Author Topic: Let's replicate the Selectric keyboard mechanism  (Read 11261 times)

0 Members and 1 Guest are viewing this topic.

Offline enthdegree

  • Thread Starter
  • Posts: 13
Let's replicate the Selectric keyboard mechanism
« on: Wed, 04 March 2020, 00:38:33 »
I own a mostly-functional Selectric II and after examining it for a while it seems within the realm of possibility to design and build a device that essentially replicates its typing feeling by re-implementing the keyboard mechanism.

The keyboard portion of the Selectric mechanically converts a keypress into a bit string. Modulo friction and spring un-idealities, each piece of the mechanism fairly directly controls its part of the keypress force curve, a cartoon of which is shown here:
http://bhtooefr.org/images/SelectricComposerForce.png

An image on the following page is helpful for understanding the moving parts involved, although if I hadn't had real Selectric guts to look at it would have been hard for me to figure out what was going on.
https://www.cryptomuseum.com/covert/bugs/selectric/#bin

Essentially I am imagining deleting everything on the Selectric past the key-restore cam and replacing hard-to-source parts with easy ones (like helical springs instead of an ornate leaf spring). What is left over is a potentially extremely nice and interesting computer keyboard. Maybe with solid housing a solenoid could be added to emulate the feeling of the typeball firing.

Has this been attempted before? You guys think this a stupid idea for a project?
« Last Edit: Wed, 04 March 2020, 00:52:27 by enthdegree »

Offline yui

  • Posts: 1082
  • Location: 127.0.0.1 (in azerty)
Re: Let's replicate the Selectric keyboard mechanism
« Reply #1 on: Wed, 04 March 2020, 01:05:07 »
from what i have read modified Selectrics where a pretty common keyboard for early computers. although it came with some majors drawbacks
1-1 key rollover, the keys get locked between key presses to make sure to only press one key at a time, and that is because of the encoding mechanism, so even if you remove the typing you will still get only 1 key rollover
2-Control keys, well those are missing
3-Mechanical to the extreme, it does make much more noise than even a model F, with that motor needed to make the whole thing work and much heavier weighting as well
4-There might be other reasons that i do not know/remember

I guess IBM had a reason to develop the beam-spring switch, they wanted the Selectric without its shortcomings.
Other than that you could do it, and you would not be the 1st doing so, linus tech tips even did a video on converting an old typewriter to usb
vi vi vi - the roman number of the beast (Plan9 fortune)

Offline Findecanor

  • Posts: 5035
  • Location: Koriko
Re: Let's replicate the Selectric keyboard mechanism
« Reply #2 on: Wed, 04 March 2020, 04:37:57 »
2-Control keys, well those are missing
Would it be possible to reuse the Lock key for that purpose?

(like helical springs instead of an ornate leaf spring).
Just don't replace any spring that is part of the key feel!

Offline enthdegree

  • Thread Starter
  • Posts: 13
Re: Let's replicate the Selectric keyboard mechanism
« Reply #3 on: Wed, 04 March 2020, 23:29:17 »
from what i have read modified Selectrics where a pretty common keyboard for early computers. although it came with some majors drawbacks
1-1 key rollover, the keys get locked between key presses to make sure to only press one key at a time, and that is because of the encoding mechanism, so even if you remove the typing you will still get only 1 key rollover

I was thinking about this a bit today and the locking mechanism shouldn't pose a problem for NKRO.

A keydown on a selectric sets an interposer into an active position and then pushes a cycle bail. When the cycle bail is pushed far enough that a latch gets past some ledge, the selectric fires the typeball with the interposer in place, while returning the interposer to its initial position. This is all without a keyup happening. While you are still holding that key is down you can go press other keys and the same described process will happen. The only thing the locking mechanism prevents is attempting to push keys down exactly simultaneously.

One difficulty I can think of is the need to develop an additional mechanism that broadcasts keyups. I am thinking keyup signal needs to be created using the keylever itself somehow, since that's essentially the only moving part. I hope an elegant solution to this can be developed.

2-Control keys, well those are missing
That's fine, I can give meta keys their own set of selector bails.

3-Mechanical to the extreme, it does make much more noise than even a model F, with that motor needed to make the whole thing work and much heavier weighting as well
This is a feature, not a bug.

4-There might be other reasons that i do not know/remember
I guess IBM had a reason to develop the beam-spring switch, they wanted the Selectric without its shortcomings.
Other than that you could do it, and you would not be the 1st doing so, linus tech tips even did a video on converting an old typewriter to usb
I don't think IBM and the guy you mention, Linus, had the same goals in mind...
« Last Edit: Thu, 05 March 2020, 22:41:35 by enthdegree »

Offline enthdegree

  • Thread Starter
  • Posts: 13
Re: Let's replicate the Selectric keyboard mechanism
« Reply #4 on: Fri, 06 March 2020, 21:59:56 »
I've been overcomplicating things... All the selector bails and interposer patterns are totally unnecessary.

Keypress-to-digital signal conversion can be done just by a sensing mechanism on the microcontroller, say, a capsense or magnetometer array right underneath the keylevers. This automatically gets you NKRO and everything.

Offline Findecanor

  • Posts: 5035
  • Location: Koriko
Re: Let's replicate the Selectric keyboard mechanism
« Reply #5 on: Sat, 07 March 2020, 04:53:45 »
Jack Zylkin sells a USB+Bluetooth Typewriter Conversion kits but not in any variety that supports the Selectric.
They have a single strip of switches aligned with the key-arms, but I am not sure about the type. There are a bunch of articles on Zylkin on the web though that might contain info on what sensing mechanism he used.

On these mechanical typewriters that the kits were made for, it was (in unmodified form) the speed of the key press that made it actuate, so there is no definite actuation point..
From what I understand, the Selectric does have a distinct actuation point though, and I think that you'd want to tune that more exactly. Maybe that would be as simple as choosing the right thickness of shims to get the right distance between PCB and the actuating key-arms.

(I'm sorry for not posting about the USB Typewriter earlier despite having seen it years ago. It was too far back in my mind...)

Offline enthdegree

  • Thread Starter
  • Posts: 13
Re: Let's replicate the Selectric keyboard mechanism
« Reply #6 on: Sat, 07 March 2020, 14:12:25 »
Jack Zylkin sells a USB+Bluetooth Typewriter Conversion kits but not in any variety that supports the Selectric.
They have a single strip of switches aligned with the key-arms, but I am not sure about the type. There are a bunch of articles on Zylkin on the web though that might contain info on what sensing mechanism he used.

On these mechanical typewriters that the kits were made for, it was (in unmodified form) the speed of the key press that made it actuate, so there is no definite actuation point..
From what I understand, the Selectric does have a distinct actuation point though, and I think that you'd want to tune that more exactly. Maybe that would be as simple as choosing the right thickness of shims to get the right distance between PCB and the actuating key-arms.

(I'm sorry for not posting about the USB Typewriter earlier despite having seen it years ago. It was too far back in my mind...)

I have a more involved project in mind... I'm thinking of designing the entire mechanism in CAD and making it from the ground up, 3D printed and die cast

And you're correct, it does have tactile feedback, here is a video of the relevant mechanism

For proper feedback there should be a separate sensor that detects when the cycle latch is released so that the keydown event time corresponds exactly to when the board fires
« Last Edit: Sat, 07 March 2020, 14:55:00 by enthdegree »

Offline Maledicted

  • Posts: 2164
  • Location: Wisconsin, United States
Re: Let's replicate the Selectric keyboard mechanism
« Reply #7 on: Mon, 09 March 2020, 01:06:13 »
I think this is a really cool idea, and some of the coolest ideas in history were also, objectively, probably pretty stupid ones. Outsource it to some German or Swiss engineers, maybe the end result will somehow come out even more complicated than the originals.

Offline enthdegree

  • Thread Starter
  • Posts: 13
Re: Let's replicate the Selectric keyboard mechanism
« Reply #8 on: Fri, 13 March 2020, 01:17:51 »
Initial progress drawing it up. Most still isn't done but this is the heart of it. Still needs a cycle restore mechanism, a keylever, a frame, springs, nuts and bolts. And after drawing those things then all the parameters need to be tweaked so that it works just in theory...

Feels a little like doing cargo cult rituals... 🙃
« Last Edit: Fri, 13 March 2020, 01:23:08 by enthdegree »

Offline Maledicted

  • Posts: 2164
  • Location: Wisconsin, United States
Re: Let's replicate the Selectric keyboard mechanism
« Reply #9 on: Fri, 13 March 2020, 07:10:43 »
Initial progress drawing it up. Most still isn't done but this is the heart of it. Still needs a cycle restore mechanism, a keylever, a frame, springs, nuts and bolts. And after drawing those things then all the parameters need to be tweaked so that it works just in theory...

Feels a little like doing cargo cult rituals... 🙃

Wow, awesome. That's looking great.  :thumb:

Offline yui

  • Posts: 1082
  • Location: 127.0.0.1 (in azerty)
Re: Let's replicate the Selectric keyboard mechanism
« Reply #10 on: Mon, 16 March 2020, 02:55:00 »
maybe going to set records, most likely going to be able to go after heaviest usb keyboard and most over-engineered keyboard. once you are a bit further into the design you maybe want to call guiness about that
vi vi vi - the roman number of the beast (Plan9 fortune)

Offline ThisNameNotUsed

  • Posts: 2
Re: Let's replicate the Selectric keyboard mechanism
« Reply #11 on: Fri, 21 July 2023, 18:57:41 »
Did this go anywhere? My 64 year old mom just namedropped this keyboard saying how much she loved to type on it. How the hell does someone just remember the name of keyboard from so long ago? She says she learned to type on it. She was even explaining it's force curve as if she knew what a force curve was.