Teletype Tales Part 3

_Teletype_Tales_ Part 3 - How I Spent My Christmas Vacation If you're still reading, don't worry, this is the last part... but it is a bit long. This year I ended up with a good chunk of "left over" vacation time. My employer implemented a "use or lose" vacation policy and my boss, a _really_ nice guy, was insistent that none of his staff was going to lose any. So I ended up on vacation for the last two weeks of 2010 despite having a rather heavy feature request list on my plate. What to do? Spend a lot of time with my family of course but after a discussion with my wife it was decided that part of my Christmas present would be several uninterrupted days of hobby time. Time to "bite the bullet" and get that teletype working. Or end up with a big pile of levers, cams, spring, cogs, etc. The day came and I jumped in with both feet. I took a few pictures and made a few notes. Then I pulled all of the cables off of the call control unit, popped out the h-plate and lifted the typing unit out. I guess I half expected springs and gears to go flying everywhere but it came out neat and easy. Even after all of the work I had done before, the filth was incredible. In addition to much more dog hair and bug bits, there was oil soaked chad and this "goop" everywhere. But now I could see the source of the goop. Under the typing unit, there was what had been a large foam rubber pad. It had decayed into crumbly dust that, when exposed to oil, turned into the goop. Clearly, more cleaning was required. Do another halfway job? Nope, take it all the way apart. Four screws later and the call control unit lifted out. Two more after that and the keyboard was free. One last screw and the tape reader was out. Now I could fully clean the subbase. Lots of detergent, warm water, gentle scrubbing and, finally, paper towels and it was squeaky clean. I still didn't have the guts to try and disassemble the typing unit or the keyboard, but with access to their undersides, I was able to clean a lot more gunk out of them with a combination of light instrument oil, canned air, paint brushes and toothpicks. Well, that had to be a good bit better. I put it all back together. Getting the h-plate back in isn't too hard once you get the knack. Plug in, local mode, a definite improvement. The carriage movement wasn't perfect but it was much better. The keypress encode/decode now seemed to match in almost all cases. Two problems remained, no line feed and no bell. I took the typing unit back out and looked carefully at the area of the line feed mechanism. It was substantially different from the illustrations in the documents. At this point, I made what turned out to be a great decision. I brought out the KSR to see if its line feed mechanism looked like the docs. The h-plate in the KSR was already out as a result of its previous misadventures. Pulling the cables to the call control unit was easy and I lifted out the KSR typing unit. Its line feed mechanism seemed to match the docs. As I looked at it, I realized that it didn't seem to be damaged in any way. I hand turned the motor a few revolutions and everything seemed OK. I decided I would give it a try, installed in the ASR. I spent a little while, giving it the same cleaning I gave the ASR typing unit. One interesting "tidbit" about this typing unit is that, sometime in its past, it clearly "ate" a ribbon. There were bits of shredded ribbon all through it. My hands were black. To make a long story short, the KSR typing unit worked perfectly in the ASR, bell, line feed and all. During the test, I didn't reconnect the reader to the call control unit because the reader control mechanism is part of the ASR typing unit and I was concerned that the reader might "run wild". So now I'm looking at the docs, wondering how hard it would be to transfer the punch and reader control over to the KSR typing unit. The punch seemed like it would be almost trivial. It was, _almost_. The punch essentially just bolts onto the side of the typing unit. There are eight small levers that control which holes are punched and there is a large lever that provides the "power". The small levers simply drop into place, connecting with push rods in the typing unit. The large lever connects to a rotating shaft in the typing unit. Here's where the trick is. The lever connects to the shaft via a sleeve. The shaft has holes all the way through it, the sleeve has holes on both sides and screws go all the way through the sleeve and the shaft. There is almost no play in the connection... almost. I took out the KSR typing unit and transferred the punch mechanism to it. I put the KSR typing unit back in, threaded in some tape and tried it. Shredded tape. What??? Everything sure looked OK. The "bit" levers looked like they were all moving correctly, the large lever seemed to be going through its motions. I spent a long time watching it. Finally, I tried putting it back on the ASR typing unit. Shredded tape. WHAT!!! What could possibly have changed? Maybe I attached the sleeve for the large lever to the wrong holes? No, those are the only holes in the shaft. That's when I noticed that little bit of play. Just three or four degrees. That couldn't possibly make any difference, could it? I held the play all the way clockwise and tightened the sleeve screws. Shredded tape. I loosened up the sleeve, held the play all the way counter-clockwise and retightened. Perfect! I moved the punch back to the KSR typing unit and, with my new knowledge, got it attached and working. After the punch experience, I spent a very long taking exact measurements of the reader control mechanism that needed to be moved to the KSR typing unit. It is a complicated little assembly, consisting of a cam controlled lever that moves based on whether a solenoid is energized and in turn opens and closes a switch as the cam rotates. It looked like the location tolerances of the whole thing would be about a sixteenth of an inch. There is a spring that attaches to the lever and keeps it pressed tight against the cam. I transferred the mechanism. Attaching the spring turned out to be the hardest part of the whole job. The lever end of the spring is "permanently" attached, but the other end has to thread through a tight space and hook over a small pin that is part of the typing unit frame. I fooled with it for more than an hour. What finally worked was to tie some plastic fishing line to the end of the spring, guide the line to the pin and use the line to stretch the spring over to the pin. I used a piece of "coat hanger" wire to coax the loop at the end of the spring over the pin and then finally reached in with a long thin scalpel to cut the line and pull it out. "There's the easy part done," I said to myself. I figured I would spend the rest of the day getting the lever, solenoid and switch lined up correctly. I put the KSR... well, now it was really the ASR, typing unit back in and fired it up. The bit gods, or maybe Rube Goldberg himself, smiled upon me. It worked the first time. Now to talk to the world. I don't have anything that is currently configured to "talk" current loop but I have a current loop to RS-232 converter. I'll spare you the details of frustratingly confusing documents that aren't clear about which wire pairs the 33 is sending on and which it is receiving on, but I will mention that the current loop to RS-232 converter is pretty old and populated with tantalum capacitors. Guess what caught fire? Fortunately, B&B publishes the schematics for the converter on the web. They also publish a nice little document that shows how to wire up a 20ma current source for the loop. In case you didn't look before, the pictures are here: http://wsudbrink.dyndns.org:8080/images/teletype/ P1010006 and P1010007 show the converter and the current source, built in an old RS-232 patch box. The pin sockets had been previously damaged, so I removed them and soldered the circuit directly to the PC board in the patch box. Oh, the external power connector on the RS-232 converter melted when the caps burned, so the patch box also supplies power to the converter. But that was it for the problems. The 33 now talks to itself perfectly in local mode and talks to the world over the RS-232 converter. I've hooked it up to one of my IMSAIs and to my Altair 680. So, that's about it, the end of an almost ten year "saga". What's left? Two things, one big but minor and one small but rather major. The big minor thing is the back panel to the stand. It's missing, it's big, but you can't see it unless you pull the unit away from the wall and walk around back. Its absence doesn't seem to hurt anything and I'm not even sure the unit originally had one as there are no scuffs or scratches where you would expect if one had been installed. Now, the small major thing is the small screw (Teletype part number TP183112) that holds the paper tape reader cover in place. It is physically small, but it is major because the cover bounces around during normal operation without it. It seems to be a rather special screw, more like a threaded rod. If I can't get an original, I will probably try to fabricate one myself. Oh, the last thing is the KSR unit. It is partially operating, no bell or line feed and under the wrong cover. I'm going to hang on to it for spares if anything in the ASR breaks. I guess I really lucked out in my recent restoration. I didn't have to dig deeply into the keyboard or the typing units... knock on wood, I can't imagine what horrors of finely tuned springs and levers lurk in them :-). I hope you have enjoyed reading this, Bill Sudbrink, January 2011