IBM 3279 Beamspring Restoration

User avatar
FXT
XT

26 Mar 2018, 22:24

Aer Fixus wrote:
Also, I'm thinking about clear coating the outsides of the housings (being sure to mask off the internals) so that if there is any plastisizer left over, it won't leech back into the new foam and ruin it prematurely. Can anyone think of a reason why this might be a bad idea (potentially bad chemical reaction with spray laquer/paint and the soft damaged plastic?) or of a better solution (do they make a plastic stabilizer?).

I tried scraping/cutting off the gooey bits around the top of the housing and it's incredibly soft right into the plastic. It cuts like butter and I can dent it with my finger nails until I cut away until the solid stuff beneath. I figure it's better to cover that up and protect it from future damage (and maintain the original shape as much as possible. It's solid, just deformable) instead of trying to remove it. The isopropyl dries it out so that it's not so sticky and makes it a bit more brittle, but not so much so that I'd be worried.
I don't think there would be any issues using clear coat on the exterior of the housings. I spray painted both metal panels of my F AT because they were rusty and I didn't notice any degradation of the plastic barrels or foam. If you're concerned maybe try 303 Aerospace protectant? I've used it with great success on worn automotive interior and exterior plastics.
__red__ wrote:
It's a 12 hour print. I'll probably abort it about 3 hours in so I can verify that the slider will in fact go and down the guide before committing the printer overnight. However, once I get one working, a whole plate of them takes the same 12 hours.

OpenSCAD.

If you want to download it, here's the code that I'm currently printing!
Interesting, I might have to check it out. I've only briefly played around with Autodesk 3DS Max back in high school, but comp sci students at my university get free access to a 3D printer, which is something I'd like to take advantage of.

__red__

26 Mar 2018, 22:52

JP! wrote: Neat idea __red__. I'd be concerned that the guides would be a bit rough textured and would negatively impact the smooth key feel.
nano.jpg
nano.jpg (265.54 KiB) Viewed 11939 times
See that? That's about 20mm tall. Printer has a layer height of 20microns. In clear resin when I print a cube, the layers are completely invisible to both the naked eye and touch.

__red__

26 Mar 2018, 23:01

__red__ wrote:
__red__ wrote:
Aer Fixus wrote: In order left to right. An in-tact module housing, a melted housing with in-tact sliding rails and slider hole, and a too damaged to be used housing.
Image
How many module-housings do you need?
Might be asking for a reason...ImageImage

Sent from my SM-N920V using Tapatalk
This is 50u, can you see the layers?ImageImage

Sent from my SM-N920V using Tapatalk

__red__

26 Mar 2018, 23:10

__red__ wrote:
__red__ wrote:
__red__ wrote: How many module-housings do you need?
Might be asking for a reason...ImageImage

Sent from my SM-N920V using Tapatalk
This is 50u, can you see the layers?ImageImage

Sent from my SM-N920V using Tapatalk
This is why I pulled it early. Not only can I see I need more clearance on the guides but now I see why the guides were triangular!Image

Sent from my SM-N920V using Tapatalk

User avatar
JP!

26 Mar 2018, 23:41

That's looking pretty good!

User avatar
Aer Fixus

27 Mar 2018, 05:52

A module housing with sticky foam residue (right) on it next to one that stayed clean.
Image

A module that has become "chalky" (bottom) and that flakes a bit when rubbed. These are the ones I'm most concerned about. Hopefully lube will take care of them since I'm not sure replacing them all would be economical.
Image

It looks like the final count is 4 housings that MUST be replaced. And 4 or 5 more that are passable but are a bit chalky and would probably benefit from replacement. I'm not sure if I will even notice the difference once they are reassembled especially since I'm going to be lubing them, but we'll have to see. And then there's like 10 or so that are only chalky on the bottom 1mm of the casing and I'm fairly certain that won't be an issue in the slightest. An issue will likely arise if they sheds the chalky dust and it gets between the sensing pad and the capacitive plate, however.

Also, I did end up cutting away most of the bad material on the tops. The bottoms were soaked in isopropyl and rubbed to remove as much as possible so they don't stick to the PCB. The PCB doesn't appear to be sensitive to the gunk, but I still would like to prevent it from getting contaminated as much as possible. I figure that the tops will just need thicker foam if I cut away too much.

__red__

27 Mar 2018, 16:49

Aer Fixus wrote:A module housing with sticky foam residue (right) on it next to one that stayed clean.
Image

A module that has become "chalky" (bottom) and that flakes a bit when rubbed. These are the ones I'm most concerned about. Hopefully lube will take care of them since I'm not sure replacing them all would be economical.
Image

It looks like the final count is 4 housings that MUST be replaced. And 4 or 5 more that are passable but are a bit chalky and would probably benefit from replacement. I'm not sure if I will even notice the difference once they are reassembled especially since I'm going to be lubing them, but we'll have to see. And then there's like 10 or so that are only chalky on the bottom 1mm of the casing and I'm fairly certain that won't be an issue in the slightest. An issue will likely arise if they sheds the chalky dust and it gets between the sensing pad and the capacitive plate, however.

Also, I did end up cutting away most of the bad material on the tops. The bottoms were soaked in isopropyl and rubbed to remove as much as possible so they don't stick to the PCB. The PCB doesn't appear to be sensitive to the gunk, but I still would like to prevent it from getting contaminated as much as possible. I figure that the tops will just need thicker foam if I cut away too much.
A little tight, so re-printing with a touch more tolerance...Image

Sent from my SM-N920V using Tapatalk

User avatar
Aer Fixus

27 Mar 2018, 17:35

__red__ wrote: A little tight, so re-printing with a touch more tolerance...
Looking good. And looks smooth enough to my eye.

P.S. Nice F107. Is it down to the bare zinc or just a trick of the light?

__red__

27 Mar 2018, 18:59

How on EARTH did you see that? nm. :-)

Yes, it is an F107 and it is bare zinc.
You actually *CAN* polish the zinc to a mirror finish, I have some spots that are so clear that they could be a mirror. The problem is that I don't have the patience to make it a *flat* mirror :-)

As such, I should probably paint it.

User avatar
Aer Fixus

27 Mar 2018, 20:38

I must have got my eye trained to pick out 4704 boards, lol.

I've been looking out for them ever since I saw Kishy's post about them whenever that was. And being my favorite series of keyboards to date (I haven't actually gotten to use a beamspring board yet), I enjoy seeing them being used and restored.

__red__

27 Mar 2018, 21:26


User avatar
Sangdrax

27 Mar 2018, 22:06

Nice. I'm glad you're able to salvage the sliders at least. Wish I had access to a UV light printer like that too. So expensive to outsource that quality.

User avatar
webwit
Wild Duck

27 Mar 2018, 23:01

#dta2018 Best project

User avatar
Aer Fixus

28 Mar 2018, 04:10

Now for some video!

An inspection of a couple modules. 1st) Good, shiny sliding surfaces. 2nd) Bad, damaged sliding surfaces 3rd) questionable due to powdery/chalky material.
Cleaning the slider hole and inspecting it
Cutting off the gooey mess from the beamspring module housing. It slices kind of like stiff, but still damp clay. That is to say, much more easily than a plastic such as this should...
And music time! Listening to The Seduction of Claude Debussy while testing each of the switches to see which ones still have sticky stuff in them and need to be more thoroughly cleaned after the first initial cleaning pass (and the second cleaning pass for switches that failed visual inspection). The answer is six of them. The chalkyness also doesn't effect how free the slider is, but it may impact key feel or cause future failure if it gets on top of the PCB. I'll have to see about that once I've finished.
And I've purchased a few replacement parts from Orihalcon as well as an Xwhatsit. Next step after I complete cleaning on the last few housings is to paint the top and bottom plates and cut the foam. Then I'll have a working assembly to debug before moving onto the case..

__red__

29 Mar 2018, 02:35

... and I now have a fully printable, functional, reliable, case for beamspring modules designed and tested and you no longer need them - lol.

NM -=- I'll upload it to github just in case someone needs one in the future!


Red

User avatar
Aer Fixus

29 Mar 2018, 02:44

__red__ wrote: ... and I now have a fully printable, functional, reliable, case for beamspring modules designed and tested and you no longer need them - lol.

NM -=- I'll upload it to github just in case someone needs one in the future!


Red
There are plenty of housings that may still give me trouble (Ater making that previous post, I've found that many still contact the plate in a few areas and I haven't had a cleaning pass for that). Some of them are still sticky after two or three cleanings as well, so I may want to still grab a few of your 3D printed ones to replace the worst ones. I'm also expecting issues when I go to put them in place due to how delicate they've become (some have formed cracks prior to my cleaning them and the walls have gotten quite thin).

And I could put them on some frequently used keys and give them a good thorough testing to see how they perform in practice as well.

User avatar
Aer Fixus

03 Apr 2018, 04:09

So, I went over all the housings again and there's a good number of powdery ones that I'm just not going to be able to replace. We'll see how they hold up. The sticky ones turned powdery after I de-stickified them. I was able to clean the holes in some of the ones I had written off earlier and they seem to slide fine now.

All in all I've been able to make all but two functional. At least it appears that way. I'll find out for sure once I get the Xwhatsit in my hands.

I'm thinking a silicone based lubricant will do well since a dry lubricant would require an already smooth surface. Especially one that doesn't shed itself...

I've got the foam and the paint ready for the next step and will be posting that soon now that I've got the mechanisms mostly worked out.

User avatar
Aer Fixus

07 Apr 2018, 06:53

Good news! It works! Sleep now. Photos later.

P.S. Touch typing on a keyboard with no keycaps is quite challenging...

User avatar
Aer Fixus

08 Apr 2018, 02:38

Photos incoming! Due to bandwitdh concerns, viewer discretion is advised.
Spoiler:
Last I left it, the board was still in pieces and I had just finished cleaning every one of the module assemblies to the best of my ability. I ordered just enough replacements from Orihalcon to do the job as I came across four total switches that had disintegrated so much that the beam couldn't actually pass the buckling point and therefore wouldn't actuate. When the modules arrived along with the Xwhatsit, I made the push to get it functional that night.
Image

Previous to the modules arriving, I scraped, sanded, primed and painted the top plate.

After much scraping I was left with this. And then there was even more scraping as I did not have access to a sand blaster and the sandpaper would just gum up from the combination rust/paint.
Image

And don't forget the underside. I spent so much time on this. There was a thin plastic layer (left over from the foam mess) and some solvent-resistant adhesive that I had to scrape off progressively as I discovered all the patches of it that I missed as I was sanding...
Image

Sanded and ready for paint
Image

Oh, look! Paint! Well, primer...
Image

Paint! For real this time! It's not perfect, but it's plenty good enough for me, will protect from rust and is generally visually appealing until you start looking deeply into the abyss. Then it reveals two or three drips and a could chips, but I mean, it's not like I'm going to be getting critiqued by the internet, lol.
Image

Oh, and don't forget the foam! It's much easier when you combine two steps. Defining where the holes need to go in the foam and priming go hand in hand.
Image

A few YouTube videos later:
Image

Now to inspect the PCB. Looks fine to me!
Image
Image

Mmmm... It looks quite fine.
Image

Excuse me for a moment...
Image

And the bottom plate. I decided that there was not enough damage to warrant repainting it:
Image

Yeah, sure. There's some bare spots where I scraped off the mysterious goo and charred remains of what I can only imagine the poor programmer's tears who must have spent so many long nights on this thing.

And the worst part is on the backside, anyway. I'm not very concerned even if it does re-rust a tad.
Image

And now. What's next?
Oh. Wow. I think I need another moment.
Image

Yep. Quite a moment. Minus the white balance.
Image

How did I keep the stems scratch free you ask? Or maybe you didn't. That's why I'm posing this rhetorical question. So that even if you didn't ask it, I can answer it. Now that we've got that cleared up:
Soft jaw parallel pliers! Look at that brass.
Image

And just a double check of the backside
Image

Is it really that time? It's that time! I had to cut the connector on the right side since it simply would not slide on far enough to make contact or even hold itself on. Mmm... Glass fiber reinforced!
Image

At that point, I made my last post, attempting to write on a keyboard with no caps and entirely from muscle memory.

The next morning, I got up bright and early to work on the...
No. A DnD session. Shucks...

Hours later, I rush home and put the caps on!
Image

Now, at this point it works! and without much issue. The biggest issue is the spacebar which binds if I push it down slowly and is quite heavy and hard to press because I'm a weak baby (hello, my names's The Spine).

But I just had to see what it would look like when I...
Image

Yes. Quite nice. It's just setting on top. No screws, not cleaned, not aligned right, yada, yada.
Finally, I leave you with what has been dubbed as the "Nervous Andrew" key which replaced the "Happy Cylon" key in the previous photo.
Image

User avatar
JP!

08 Apr 2018, 02:58

Excellent work so far. This one deserved being brought back to life. Do all the keys feel nice? What do you plan on doing as far as paint for the case? Also how bad is the plastic cover?

User avatar
Aer Fixus

08 Apr 2018, 03:20

JP! wrote: Excellent work so far. This one deserved being brought back to life. Do all the keys feel nice? What do you plan on doing as far as paint for the case? Also how bad is the plastic cover?
The keys aren't perfectly consistent, but they are very close. Some sound slightly different, but even then, not by much. All my Model F boards have variations in the spring sounds and even feel in some instances (over a single board). This is about as pronounced as those variances. Maybe even less so.

The case is definitely going to be painted black. I'm going to try to find a black paint with a blue base to match the Selectric caps I've got, but even if I can't do that, I'll paint it black anyway. The current colors I have just don't work with the beige case color. As well as the fact that the door damage prevents me from keeping it the stock color anyway as I don't imagine I'd be able to do the original colors justice without significant effort.

The door appears to be in fine shape. Well, restorable shape. It's not irreparably damaged. The damage does go quite deep into the surface, so I'll need to fill it. Some Bondo should do the trick nicely. I'd consider powder coating the entire case if I could acquire a metal door replacement.

The tab that holds the door shut is quite melted and I'm going to attempt to replicate and replace it. Potentially with a spring steel version, but I'll have to see about that one. I'd like to keep the door and the book there as a minor novelty. In the event I cannot replicate the tab, I'll just remove the spring that opens the flap and it will finally be able to close on it's own which will be good enough for me right now.

codemonkeymike

08 Apr 2018, 03:59

Amazing work so far, it is a real inspiration to those also trying to pull off a restoration. If you happen to find someone to make replacement door for the 3279 I would be interested in also buying one.

User avatar
Khers

08 Apr 2018, 10:14

Oof, all the successful work done in this thread makes me feel bad about not having brought my beamer back to life yet. That said, amazing job!

User avatar
JP!

08 Apr 2018, 14:41

@Aer Fixus This thread may be of interest to you.
photos-f62/ibm-3278-restoration-t9365.html

green-squid

08 Apr 2018, 14:55

JP! wrote: @Aer Fixus This thread may be of interest to you.
photos-f62/ibm-3278-restoration-t9365.html
Great restoration, but the guy just ripped it from the terminal. Could've restored that one as well :(

green-squid

08 Apr 2018, 15:13

But anyway, Aer Fixus, I would like to congratulate you on your amazing restoration project! This is the best restoration job I've seen on DT. You absolutely deserve a Golden Wingnut for this :)

Also, You should paint the case in a fun color like yellow or green or blue or red or purple!!!! It would be the cherry on top!

Image

User avatar
Myoth

08 Apr 2018, 15:23

No.

green-squid

08 Apr 2018, 15:24

Yes.

It's his choice, but what's wrong with it? :roll:

It would turn out great if he doesn't do it with a paint can :lol:

green-squid

08 Apr 2018, 15:34

Image

Selectrics also came in many colors back then.

Not yellow or very FUN colors, but typewriters with those colors can also look great. Again, it's up to him, but ya know, he could spice it up! ;)

User avatar
Myoth

08 Apr 2018, 15:42

green-squid wrote: Yes.

It's his choice, but what's wrong with it? :roll:
do you follow the "ugliest keyboard thread" ?

have you seen just_add_coffee's displawriter ?

no ? wonder why ?

it got posted in this thread, now the picture have been deleted from his imgur because everybody was making fun of it, moreover, it was posted in the "UGLIEST keyboard thread"...

I don't think anybody would like to be that shameful about his keyboard.

But yes, it's his choice after all.
green-squid wrote:
Spoiler:
Image
Selectrics also came in many colors back then.
Those are much duller colors than the ones you mentionned, and they were in classy colourways. Not Black/Deep Blue on "Yellow/Green/Blue/Red/Purple" ...

Post Reply

Return to “Gallery”