in concert with manuals (and probably some desperate phone calls to tech support or your geekiest compadre)…

I recall these from some of the much later PC clones I owned. (I also recall loathing them.) There are two banks of these baby-blue baddies on the 5150 board. 

Also on the board: More RAM. Here it’s soldered down. We’re a ways from the days of ubiquitous DIMMs. Memory densities weren’t nearly where they needed to be to get enough RAM onto sticks that size...not even close. 

And here’s the main motherboard power connector, I detached the power supply's main lead from here before yanking the mainboard. It’s foreign but a clear ancestor to the standard ATX 20+4 pin connector we know today.

The one thing that is conspicuously missing, versus a motherboard of any remotely recent vintage: Where’s the CPU, the CPU socket, and the heat sink? Well, that took some scouting, but we finally found the chip…

Good old Intel 8088, a 4.7MHz CPU on the board. The socket next door, empty, was meant for a math co-processor. You’ll note no heat sink or even an auxiliary fan blowing near by. The 8088 on the board was just another chip. (Michael Miller has lots more on the decision to make the Intel 8088 the core of the Model 5150, if you're interested in the backstory.)

Then there is the array of ISA expansion-card slots...

Finally, here's the real heart of the PC: the bank of 8-bit expansion slots (later to be dubbed 8-bit ISA slots in retroactive fashion). The open architecture that IBM opted for in the original PC spun up a whole ecosystem of expansion cards and more that, in part, gave the IBM PC so much of its momentum. (Michael Miller also discusses, in more detail here, the decision to go to an open architecture.) Later IBM PCs would include more slots, and then IBM would flirt with the proprietary MCA architecture in the latter part of the 1980s. But here you can see the start of the throughline from the original slots to 16-bit ISA and EISA, to PCI and VESA, and to the PCI Express we know today.

A Hunk of Bonus Hardware

Now, this piece wasn't standard issue, but this hard drive was stuffed, unceremoniously and uninstalled, inside the chassis. We cleaned it up a bit for its star turn…

It’s a Seagate ST-225, a 20MB drive that interfaced with the system via a dedicated ISA controller card. Notable is how the hard drive’s front face actually occupies a full 5.25-inch bay. Interestingly, ours came with a dot-matrix-printed test report Scotch-taped to the top. 

Fresh-faced and youthful...in 1988. Weren't we all?

Finally: Would It Boot?

Oh heavens, no. No, this museum piece would not boot. I tried, admittedly half-heartedly, for half a day. I didn’t have a compatible IBM PC keyboard handy, which was a wee hurdle, but I'd have called it a victory getting a flashing cursor on the screen. (Even if it booted to the Model 5150's built-in BASIC, all I'd be able to do is stare at it agog.) Nor did we still have the original DOS diskettes. Indeed, I never even established whether those floppy drives were single-sided or double-sided mechanisms inside. (That would matter, too.) All I got was the angry bleats when I switched it on. An early version of “Keyboard not found”? An obscure hardware fault? Misconfigured DIP switches?

The bright spot, of course, is that it blurped and bleated at all; the power supply still spun up. And at least our museum-model 5150 is cleaner and in better repair than it has been for the whole of the 21st century.

Will we get it up and running? Let’s make a date for the IBM PC’s golden anniversary to get it all ship-shape, shall we? We’ll make every effort to show it off to you in the hyperconnected virtual metaverse of the time. No doubt, though, the story we concoct will be cooked up on the great-great-great-grandchild of the humble Model 5150. August 2031, here we come.

For more, check out PCMag's full coverage of the 40th anniversary of the IBM PC:

• The Golden Age of IBM PCs

• The IBM PC at 40: Lessons Learned

• Why the IBM PC Used an Intel 8088

• Why the IBM PC Had an Open Architecture

• The IBM PC's Killer Apps: Where Are They Now?

• Project Chess: The Story Behind the Original IBM PC

• The Rise of DOS: How Microsoft Got the IBM PC OS Contract

• Here's What Bill Gates Told PCMag About the IBM PC in 1982

• 40 Years Later: Read PCMag's First Impressions of the IBM PC

• Spec Showdown: The Original IBM PC vs. a Lenovo ThinkCentre AIO