Replacement Foam Mat for IBM Model F Keyboards?

[Hypersphere]

Interesting. I suggested O-rings or something like soft landing pads in one of my posts, but now I cannot remember which one. I was thinking of something that would fit directly onto the barrels.

[Parak]

Seems that mcmastercarr has the atypical 50A o rings at fairly reasonable prices - http://www.mcmaster.com/#2418t121/=sfnke9

Hmm, may be worth a try..

[Hypersphere]

Seems that mcmastercarr has the atypical 50A o rings at fairly reasonable prices - http://www.mcmaster.com/#2418t121/=sfnke9

Hmm, may be worth a try..

Looks like Amazon has them as well for a similar price.

However, I wonder if O-rings would work. Looking at a barrel from an XT, I see that the cylinder is not centered in the rectangular base. Instead, one edge of the cylinder is right up against one edge of the rectangle. Thus, an O-ring would extend beyond the rectangle. Would this cause interference between barrels, or would this be somewhat equivalent to having a foam mat covering the entire back of the front plate?

A quick search for even softer O-rings turned up nothing softer than 50A durometer, but perhaps there is some product of a different name that would work.

[Parak]

Yes, they extend - but this is actually both good and bad - good in that it provides additional padding for the barrel on top, so that it's even like with a mat. Bad in that the thickness of the o ring of row 2 would extend a little to row 1, essentially - so my idea of using thicker orings for the middle rows did not work as well as I had hoped.

[Hypersphere]

Yes, they extend - but this is actually both good and bad - good in that it provides additional padding for the barrel on top, so that it's even like with a mat. Bad in that the thickness of the o ring of row 2 would extend a little to row 1, essentially - so my idea of using thicker orings for the middle rows did not work as well as I had hoped.

Are you going to try a uniform diameter of O-ring in a more compressible neoprene (e.g., 50A)?

BTW, I am very favorably impressed by McMaster-Carr. They have an excellent web site for finding products. In addition, I emailed them with a question, and they got back to me within a couple of hours. I asked about soft O-rings. They said that 50A Buna-N is the softest ones that they sell. However, they also sell neoprene cord stock in 30A, 40A, and 50A, from which you can make your own O-rings by gluing the ends together with Loctite 404, part number 7569A22. Worth a try, or should we use foam sheets instead?

Yesterday, I punched holes in a sheet of the 2mm-thick art foam. I made the mistake of repunching some holes that were slightly off-center. As a result, some of the holes have little shards of foam that could break off. I will need to clean these up before using this mat in a keyboard. The 2 mm thickness also seems a bit thin, but I have not yet tried to reassemble the plates (I am waiting for the first coat to dry; then I want to sand and apply another coat).

I have some art foam that is 1/8-inch thick (0.125-inch or about 3.18 mm). However, this is in the form of fairly stiff board, and it might not be sufficiently compressible.

Thus far, it appears that the extra soft neoprene at 3/32 or 1/8 inch would be the best material, but I would like to compare this with silicone foam/sponge, which should arrive this week.

[Parak]

There's another experiment that I've already tried with some success, but need to retry tonight. Basically, thin strips of soft silicone rubber, 1/32, width of the portion of the barrel from the bottom to the barrel itself. I had already punched a silicone sheet of the same material earlier in the thread, but the middle rows were once again a bit loose, so layering of the strips was attempted. Basically, the same issue was encountered where the middle row being solid raised the plate far enough that row 2 and row 4 would get wobbly again. Even KS top shell pressing into the plate curve is insufficient for the keys beyond the two or three next to the sides. Meh.

[Hypersphere]

The 3/32" Buna-N 50A durometer O-rings arrived this morning from Amazon. They are quite hard, not resilient like neoprene or silicone foam. I don't think they will be suitable as a foam layer replacement.

The silicone foam from McMaster-Carr arrived. Unfortunately, the different types are not labeled, so I will have to try identifying them correctly. However, it appears from initial inspection that the white 1/8" foam might be the best material; better than white 3/32" or red 1/8". There is also some gray 1/8" that might be sufficiently resilient. However, these materials are all very expensive compared to neoprene foam.

[vivalarevolución]

The 3/32" Buna-N 50A durometer O-rings arrived this morning from Amazon. They are quite hard, not resilient like neoprene or silicone foam. I don't think they will be suitable as a foam layer replacement.

The silicone foam from McMaster-Carr arrived. Unfortunately, the different types are not labeled, so I will have to try identifying them correctly. However, it appears from initial inspection that the white 1/8" foam might be the best material; better than white 3/32" or red 1/8". There is also some gray 1/8" that might be sufficiently resilient. However, these materials are all very expensive compared to neoprene foam.

Mcmaster Carr usually numbers their foam to correspond with the order list. Are there stickers with numbers on he foam anywhere?

[Hypersphere]

The 3/32" Buna-N 50A durometer O-rings arrived this morning from Amazon. They are quite hard, not resilient like neoprene or silicone foam. I don't think they will be suitable as a foam layer replacement.

The silicone foam from McMaster-Carr arrived. Unfortunately, the different types are not labeled, so I will have to try identifying them correctly. However, it appears from initial inspection that the white 1/8" foam might be the best material; better than white 3/32" or red 1/8". There is also some gray 1/8" that might be sufficiently resilient. However, these materials are all very expensive compared to neoprene foam.

Mcmaster Carr usually numbers their foam to correspond with the order list. Are there stickers with numbers on he foam anywhere?

No, unfortunately, I did not see any stickers. However, I think I will be able to sort it out. In general, I like the white or gray silicone foam better than any other materials I have seen (including neoprene foam and art foam); however, the silicone products are very expensive.

[Hypersphere]

Finished painting the barrel plate on an XT. Installed a new foam layer consisting of 1/8"-thick white silicone foam from McMaster-Carr. I had considerable difficulty getting the plates to slide back together, despite using spring clamps to hold the plates together and a bar clamp (with a straight holder rather than a swivel holder). Finally got the plates to align sufficiently so that I could fold the tab back over to lock everything in place.

It could be that the 1/8" foam is somewhat too thick or that the surfaces provide too much friction to allow the plates to slide readily. If 1/8" is too thick, then the 3/32" or 2 mm thickness could be used. Regarding sliding friction, the surface of the art foam seems to have less friction than the surface of the silicone foam. The art foam is also much cheaper than the silicone foam or sponge.

Because I did not use the art foam that I had punched as the foam layer, I decided to put art foam between the outside of the back plate and the case to provide some additional sound deadening.

I used the fishing line trick to keep the spacebar hammer in place during the realignment of the plates. However, I used black multifilament line instead of clear monofilament line. The black line give me some trouble pulling it out after assembly; I think the monofilament clear fishing line would have less friction and would be easier to remove.

Despite these problems, the refurbished XT works just fine -- I am typing on it now, using a Hagstrom converter box to connect the keyboard to my Mac. I plan to install a Teensy and Soarer's converter later.

This is my second XT. I did not do any restoration on the first one other than to do a thorough cleaning of the keycaps, case, and barrel plate (without separating the plates). The first XT seems to have a somewhat higher-pitched click with each keystroke, but it does not have a "ping". On the other hand, the second XT, which I have refurbished, has a pinging sound in many of the keys. I am considering doing a "floss mod" to see if this alleviates the pinging.

However, the spacebar on the second XT seems smoother and quieter than the one on the first XT; this might be because I lubed the stabilizer spring with Mech Lube 2 at all its contact points, but it might also be a function of the different composition of materials within the case of each keyboard.

Looking forward to my next steps: installing a Teensy and Soarer's converter in the XT, doing an ANSI mod on an AT and F-122, and refurbishing my IBM Displaywriter -- including installing Xwhatsit's controller!

[JBert]

I used the fishing line trick to keep the spacebar hammer in place during the realignment of the plates. However, I used black multifilament line instead of clear monofilament line. The black line give me some trouble pulling it out after assembly; I think the monofilament clear fishing line would have less friction and would be easier to remove.

Soon after I proposed this trick, someone came up with Teflon ribbon/wire as a supposedly better alternative. I haven't tried it, but it should in theory have the least possible friction. I believe the real trouble lies in getting some.

[Hypersphere]

I used the fishing line trick to keep the spacebar hammer in place during the realignment of the plates. However, I used black multifilament line instead of clear monofilament line. The black line give me some trouble pulling it out after assembly; I think the monofilament clear fishing line would have less friction and would be easier to remove.

Soon after I proposed this trick, someone came up with Teflon ribbon/wire as a supposedly better alternative. I haven't tried it, but it should in theory have the least possible friction. I believe the real trouble lies in getting some.

Thanks for chiming in. I should have included a reference to your post about the fishing line. I doubt if I did it correctly, although it seems to have worked. I ended up with the line at the bottom of the keyboard instead of at the top.

As for a source of teflon thread, I haven't seen any. Maybe we would have to feed tetrafluoroethylene to spiders and harvest their teflon silk.

[wcass]

Oral-B Glide dental floss is teflon.

[quantalume]

Oral-B Glide dental floss is teflon.

That's good to know. No wonder why it's so expensive, for dental floss, at least.

Thanks for documenting your findings, Hypersphere. I have some 3/32 silicone that I'm going try when I get back from vacation.

[JBert]

Oral-B Glide dental floss is teflon.

Ah, that makes sense. I think it was a user named "skcheng" on GeekHack who proposed it, and he was known to be a dentist.

[Hypersphere]

Oral-B Glide dental floss is teflon.

Thanks for letting us know!

[Hypersphere]

Oral-B Glide dental floss is teflon.

That's good to know. No wonder why it's so expensive, for dental floss, at least.

Thanks for documenting your findings, Hypersphere. I have some 3/32 silicone that I'm going try when I get back from vacation.

You are welcome. Let us know how the 3/32 silicone works out for you.

One of the silicone varieties that I have is a closed-cell gray foam that has a smooth skin on both sides. I think that this might work better than the material I used with respect to sliding the plates back into position -- it should have less friction. Unfortunately, this material is less compressible than the one I used, and it is 1/8; I don't think it is available in 3/32.

[wcass]

I finally got around to making a template for 122. Due to the size of the pad, i'm publishing it two ways - one uses two sheets of legal (8.5x14) - the other as ArchC (18x24).

Model F 122 pad(legal).pdf

(77.64 KiB) Downloaded 478 times

Model F 122 (archC).pdf

(56.33 KiB) Downloaded 405 times

[Hypersphere]

@wcass: Thanks for the F122 templates! I appreciate the larger all-on-one-sheet size. I can get them printed at a nearby FedEx office.

Let us know if you decide to make an XT template.

[vivalarevolución]

Great news. Thanks for the templates. This will make things easier for any of us doing a foam mat replacement. I will have to save these on my computer.

[orihalcon]

Couldn't it be possible to add some curvature of the center of the top plate so that it matches the bottom plate making a uniform thickness foam compress more evenly? We'd only be adding like an extra 1/16 to 1/8 inch of depth which shouldn't really change how it fits to the bottom plate from a width standpoint. Problem is that it would be hard to create a consistent smooth bend across the entire plate since it is kind of long.

Probably could be done with some sort of jig that stabilizes the top and bottom edges and clamps that depresses a long heavy steel rod down the center (pulled from the sides with either C-Clamps or pressed from above with a hydraulic press).

You'd almost have to have a milled anvil and backplate die that is a few inches wide and then hydraulically press the plate every few inches moving it down the length of the plate

Any thoughts, or does it sound too risky?

[fohat]

Couldn't it be possible to add some curvature of the center of the top plate so that it matches the bottom plate making a uniform thickness foam compress more evenly? We'd only be adding like an extra 1/16 to 1/8 inch of depth which shouldn't really change how it fits to the bottom plate from a width standpoint. Problem is that it would be hard to create a consistent smooth bend across the entire plate since it is kind of long.

Are you talking about new plates, for "new" Model F creations?

Otherwise, on the F-122, the 3 bolts I recommend adding in my guide help pull those plates together at the center line of curvature.

[DaCiRo]

Thanks to Wcass who was kind enough to share the Keyboard template for the IBM Model F AT keyboard I could finally restore mine after few failed attempts of replicating the template in my own by spraying black paint over the old putrid original keyboard foam sheet.
The Spay over foam try never worked, the wholes where never perfectly aligned but Wcass template is accurate to the mm... amazing!.
I used 2mm art foam from a stationary shop in Tokyo city center. and it seem to be the perfect size. I tried to fit in a 3mm on but was too thick. Nothing in between, I wasn't able to find a 2.3mm~2.4mm sizes he in Tokyo, the one thicknesses I came across were 2mm, 3mm, and 5mm
For the holes I used 12mm and 3mm hole punchers, the 12mm for the main barrel hole and the 3mm for the fitting pins. These holes needed to be very accurate as you know otherwise the foam will fold or overstretch when fitting in the barrels. Overall it took me 4 hours to punch the 168 holes (84x12mm + 84x3mm), including preparation and cutting the spacebar shape. That is too much time for such a simple operation but I took my time to try match the hole cuts to the template accurate to +/- 0.5~1mm,....and it really paid off. Some pictures below:

Wcass Template TOP . Foam Cut BOTTOM

AT template - Foam cut.jpg (27.07 KiB) Viewed 13283 times

Template and Foam - Close Up

Template and Foam Close UP.jpg (26.66 KiB) Viewed 13283 times

Template, Foam and Tools

Template, Foam and Tools.jpg (27.55 KiB) Viewed 13283 times

Tools Only

Tools only.jpg (26.19 KiB) Viewed 13283 times

[/img]

Foam cut set over plate

Foam cut over plate.jpg (23.17 KiB) Viewed 13283 times

Laying down the barrels

Laying down the Barrels.jpg (24.31 KiB) Viewed 13283 times

All Barrels in Place, next is springs

All Barrels IN - Nice!.jpg (26.46 KiB) Viewed 13283 times

Close up - Perfect barrel alignment

Perfect allignment All Barrels in.jpg (21.47 KiB) Viewed 13283 times

and now the springs

Now The Springs!.jpg (27.19 KiB) Viewed 13283 times

All Springs in - Almost completed

All sprigs in - Almost completed.jpg (27.78 KiB) Viewed 13283 times

Plate itself - rust free, cleaned, dusted, primered, and painted White

Plate, rust free, cleaned, dusted, primered, and painted White.jpg (26.46 KiB) Viewed 13283 times

Kept unpainted the plate's date - plastic window sticker on top of date

Plate . Kept the original printed date visible protected with a plastic window sticker.jpg (23.8 KiB) Viewed 13283 times

Big Thanks to Wcass for his template which allowed me to resuscitate this beauty

[quantalume]

You did a great job on that foam. How did you attach the template to the foam? Tape around the edges?

[DaCiRo]

I simply stapled the template print onto the foam 2cms above to the cutting margins.
4 Staples on Top and Bottom longitudinally and evenly distributed and 2 staples on each Right and Left side. That was enough to keep it fixed without the need to use any glue or tape.

Rather than the fixing method to the foam what I realized it was most important to keep the holes aligned while punching them was for the the template print to be thick enough to sustain the surface deformation of the hammer hole punch with every hit, and to have a thick base board (and I mean thick) to punch the holes on.

Regarding paper I made 3 prints:
1x Regular A3 paper (0.1mm) - Discarded right away, too thin and too light, even to sustain shape after stapling it
1x Transparency (0.1~0.15mm) - Ok but not really good deformed too much
1x Thick paper (0.3mm) - The best, minimal deformation

Transparency Paper

Transparency paper.jpg (93.63 KiB) Viewed 13223 times

Thick Paper

Best Paper - Thickness 0.3mm.jpg (88.68 KiB) Viewed 13223 times

Thick Paper - Paper Thickness (approx 0.3mm)

Paper thickness.jpg (70.48 KiB) Viewed 13223 times

The Regular paper and the Transparency deformed too much on pressure to be as accurate as I hoped. See the waved deformation on the top printed lines as a reference. The Thick Paper (0.3mm) turned out to be the best requiring not much strength to perforate the paper and foam in a couple of strokes and with little deformation.

Quick Pressure test - See deformation of the top printed lines to get an idea

Transparency - Pressure deformation

Transparency - Pressure deformation.jpg (91.8 KiB) Viewed 13223 times

Thick paper - Pressure deformation

Thick paper - Pressure deformation.jpg (82.49 KiB) Viewed 13223 times

The Base board
The last but not least, A thick base board. I tried using spare shelf trays that were 1cm thick and half the length of the print. It wobbled and vibrated too much with every hit. A big difference when I moved to use a thicker board.

Thick base board

Thick base board.jpg (90.62 KiB) Viewed 13223 times

Base board Thickness - just to get an idea

Base board Thickness.jpg (94.88 KiB) Viewed 13223 times

Thick base board - The Giant Octopus attack

Thick base board - The giant Octopus attack.jpg (137.11 KiB) Viewed 13223 times

Maybe I punched them in too hard

[quantalume]

Thanks for sharing all those details. I used a thick piece of soft lumber too, but I used a smaller piece and had to move it around frequently. I like the idea of using a larger board. I have a couple of foam pieces I need to punch out where I've already painted the top plate, so I'll be using the template method. The template is larger than A3; how did you line up the two halves?

[DaCiRo]

I actually discarded the A3 print because it was a 0.1mm regular paper. However the template fits for the most part on an A3 print no problem. If you print it centered all barrel holes will show up. You will only lose about 1~2cms on the Right and Left margins (see image below) but since no holes are printed in that area you can simply add these to the foam cut.

A3 print(11.69' x 16.53')

A3 print

A3.jpg (486.75 KiB) Viewed 13172 times

I used a Tabloid size print (11' x 17') which is perfect for printing the keyboard template real size in one single sheet, including the cutting margins.

Tabloid print (11' x 17')

Tabloid

27.94cm x 43.18cm.jpg (397.85 KiB) Viewed 13172 times

A3 and Tabloid print side by side

Tabloid LEFT, A3 RIGHT

side by side.jpg (737.99 KiB) Viewed 13172 times

[0100010]

Those O-rings are an interesting idea. I was contemplating using a split lock washer myself - but figured their use would probably crack the barrels.

When I rebuilt my 4704 107 recently, I tried 2mm craft foam across the entire plate - found that it wasn't thick enough for the home row. Tried 3mm craft foam across the entire plate - found it was too thick for the top and bottom rows. I then cut the home row strip out of both pieces, and tried the 2mm foam on the top/bottom with the 3mm foam on the home row. This ended up sort of cantilevering one edge of the barrels on the home row.

I went back to the store and picked up one foot of automotive headliner foam, which came on a roll 54" wide - and is about 1/4" or 3/16" thick and very soft. It was a pain to cut holes in compared to the craft foam (don't have a punch) - but it applied an even pressure on all of the barrels across the plate. I also found I much preferred the key feel with this softer foam compared to the much denser craft foam.

[Parak]

Yeah, one would need really soft orings to be able to do this - 30a or less. It does seem like mcmaster 8647K21 or similar will be the way to go for now.

[dorkvader]

However, they also sell neoprene cord stock in 30A, 40A, and 50A, from which you can make your own O-rings by gluing the ends together with Loctite 404, part number 7569A22.

Hmm, this could work, but instead of making o rings from cord stock, we just lay down a line of it on each side of the barrel, then compress.

Depending on price I may give it a shot. My favourite F122 keyboard has been in pieces for almost a year now and I want to use it again.

Here's all I came up with on mcmaster:
http://www.mcmaster.com/#8466k11 (the 1/8" 30A durometer neoprene cord mentioned above)
http://www.mcmaster.com/#5234k93 (1/8" OD tube. Depending on a number of factors this could be good. You can even fill it with liquid and glue the ends if you need more pressure holding the plate. Durometer hardness is shore A35! Wider ones are available too.