November 1990 NEVER RELEASED OFFICIAL TI PERIPHERALS: THE WAFERTAPE DIGITAL TAPE DRIVE a hands on description by Charles Good Lima Ohio User Group The WAFERTAPE DIGITAL TAPE DRIVE was supposed to be a step up from cassette data storage. The device is totally under the control of the computer (no manual rewinding or keeping track of cassette counter numbers). A directory on the tape allows the computer to automatically advance the tape to the beginning of any desired file. Recording data digitally (as 1 or 0, on or off) rather than as part of a continuous spectrum of sound frequencies as is done on a regular sound cassette recorder such as T.I.'s Data Recorder, was supposed to be a more reliable way of recording and retrieving data. Had this device been released in 1983 after its debut in January of that year, it and the also never released HexBus interface would have formed an inexpensive mass storage upgrade (compared to using a cassette recorder) for the 99/4A at a time when a new full PE box (SSSD drive, 32K, disk controller) cost anywhere from $550-$1200. The last 99/4A catalog published by TI in the fall of 1983 lists the Wafertape drive for $139.95 and the HexBus interface for $59.95. The wafertape drive was shown at the Consumer Electronics Show in January 1983 together with the first showing of the CC40. It was to be the major mass storage device for the CC40, and is described and pictured in the user guides that come with most CC40 software and peripherales. Unfortunately these user guides now also come with an addendum sheet that states "The Wafertape Digital Tape Drive is not available." The non release of the Wafertape Drive left the CC40 totally without a mass storage device until the 1986/1987 introduction of the Mechatronic HexBus Quickdisk drive. This lack of a mass storage device probably killed most consumer interest in the CC40. I recently purchased for $100 a working Wafertape Drive, serial number 0000007, ATA3883. It is my understanding that the ATA number is a date code, indicating in this case manufacture in the 38th week of 1983. Most 99/4A hardware and software modules have an ATA number. If my understanding of ATA numbers is correct, my Wafertape Drive was not one of those shown at the January 1983 exhibit. I personally know of one other working Wafertape drive, serial number 0000095. I have been told that there are about 10 or 11 working Wafertape Drives in private hands and others that don't work. The Wafertape Drive is comparable in size to other TI HexBus peripherals, measuring about 11.5cm wide, 14.5cm deep and 3.5cm tall. It is designed to be stacked with the HexBus RS232, modem, and Printer/Plotter. It is battery powered with 4 AA cells. You can also use an AC adapter. The front has an on/off switch and a slot for inserting the wafertape. On the back are two HexBus ports, an AC adapter jack, and a rotary switch for setting the device number. The switch on mine has positions 0-9. However, only positions 0-7 work, corresponding to devices 1-8. By accessing each HexBus peripheral individually by number, a single CC40 could control up to 8 Wafertape Drives, if one could somehow gather that many working drives together in one place. Wafertapes come in a cartridge measuring 68x40x5 mm, about as big as the miniature cassettes sold these days for small tape recorders. The top is clear plastic and the bottom is black plastic. T.I. calls these cartridges "wafers". They fit easily and snugly into the slot in the front of the Wafertape Drive. Inside the wafer a dark colored magnetic tape 1.7mm wide (very thin) is wound in a continuous loop, in the same manner as the tape of an "8 track" music cartridge. The tape moves only in one direction and its ends are attached to each other with a piece of reflective silver tape. T.I.'s last 99/4A catalog lists 50 foot ($7.95), 25 foot ($6.95), and 10 foot ($5.95) wafertapes. Five foot wafertapes are also mentioned in the Wafertape User's Guide. Long tapes store more, but it takes longer to find the beginning of a specific file. Official T.I. wafers have a little sliding panel that covers the exposed part of the tape at the edge of the wafer. This slides open as the wafer is inserted into the slot on the Wafertape Drive. I have one such T.I. wafer and I also have some wafers without a T.I. label and without the sliding panel. I wonder who made these "generic" wafers?^^I do know that other wafertape drives were planned or actually sold for other devices. A wafertape drive for the Tandy 100 laptop computer is described in the March 1984 issue of Creative Computing. Also, see the comments below from Tony McGovern about another wafertape device. Maybe my "generic" wafertapes were not made specifically for the TI Wafertape Drive. The method of write protecting wafers is unusual. When I first got my Wafertape Drive I thought it was defective because I always got a "write protected" error message whenever I tried to initialize a wafer. I could find nothing resembling a write protect tab on my wafers, and looking inside the Wafertape Drive slot revealed no evidence of a mechanical pin associated with the "remove the tab and it is write protected" system of protection typically found on most audio and video tape cartridges. It turns out that you have to put a "write enable" paper sticker at a specific location on the top of the wafer in order to write to the wafer. Anything white will do. This "write enable" piece of paper is OPTICALLY sensed from above by the Wafertape Drive. In the absence of the sticker, the optical sensor sees through the transparent wafer upper surface and does not get a reflection off of the black wafer bottom. A second optical sensor detects the silver end/beginning of tape marker. This marker is the only reference point the wafertape drive has to tell the relative position of everything else on the tape. On "8 track" continuous loop music cartridges, the end/beginning of tape marker is detected electronically. This silver marker on a wafertape is detected through a window in the center of the wafer's transparent upper surface. If the "write enable" sticker you are using is too big and covers this window, then the wafer will not be usable as I finally figured out after lots of frustration. When you initialize a wafer with the FORMAT command of the CC40, the tape is advanced until the marker is optically detected, and then the Wafertape Drive prepares a new tape directory area at the beginning of the tape. The entire tape is not magnetically encoded with FORMAT, just the directory area. Software designed to directly read the contents of a wafertape directory reveals that if a previously used wafer is reFORMATted, the old file names are retained in the directory area, but all file lengths are set to zero. The time required for a FORMAT depends on how close the tape is to its beginning when FORMAT begins. Even a 50 foot tape formats very rapidly if it is already almost at its beginning. A directory has room for 16 files irrespective of tape length. Of course short tapes may not have enough room for 16 files if they are of significant length. Files are written to the tape sequentially, with the directory keeping track of the file name, number (0-15), and length. Apparently the Wafertape Drive locates specific files by file number, counting the End-Of-File indicators that pass by the read/write head as the tape is advanced to the start of the desired file. Only the last file can be overwritten by another file of the same name. If you write a file with the same name as a file already on the wafer that is not the last file on the wafer, the old file's directory reference is deleted and the new file of the same is written to fresh space after all the other files currently on the tape. (I hope you understood that.) File types supported include PROGRAM, and INTERNAL or DISPLAY data files. RELATIVE files are not supported, and you can't open data files as APPEND. Although this is not made clear in available documentation, I think that when data files are read from wafertape, the entire file is read into computer memory for manipulation by the controlling program and then later if necessary written back to wafertape. This limits the size of data files. Only one wafertape file can be opened at a time. To compare the speeds of the Quickdisk and Wafertape drives, I timed the SAVE and OLD of a 15300 byte text file from Memo Processor using a newly FORMATed wafer and disk. Wafertape: SAVE, 4 min 25 sec; OLD; 3 min, 10 sec. Quickdisk: SAVE, 2 min 15 sec; OLD, only 38 seconds. Why wasn't the Wafertape Digital Tape Drive ever officially released? It just did not meet T.I.'s standards for reliability. It does not work well on battery power. Even with four newly installed, fresh alkaline AA batteries, you almost always get an I/O error 25 (low batteries in peripheral) when you try to load something, and you often get the same error when you try to SAVE while on battery power. The Wafertape Drive only works with much reliability when operated with the AC adapter. Apparently the speed at which the tape crosses the Wafertape Drive's read/write head is critical, and variations in this speed are not tolerated. With any battery, continuous power drain results in a voltage decrease compared to the initial voltage put out by the battery. Such a voltage decrease slows down the drive motor. Also, as the Wafertape Drive operates it turns itself on and off several times as it loads or stores data. Starting an electric motor requires an immediate surge of extra current compared to the current needed to keep the motor operating at constant speed once it has started. It is possible that the Wafertape Drive's AA batteries are not able to maintain constant voltage with all the required on/off cycles. In addition to the battery problem, it is sometimes possible to write data beyond the end of the wafertape and wipe out the directory on the other side of the reflective end/beginning of tape marker. This renders all files on the tape useless. I have managed to overwrite the end of a wafertape on two occasions. At other times, when I deliberately tried to do this, the CC40 would not let me write past the end of a wafertape. I sometimes get I/O error 6 (device error, try again) with the Wafertape drive for no reason I can determine. Sometimes trying again doesn't work. I never get these error messages when using my Quickdisk drive. Finally, I suspect that wafertapes are not as durable as disks. The tape is very tiny and is subject to a lot of physical movement and twisting as it moves within the wafer. I suspect that with time the tape may break. I know for example that I have had a higher percentage of my "8 track" music tapes break compared to my reel to reel cassette music tapes. Let me quote from a letter received recently from Tony McGovern of Australia, senior author of FUNNELWEB. "Wafer-tapes were always a disaster area! I think they appeared in one of Sinclair's UK machines. The other place they appeared, also in the UK, was in an abomination produced by ICL sold here, a computer phone -a combination of low end PC with modem/phone all built in- but no disk drive, only the wretched unreliable wafertape. Telecom Australia probably has a warehouse full of these things that they would rather not be reminded of. They had a great marketing campaign to sell these several years ago and no one wanted the turkeys." I do use my Waertape drive. It isn't that unreliable. I have a little briefcase in which I can keep my CC40, the Wafertape Drive, my HexBus printer/plotter, a power strip, and all the necessary AC adapters all pluged in and ready to go. I can open the briefcase and plug in the power strip and use the peripherals as they sit in the briefcase. I can also use the printer/plotter and CC40 while still in the briefcase using battery power, but I have given up trying to use the Wafertape Drive with battery power. I keep my HexBus RS232 and Quickdisk drive pluged in (these two devices both REQUIRE AC power) next to my 99/4A. If possible, anything saved to wafertape ALSO eventually gets save to disk with the Quickdisk drive, which I find to be quite reliable.