When we look at our computers our minds are usually only on the finished product. Sometimes it’s interesting to look more closely, and to also check some of the TIMEX and Sinclair News Releases, made before actually selling a particular item. Doing so we see that our computers might have looked and acted quite differently from the way they do today.
This shouldn’t be too surprising. When a group of people set out to design a new product, their initial impressions are often very different from the final result. This is because the design process is itself a learning experience. Although different people in the design group may want to add different features, some unforseen gadgetry almost always has to be added to make it work together. There are times when some options absolutely won’t cooperate, in which case something has to go. In the middle of the design, it may turn out that some unplanned feature would be convenient, causing further changes. And of course, five minutes after the first hand built unit is working, in the middle of the celebration, someone from marketing will call to tell them that it has to be redesigned to include some new function, or it won’t sell. Then once that’s done, the redesigned unit won’t fit into the case….
It goes on and on (and you wondered why it sometimes takes so long for an announced product to reach the marketplace.)
It doesn’t even end once finished units start rolling off the production line. Once in the hands of actual customers, numerous never-before-seen bugs will come screaming out of the machine matched only by the screaming of the users. Design changes for later issues of the product may both correct the bugs and make it cheaper to manufacture. The Sinclair Interface One, which among other things controls the Spectrum microdrives, is already in its second issue! The ZX-81, TS1000, the RAM pack, and the British ZX-Spectrum all have at least three issues! And if you’re following the story of Sinclair’s QL computer, you’d know that you NEED a (working) QL just to keep track of the changes flying about. The TS2068 wasn’t around long enough to really get beaten around, but the long list of software bugs detailed in the TS2068 Technical Manual suggests that a lot of the system code would have gotten fixed up in later versions. The printed circuit board would likely have been redesigned as well. There’s a “kluge board” plugged into an IC socket, with extra wires strung around inside; you see the board quite clearly if you look into the expansion slot. The life of the TS2068 may have been short, but not so short to escape last minute engineering changes!
There are some interesting stories surrounding the ROMs for Sinclair’s ZX-81. Many SINCUS “old timers” probably remember the early ZX-81’s that couldn’t count. (If you have an old ZX-81, take the square root of .25; you should get .5, but you may instead get something over 3, in which case you have an old ROM.) Very few people are aware of an earlier version, never released for sale, that contained PRINT TO, DRAW, UNDRAW, READ, DATA, and RESTORE commands in its BASIC. ROMs with this code were made up, including one sent for review to SYNTAX (see SYNTAX, Dec. 1980).
It might seem strange that these commands were removed from the production version; early users complained bitterly that READ, DATA, and RESTORE were missing. If we look in the same issue of SYNTAX, at the list of commands available in the ROM, we find that five other commands are conspicuously missing FAST SLOW, LLIST, LPRINT, and COPY. The early ZX-81 designs didn’t have a SLOW mode or support a printer! Perhaps some market forecast predicted better sales if a printer were also available. Then again, someone may have just figured that after they sold a lot of computers, there’d be an automatic market for a lot of printers. The added appeal of the SLOW mode is obvious to those who have used it. In any case, the present ZX-81/TS 1000 ROM is absolutely full; I know of only one (!!!) unused byte in its entire 8K! In order to get the new commands in, something that wasn’t absolutely necessary had to be deleted Those who never bought printers might resent the loss of some very useful commands, but that’s the decision that was made.
Speaking of ROMs, an interesting mystery surrounds the ROM socket in the ZX-81. Altho the ZX-81 ROM is a 24 pin device, it was plugged into a 28 pin socket, meaning that something larger than the standard ROM can be inserted. For a long time there were all kinds of rumors. The most popular said that Sinclair would soon introduce a mini-disk drive and a 16K ROM to support it, which would just plug into the “oversized” 8K ROM socket. The answer was in the printed circuit board, but I don’t know anyone else who checked it out. The additional signal needed to control a 16K ROM is not connected to the socket; you can’t just plug in a 16K ROM. Also, the 8K ROM pinout is a bit different from the standard 16K pinout; other rewiring would be needed. There are certain EPROMs that can plug into the socket, but only part of their memory would be useable with it wired as it is. This suggests that an EPROM with a “test program” might have been plugged in at the factory, to check out the circuitry.
But it wouldn’t be any fun to leave the story there. It’s interesting to note that Sinclair’s original Issue I printed circuit board has a few circuit traces that just stop without being connected to anything!!! They stop right next to the ROM!! Tracing around these lines, I’ve found that it’s possible to cut a few printed circuit lines, add a few (very, very short) wires, and reconfigurethe board for a 16K ROM or EPROM, or an 8K EPROM. (The exact cuts and additions depend on what you want to plug in.) The operation is somewhat more complicated than many ZX-81 owners would feel comfortable doing, but it’s possible that Sinclair had planned to modify the machines, for a fee.
It seems highly likely that someone at Sinclair wanted to keep the option to someday manufacture a 16K ROM version of the ZX-81. The printed circuitry couldn’t be laid out in a more convenient way. Just the tiniest bit of extra printed circuitry would make the necessary connections, and no existing lines would have had to be moved. Making the necessary line deletions is a piece of cake. But whatever their plans it wasn’t to be.
In case anyone’s interested, the TS1000 uses the Issue 3 printed circuit board, which is laid out very differently. It still uses a 28 pin scheme in the ROM area, so it may also be possible to do “quick changes” on it. (I’ve not checked it myself; in between the time I checked out the Issue 1 circuit board and the time my vision returned to normal, I vowed to never, never, never do that again.) Note that this is all since I’m told TIMEX soldered the ICs directly to the printed circuit board after that. It would be pretty much impossible to change memory chips under those circumstances.
The ROM rumor also mentioned a Sinclair mini-disk drive. This could have had some basis in fact, since Sinclair did develop a small disk drive before coming out with their continuous-tape-loop microdrives. Every other disk in the universe contains numerous concentric tracks, requiring expensive and fairly precise head positioning gadgetry in the disk drive. It also allows the head to be positioned quickly on the proper track. The Sinclair disk contained a single track, that spiralled in the center like the groove in a phonograph record. This probably allowed the head positioning mechanism to consist mostly of an injection molded plastic template, impressed with the shape of the spiral track. This would lengthen the time required to “find” the correct data on the disk, but it would also cut the cost of the drive. However, the continuous tape loop in the present microdrives should not be any slower than the Sinclair disk scheme, and doesn’t require the head to be moved at all. Probably for this reason, the disk was announced as a technology development, was given some small mention in the electronics trade publicationsn and it silently vanished. This seems all for the best, despite the many problems that present microdrive users are having. Nevertheless, had it been developed soon enough, even American users might today be plugging strange little disks into strange black plastic disk drives.
Here’s an interesting thought. If you wanted to build a TS1000 add-on that plugged into the expansion slot, but was very tall and wide, you’d have to find a good way to secure it to the computer, or you’d crash anytime you touched it. If you’ll look under your TS1000, you’ll find a strange rectangular indentation in the center rear. On each side is a hole, reinforced with plastic. That indentation will readily accept a mounting bracket from a large, bulky add-on, and it can be attached with two self-tapping screws in the holes.I’ve used this for one of my larger projects, and it works beautifully. It can also be found on ZX-81s and so is a Sinclair design. I’m grateful to the folks at Sinclair for inluding it, but what was in their minds when they put it there?
The “things that might have been” in the TS2068 are a whole different story (and the subject of another article). TIMEX announced some of its plans; we’ve learned a lot by working with what they left in the machine, and details on some comparable Sinclair products are readily available. Since Gary isn’t fond of extra long articles, we’ll leave all that for the next time.