So about a year ago I bought an IBM Electric Typewriter Model 85 (which I refer to as an F ET, for Electric Typewriter) from ebay and it unfortunately got damaged in shipping. The platen was bent slightly (luckily didn't effect operation) the barrel for the tab key had a large chip missing and the flipper wasn't seated correctly anymore, the upper case had some decent cracks in it, and the bank of buttons to the left of the code key had it's mounting broken off on both sides. Everything still operated, but cosmetically it didn't look so hot.
Spoiler:
How does this even happen?
The button module appears to have undergone no changes from the ones on the model 75, the PCB is clearly an older design and the older carbon resistors clearly indicate that.
So I started with the really only thing I could do, which was attempt to glue the tab barrel back together as luckily the missing piece was in the bottom of the typewriter. That sort of worked, but wasn't especially smooth as the alignment wasn't perfect. I suspected that would be the case and was surprised that I got it to work at all if we are being honest. After that I needed to remove the barrel frame, re-seat the tab flipper, and bolt mod the whole thing. However, the bolts I have for bolt modding are not the correct length for the thick PCB that IBM uses as these bolts are sized for membranes. After doing the bolt mod I took another, larger bit and carefully countersunk all the holes to try and gain enough clearance to get the nut on the bolt. This didn't really work so I tried one more trick. I used my soldering iron to press the head of the bolts into the plastic ever so slightly. This worked very well and with that done I was able to reassemble the thing and test it out. Everything worked first try and I was very happy with my progress. Cosmetic improvements could still be made but everything was perfectly functional.
Spoiler:
There is a rubber mat that is normally glued to the back of the plate, but the glue is well past expired at this point. What I should have done (and what I'll do in the future, is get very long bolts and larger washers and use the washers to hold the rubber mat onto the back plate. I would toss the mat, but it has the paper label on it and I would like to keep that with the module.
Fast forward to a few months ago. I had removed the keyboard module hoping to make a controller for it with pandrew's assistance and I picked it up to test it out and part of the barrel plate fell off! this was weird since I hadn't been using it at all since I bolt modded it and it was just sitting there on the shelf.
Looking at the pictures it's clear that the holes from the bolt definitely weakened the right side, the left side of the barrel plate doesn't appear to have suffered any ill effects from what I can tell. My best guess is that the stress of taking the on/off toggle off of the barrel plate a few times was not good for it. There aren't a lot of bolts on that side of the barrel so the few that are there are the would be taking most of the force themselves. I would caution against removing that toggle more than you have to. I think something that might help is an end cap to the assembly that prevents the plastic frame from flexing too much when the toggle is removed, possibly a metal clip or something of that nature. That way the force of the toggle's removal isn't being absorbed by a plastic plate with holes in it. At some point I will model and 3d print something for each side of the module and hopefully that will work. Another possible fix is to make some adjustments to the toggle itself that makes removal easier so that less force is at play to begin with.
And that brings us to today. I got an ET 95 that had mechanical issues but was in good cosmetic shape. It arrived entirely undamaged and I was able to swap the memory board, platen, keyboard, and case onto the ET 85. I kept the lid from the 85 as the 95's badge was in slightly worse shape and the 85 has the cool service/sale sticker which I try to keep if I can. The result is a cosmetically and functionally sound typewriter that even though it started out as an 85 it is basically an 95 at this point. I've been calling it an 85+10 personally and I'll get a small "+10" label to slick next to the 85 on the badge here shortly.
All in all the thing is my best typewriter. perhaps some of my Wheelwriters or my Actionwriter have more features, my blue Selectric II looks nicer, and my brother portable is way lighter, but this one has much more presence for sure, and with the cap BS keyboard it is for sure the most pleasant to type on. None of the ink or correction ribbons I have work very well so I need to track down some of that.
Here's some pics of various parts of the beast:
Spoiler:
Interesting that the code key has a T mount when it is mounted on a bar
weirdly enough the mount is perfect for Cherry gold crosspoint. Makes me wonder if IBM intended to use cherry switches at some point? The model 50, 60, and 75 were precursors to the model 65, 85, and 95, and I can't find any hard evidence on what those use so it is possible (albeit unlikely) that they use a cherry switch. I suspect the code key was designed before the keyboard mechanism was settled on for the 50/60/75 and cherry may have been discussed before being dropped for reed switches of some sort. No pics of those switches have yet been found to my knowledge.
but as you can see IBM went with a discrete reed switch here. The manual for the model 50/60/75 show a dial next to a thinner code key for various features, as well as a dial next to a thinner no/off toggle but I am not aware of any such thing for the 65/85/95. I don't recall seeing anything about that in the service manual at least.
they actually bothered to doubleshot the power switch, which is a bit excessive, but for how much these were back in the day I guess it you get what you pay for.
and even more doubleshot caps, these don't even have a mount, they just sit above the switches and consequently they wobble like hell. They're more buttons than keys so that isn't a huge issue I suppose.
The 2 machines in a state of partial part transfer.
And while I had everything apart I removed all the gross case foam. These machines are a delight to work on, 2 latches on either side release the top and bottom cases, remove the platen and the top case comes right off, then the machine can be lifted out of the bottom case without needing to unscrew anything. They were designed for easy serviceability and it shows. Very pleasant to work on. What was not pleasant was the foam which was held to the top case with double sided tape and in the bottom case was affixed to a foil liner. The foil could simply be lifted out and thrown away but the rest of the foam had to be removed by carefully peeling the tape off of the case. Foam removal took up about 80% of the time I spent repairing the 85.
and the caps are stabilized with very primitive stabilizers and shallow stepping to encourage typists to only type on the keys where they are most stable. In reality typing anywhere on the stepped parts is fine, the keys bind slightly but still go down well enough. It is my opinion that this keyboard was designed either shortly before or shortly after the XT. If it was designed before it could explain why the XT does not use this style of stabilizer as these would have shown this style to be non-ideal, and if it was designed after it could explain why the XT doesn't have stabilizers at all as they would not have been developed yet. Either way it is very clear to me that IBM was not satisfied with the stabilizers used here and did not use them at all on the AT, instead opting for wire stabs for most keys and a rod stab more reminiscent of a model M rod stab than what is used on the ET. These are a very interesting look at the evolution of the buckling spring design for sure. In addition to the weird stabilizers used they also employ O rings, which apart from a few Wheelwriters I have not seen IBM use before. The O rings are pretty hard on both the ET and the wheelwriter module and don't effect the bottom out feel but do make the sound a but nicer IMO.
And here we have the switches that the Power and on/off toggles are attached to. Neat to see the IBM branding on them. If you count these and the code key then there are 3 different IBM switches in this machine.
Here are some choice diagrams from the service manual:
Spoiler:
The on/off toggle is not actually connected to the keyboard matrix. See the next pic for how it actually operates
I have personally never seen an iso version of a 65/85/95, but I bet they would be just as nice.
And here she is all fixed/cleaned up. I would still like to bolt mod this keyboard as well, but until I have a good solution for the cracking around the on/off toggle I'll hold off on that. The mounting brackets for the buttons left of the code key is also damaged on the 95 so I still need to repair that and make it sit right in the machine, some design and 3d printing is likely in order there as this plastic seems like it is brittle and a replacement for it would likely be helpful for other folks with the same problem.
“When I saw those x-rays, my response was, ‘holy fuck.’ They really were bugging our equipment. I was very excited, but no one was around to tell the news. My wife was an NSA employee, but I couldn’t even tell her because of the level of classification.”
