The Tandy 1000 vs PCjr is a natural comparison. The Tandy 1000 is exactly the computer the IBM PCjr should have been. That means there are similarities, but it also means significant differences. So here’s what they were. It goes a long way toward explaining why one was a notorious flop and the other was a raging success.
Any discussion about the Tandy 1000 vs PCjr has to start with the keyboard. The PCjr keyboard isn’t necessarily the worst keyboard of all time, but at IBM’s price, people expected something less like the keyboard on a $49 Timex-Sinclair computer. At the time, IBM was the market leader in typewriters. They knew how to make a good keyboard. And the PCjr came off as the making an intentionally bad keyboard.
Even the improved second version which was a full travel keyboard was not up to IBM standards. It was in the same league as a modern $15 keyboard from your local big box store. Depending on the keyboard, maybe a little better, or it might be worse. Not what people expected from IBM. And the keyboard layout was a bit odd. This led to some incompatibility.
The Tandy 1000 keyboard isn’t perfect either, but was much improved. It used a non-standard connector, so even though third-party keyboards would fit, they didn’t work. Today, adapters exist so you can use a common PS/2 keyboard with your Tandy. This was a common theme. The Tandy 1000 wasn’t perfect either. But it was good enough to gain acceptance where IBM didn’t.
The Tandy 1000 keyboard layout is a bit odd if you are used to modern keyboard layouts. I like the modern PS/2 layout better, but that didn’t exist to copy at the time. It is better than any of the other standards that existed at the time just not as good as the PS/2 standard was.
But the other thing about it was the quality. Once you get past the slightly odd layout, it’s a high quality keyboard. It is a long the lines of a Cherry MX red mechanical keyboard. If you like clicky keyboards, it isn’t for you, but I find it just as nice to type on as an Apple IIe keyboard.
The PCjr was infamously not 100% compatible with the IBM PC. If your software stuck to using the ROM BIOS routine for everything, it would probably work on both machines. But graphics required programming the hardware directly, and that was a problem. The graphics weren’t 100% compatible.
The PCjr had the same problem as the very early IBM clones that ran some software but not others. That was understandable from an early clone. Not so much from IBM itself.
Tandy went out of its way to not call the Tandy 1000 PCjr compatible. It was, but when they were designing the machine, they said when they had to choose between PC or PCjr compatibility, they opted for PC compatibility.
So that meant the Tandy 1000 had graphics and sound that was highly compatible with the PCjr, but also with the PC. Theoretically, the Tandy 1000 was not 100% compatible with either, but it was close enough that most software worked. If you bought your software at Radio Shack, of course it worked. And the Tandy 1000 became so popular that software developers couldn’t ignore it. The machine made its own market. Within a couple of years, software went from saying “IBM” or “IBM and 100% compatibles” on the box to saying “IBM/Tandy and 100% compatibles.” No other specific brand got second billing with IBM like that in the 1980s.
The PCjr ran at 4.77 MHz, but it shared memory between the CPU and video, which introduced overhead. Effectively it was 50-67% the speed of a real IBM PC. A PCjr with some memory expansion and a trick to keep the CPU out of the shared memory helps this, but suffice it to say this trick was not something everyone knew to do in 1984.
The Tandy 1000 didn’t have the shared memory overhead. So it outperformed the PCjr, but its lack of DMA meant it still lagged somewhat behind a real PC. The difference was along the lines of 10 percent, depending on the benchmark. Some benchmarks can’t tell a difference. But once you added memory expansion, which included a DMA controller, the Tandy 1000 closed the gap.
Second-generation models ran at 7.16 MHz, making them nicer to use than either a PCjr or a PC. And starting with the Tandy 1000TX, Tandy started using 286 CPUs in their higher-end model. They still had an 8-bit bus, so they weren’t AT-class machines, but they outrun 8088- and 8086-based machines.
In both cases, you can get a bit of a speed boost by swapping in an NEC V20 for the Intel 8088.
So in the IBM PCjr vs Tandy 1000 speed battle, Tandy wins. As long as you keep the fight fair. Neither was ever considered a high-performance machine, but the Tandy was more pleasant.
The PCjr had limited expandability. IBM integrated the most important ports onto the motherboard, which meant it didn’t need five expansion slots like the IBM PC had, and their solution for the expansion it did need was to use plug-in sidecars that you could change off the side of the machine. It worked, but it wasn’t as elegant as the design they used for their business computers. To add a second disk drive or a hard drive, you had to buy a third-party peripheral that stacked onto the top.
The original Tandy 1000 had ISA expansion slots. The computer was not as big as the original IBM PC, so some cards did not physically fit, but smaller form factor PCs quickly became popular, so that problem soon solved itself. Manufacturers figured out how to shrink their boards to accommodate the smaller form factor.
Within a couple of years, Tandy decided to split the line, offering desktop computers with ISA expansion slots like the original Tandy 1000, but the two all in one machines, the EX and HX, worked a little differently. Instead of slots, they had a bus connector that could take up two three cards. Electrically it was compatible with ISA, but it used a form factor similar to a parallel IDE hard drive connector.
It was a smart compromise. You could add expansion cards to those machines to add memory or a modem, but the expansion was limited enough that people wouldn’t buy the $600 model and try to use it as a serious office computer. For that, they were probably going to buy something that had expansion slots. Maybe that was a desktop model, or maybe it was one of Tandy’s business oriented computers.
This strategy allowed Tandy to sell to both markets. The all-in-ones were very good home computers for their time. They were mediocre business computers at best. IBM’s goal was to not compete with itself. Tandy had the same goal and executed better.
IBM sold somewhere between 270,000 and 500,000 PCjrs. It was only on the market for about 2 years. While 1983 and 1984 were very different from today, Commodore sold 2 million units both of those years. IBM had the capacity to build that many machines, and with good execution, they could have sold that many. Instead, it was a flop for the ages.
Tandy never disclosed sales figures. But at the end of 1984, they said the Tandy 1000 was selling so well they couldn’t make them as fast as they could sell them. And we do know the Tandy 1000 was the best selling home computer from about 1986 to 1989, outselling even the Commodore 64. During those years, the C-64 was still selling 900,000 to a million units a year, so that means Tandy was selling at least a million Tandy 1000s per year. Based on their revenue figures, they may have approached 2 million in their best years. I would estimate the Tandy 1000 may have sold 10 times as many units as the PCjr.
So there’s little reason to wonder what might have been. If the PCjr had been just a little more Tandy 1000-like, IBM could have had some of that market. So that’s why Tandy won the IBM PCjr vs Tandy 1000 battle. The PCjr is mostly remembered as the butt of jokes. But almost any time someone hears I collect vintage computers, they want to talk to me about their Tandy 1000 and how they wish they’d kept it.
By any reasonable measure, the Tandy 1000 was a raging success. But Tandy wasn’t able to follow up on it, and they actually ended up leaving the computer manufacturing business in the summer of 1993, almost 12 years before IBM left the PC business.