Jul 86 History
|Column Tag:||Historical Computing
The Famous TV Typewriter
By Dick Heiser, Industry Pioneer, On the Great Peace March
A Cooperative Venture
Sometimes the teamwork that makes a successful product is hidden, sometimes it's visible. In the case of the Macintosh, it's both hidden and visible- the names are molded into the inside of the case! The tv typewriter (TVT) was the work of four independent people. Don Lancaster, Dan Myers, Josef Rosenthal and I worked in succession to design it, to produce it as a kit, to modify it into a proper computer terminal, and to document the information. Each of us brought the product closer to the customer by making it easier to build and more useful.
Mother of Invention
The Altair personal computer didn't include a built-in video terminal. Video terminals sold for $1500 or more, until Lear Siegler introduced their $995 "dumb" terminal. But even that Lear Siegler terminal was an expensive accessory for a personal computer. Used Teletype model 33 teletypewriters were expensive too; good ones sold for $800 or more. Teletypes are like Volkswagens- they're slow and noisy, but they hardly depreciate, and they last forever.
Everybody who built an Altair needed a low-cost terminal for it. The "TV Typewriter 2" from Southwest Technical Products Company emerged as a probable best buy. [I had one. It was great! -Ed.] The TVT-2 kit cost about $275 by mail. Adding a keyboard, power supply, and serial interface raised the price by another $100.
Don Lancaster - Designer
"Don Lancaster writes books". That's how the designer of the TVT describes himself. He has written about a dozen electronics books. Many are still in print long after lesser technical books are gone. He's also a volunteer fireman and occasional forest ranger.
His Incredible Secret Money Machine is a wonderful career guidance book. Basically he suggests:
Work for yourself.
Do something you can throw yourself into wholeheartedly.
Maximize the value you personally add to your product.
Look for leverage and royalties.
Don't invest lots of money in a new business.
When I referred to this book at the 1984 Hackers Conference, the audiance interrupted to cheer. Lots of top hackers know this book! Lancaster illustrates his advice by talking about himself. As an example of leverage, he explains that when he designs an electronics project, he writes a magazine article about it, licenses the kit for a royalty, and uses the articles as a book chapter. Don's TTL Cookbook and TV Typewriter Cookbook were best-sellers in my store.
Don Myers warned me "Don doesn't like to chat. In fact, he doesn't always answer his phone. He's as likely to sleep days and work nights as the other way around." When I called Don's unlisted number, he turned out to be quite friendly and open. He even told me what he planned next for the TVT's number 3, 4 and 5! He planned to lower the cost, and add color, to develop a terminal more like an Apple II. I tried to talk him into more characters per line and lower-case letters, because of my interest in computer text editing. Don was on a better track than I was.
The TV Typewriter-2 displayed 16 lines of 32 upper-case characters in a 5x7 dot matrix. The main circuit board was about 11 inches square, with a daughter board for screen memory, and optional boards for serial or parallel interface and manual or coded cursor movement control. It had connectors for a parallel ASCII keyboard, power supply, and video monitor.
The TV Typewriter-2 was an improved design. Lancaster used considerable expertise to generate an accurate, standard video signal. He was proud of the rock-steady video. The TVT was not really a computer terminal, however. For instance, new lines of text overwrote old text, so the ends of longer old lines showed alongside the new lines. The TVT was a video generator, suited for titling and announcements via closed-circuit tv, like the ones in hotels, theater lobbies, and airline terminals.
Dan Myers - Kit Producer
Dan Myers presided over the Southwest Technical Products Company in San Antonio, Texas. His company paid a royalty to Don Lancaster for the right to sell TVT circuit boards and kits. He advertised in magazines like Popular Electronics, because the TVT was an electronics project rather than a computer peripheral.
Dan took the printed circuit layout from Don Lancaster and had boards made. He prepared bags of integrated circuits, resistors, capacitors, and connectors. He printed a brief manual with layout diagram and schematic. He also bought the rights to a parallel ASCII keyboard design.
I've never seen Dan in person, but we've learned a lot about each other by talking on the phone and by doing business together. Dan is a "no frills" businessman, but he won't cut corners that he considers important. He wants to hold down costs so he can sell low cost products. Don Lancaster warned me to expect very short discounts, maybe ten percent or less, from Dan. I felt very lucky to negotiate a twenty percent discount, and preferred to do all of my business directly with Dan. Sometimes I wondered if he was a one-man show back then.
Unlike MITS and Heathkit, Dan charged extra to fix your machine if it didn't work. Dan and Don were well-matched to collaborate: they both wanted to control costs first and add features second.
The following year, Dan bought rights to a Motorola 6800 based computer design. He evaluated it on the basis of memory size, clock speed, instruction set, and price. His 6800 computer was a fantastic hardware value. He advertised it in the major computer and electronics magazines, and signed up dealers. It was a brave design. Unlike the MITS and IMSAI computers, it had no rows of switches and flashing lights. It used a "software front panel" which was the Right Stuff; better and cheaper.
His one big mistake was to apply his "no frills" attitude to software. Too bad. As Portia Isaacson once said, "After you consider the software available for a particular computer, the next most important consideration is the color of the cabinet!" Software was a hard lesson for Dan.
The TVT Kit
When you opened the five cardboard boxes for your TVT, serial interface, keyboard, power supply, and cursor control board, your emotions reflected your electronics expertise. The printed circuit board was a first-class piece of goods; fiberglass, two-sided with plated-through holes. You had to identify the parts by the codes printed on them. The only decoding information was a layout diagram for chip replacement, and a color-code list for resistors. Do you like Molex connectors? One early customer put the connectors on backwards, but they're cheap and they work.
The keyboard circuit was plated on only one side, so you had to solder on perpendicular "bus bars" to stiffen it and to supply the horizontal traces. Too little solder and the bus bars came loose. Too much heat and the plastic keyswitches would melt. It was definitely a challenge for newcomers!
The keyboards weren't good enough at first. More expensive keyswitches, more special characters, and improved circuit board layout helped, but these low cost keyboards were susceptible to static electricity burnouts.
Left as an exercise for the kit builder was the problem of boxing this collection of boards, figuring out which polarity of the keyboard strobe the TVT expected, and finding a video monitor. [Hitachi made a great little 9 inch montor, the 905, that was very popular. -Ed.]
Josef Rosenthal - System Engineer
Joe worked as a programmer for the System Development corporation (SDC) in Santa Monica. He built an Altair and a TVT early in the product life cylce. That means I sold him the stuff before it worked together properly.
Smiling Joe looked at the misfits between the computer and the TVT as a puzzle. He came by the store often to find out the latest developments and to share his discoveries.
Three problems fascinated him. First, the TVT was slow: its serial interface operated at only 10 characters per second. It was probably intended to connect TVT's to the teletypewriters rather than to computers. Joe wanted the interface to work ten times as fast.
The second problem has been mentioned above: ends of old lines of text remained on the screen beside the new lines. Listing a Basic program resulted in a gibberish mixture of text. Joe worked on a way to erase each line or screen before starting to write on it, or to erase to the end of each line before responding to a carriage return character.
The third problem involved taking advantage of a design feature. Don Lancaster provided two screens of video memory, and the TVT filled them alternately. We wanted to be able to switch screens, to review previous data.
Joe was the only one of us who worked for free. He was an enthusiast: excited, happy, full of possibilities, curious, experimental. He solved all three problems.
Speeding up the interface turned out to be easy. Joe examined the schematic, discovered a divide-by-nine counter in the clock circuit, and jumpered it out. The 110 baud circuit (10 characters per second) would now operate at 990 baud, an unlikely speed. No problem: Joe calculated what combination of jumpers would make the counter in the Altair's serial interface work at 990 baud, and viola, the nine-times speedup worked!
Another magic jumper solved the jumbled lines problem. Joe ran a jumper from the linefeed code to a circuit point which caused erase-to-end-of-line. Sure enough, it worked! Each time the terminal advanced to a new line, the new line was cleared. Now the TVT worked as a proper video terminal.
Joe further experimented with the screen switch logic. He could have added integrated circuit chips to switch from one screen to the other and vice versa. Instead, he directly wired a keyswitch to a screen selector toggle input. Each time you pressed that key, you delivered a random jolt to the selector toggle. No debouncing! I thought it would feel frustrating to operate a random switch, but it worked fine! If the screen didn't switch, I'd hit the button again. No sweat.
Joe also assigned key codes to make the cursor controls work right, and figured out keyboard jumpers to accomplish this. A few more jumpers supplied special characters like the caret or up-arrow, used in Basic programs. Joe accomplished magic: something for nothing. Not adding parts is a characteristic of elegance and of good hacking.
Dick Heiser - System Integrator
I was the last member of this "gang of four." As the person responsible to the customer for a successful project, I wanted to tie together what Don, Dan, Joe and others had discovered. At IBM, I'd learned that, "you're not through until the paperwork is done."
Earlier, I'd written a useful booklet of tips for building the Altair computer kit. The TVT booklet suggested which kits to obtain, how to orient connectors, gave advice on making the keyboard strong and reliable, and illustrated cuts and jumpers for strobe polarity, clear newline, screen select, special characters, cursor control and serial speed-up. It showed how to jumper the MITS serial board for 990 baud.
I had to revise this booklet as we learned more, and when improved keyboards became available.
In addition to the booklet, we sold a kit of extra parts, including chip sockets, keyboard cable, fuse, switch, screws and grommets. We also tried to sell cases for the TVT. Unhappily, metal-bending is a low-tech art, and we could not find a supplier able to consistently deliver quality. We helped people obtain a video monitor, too. We reprinted an article on converting tv's and we sold converted tv's and closed-circuit monitors.
The manual overcame stumbling-blocks associated with the kit, and showed how to make a good $400 video terminal. The design and the chips were good enough that, armed with the advice in our notes, most customers were successful in building a TVT that worked right. Customers found the improved TVT easy to build. Our confidence at the store increased because our recommendations worked so well.
Using the TVT
For several months before we bought a cash register, I balanced the receipts with the comptuer. I used the TVT to operate the Altair, with a simple Basic program.
The video terminal gives a computer its "feel". It's "where the rubber meets the road." The TVT had a good feel. It's big letters (32 characters per line) were highly visible. Sometimes I like lots of little characters on the screen for context, but usually big characters are visually more relaxing. Naturally, the Macintosh is superior here, allowing either size. See if you like writing your draft copies with a larger than normal font next time.
The TVT's full-stroke keyboard allowed confident typing. Computer wizard Rick Shiffman would bang the keys so hard that weak keyboards would simply break. Thus Heiser's Fifth Law: keyboard bashing reflects user confidence- or poor software design!
Our modified TVT was the "state of the art" for only a year. Then, memory-mapped video boards for the Altair came along. These character-oriented displays worked like the Apple II. A dedicated area of computer memory held ascii codes to drive a character generator.
Polymorphic Systems and Processor Technology produced good memory-mapped video boards. This was before the era of the CP/M operating system, so softwre device drivers were not used. Each program had to be individually modified for memory-mapped video. Soon, memory-maped video took over, offering speed, compactness and economy, better suited to the powers of personal computers.
In it's day, the TVT was a key to making the Altair work right. I'm proud of the result, happy that our collaboration worked out so well. Besides which, it was fun!