(Model MF) Remodeling the Model M (aka.. the Mara)

[lot_lizard]

For the stabilizers, are you going to add tapped holes for the other horizontal locations too? They would need to be there for:

(Both)
Left shift
Right shift
Backspace
ANSI enter
ISO enter?

FEXT
Numpad 0

I'm glad you brought that up. What is everyone's expectation of using the KEYS from an F? My original assumption was that we would just pull the keys from the M that it is going into, and account for the possible stabilizers that the early M's had (the emdude catch), but talking about an F version of this replacement stabilizer... the thought occurred to me about using F keys and I never remembered to post the question of likelihood.


The one caveat... the more we add to this, the more the cost goes up (not much mind you, but some), and I only want one version of the board to avoid complexities. We could have a pattern made up that you could use as a stencil if you ever wanted to add F key stabilizers maybe. Or if it seems like something a reasonable percentage would want to do, we could add it out of the gate. We want to avoid scope creep, but grab low hanging upsides along the way where they make sense

Is there any way for the screw-in stabilizers to attach from the top, rather than underneath?
...
It'd be a lot easier to reconfigure for F versus M wire stabilizers if they could be attached and detached from the top.

You should be able to leave the F stabilizer clips in place if you went back and forth with most M caps. There would be very few M caps that would actually use wire stabilizers (they use the nylon barrel insert in most cases), so they shouldn't need to be swapped. I have actually been thinking that we might just use some sort of sleeve insert (think a bent U) to slide into the stabilizer clip to turn it from a M into an F diameter. It is probably the cheapest approach, and to your point, could be changed without pulling the board apart. Should be 28 gauge steel (.5 mm). I could always make more of them myself if we ever ran out (the sleeves) with my press out back since the metal is so thin. In theory, they would work in M plastic barrel frames if you wanted to use F key on it. Adapters = flexibility in ways they weren't intended

Attaching from underneath will be cleaner for a couple of reasons. The metal of the clip will have more girth than the metal of the top plate, and we might even get away with not having the holes be visible if we use a flat head tap (super clean look). Not sure just yet.

[Rimrul]

I just checked my current M (not intended for conversion) and it has stab points on the oposite side of the ISO enter ccompared to the F style ISO enter. It uses the Nylon inserts for the horizontal keys and stabiliser wires for the vertical ones.

[Techno Trousers]

Thanks, Rimrul. That's valuable input on an ISO Model M (I've never seen one of those myself).

I thought it might be helpful to list all the types of cap sets that could be transplanted (please chime in, everyone, if I missed anything or you have anything to add). I don't know if you'd want to run a poll on this, lizard, but if you're concerned we're coming up with too many options, I'd be okay with workable adapter-type parts for M-F cap conversions.

SSK - As far as I know, these ALL came with the nylon barrel insert stabilizers, except maybe in the case Rimrul is talking about where they also use a wire stabilizers for ISO enter and vertical numpad keys. I don't have my SSK barrel plate here so I can check that.

101 Ms - Early models (late 1985, and at least part of 1986) will have the large wire stabilizers for all 2+ u keys. My 1390131 from April 1986 has these. From late 1986 or early 1987 and onward, all of the caps use barrel insert stabilizers, except again perhaps for the ISO enter and vertical keys? So anyone wanting to convert one of these keyboards will be needing M-sized wire stabilizer points on the left shift, right shift, backspace, enter (ANSI/ISO), numpad 0, numpad +, numpad enter.

F caps - those wanting to use F caps and/or space bars will need an accommodation to accomplish that, either F-specific stabilizers, or F cap to M stabilizer wire adapters.

From my personal perspective, I happen to have a couple of F-122s that are being sacrificed for this project. So I have some choices on using the really cool caps from those. These are the wire stabilized ones that I'm seriously considering using, just because they are symbol-only, and I think will look totally awesome:

Right shift -- the F-122 right shift has only the "hollow up arrow" logo on it, no "Shift". It looks nice and clean.
Backspace -- this also has just the left arrow on it, but the tail is much longer
0 -- I mentioned before that the 0 on the F-122 numpad doesn't have the alternate legend printed on it.
Space bar -- I'm hoping to compare the M and F space bars head to head, and make a final choice about which one I prefer (if there's any difference).

I also want to use the F Tab (long right arrow graphic) and Caps lock (lock image) from the F-122 caps, but those aren't stabilized anyway. Finally, I'm considering using the short left ISO shift to open a third modifier key where the right side of left shift would be. Again, that doesn't affect stabilizers, except that I wouldn't need a left shift stabilizer point.

I think that covers everything. Again, if anyone has input on this, now's the time to pitch in.

Thanks a bunch for considering all the nitpicking, lizard!

[lot_lizard]

I just checked my current M (not intended for conversion) and it has stab points on the oposite side of the ISO enter ccompared to the F style ISO enter. It uses the Nylon inserts for the horizontal keys and stabiliser wires for the vertical ones.

Your point echos loudly how many options were in these original M's. Let's plan on making all of the holes in the top plate for the stabilizers on the keys mentioned. Again, its pennies to add, then we are covered. For those of you that might be worried about extra holes, it won't impact anything structurally. As for debris, we are going to be using a 2-part foam instead of the traditional. I haven't worked out the specific materials yet (probably next week I will go visit the foam plant a few miles away), but have had conversations already to prep them. The top layer of the "foam" (really rubber) will be a more laytex debris shield/sound deadening type material that will be the full size of the top plate (including sandwiched at the top/bottom plate intersections). The second layer (touching the barrels) will be a more tradtional closed cell foam that will set INSIDE of the offset bends (never touches the bottom plate). The current goal is that this is 4mm+ thick (both combined) before compression as opposed to our traditional 1.5-2mm approaches... Thank you screw fasteners . We should end up with something pretty special in the end.

Attached is a quick mockup of the F adapter sleeve I mentioned previously for the stabilizer clips. The 90 degree angles would be closer to 95-100 to retain outward extension against the stabilizer clip

F sleeve adapter in orange

stabilizerClipFSleeve.png (7.87 KiB) Viewed 4864 times

[Rimrul]

My mid 1986 1390148 (ISO-DE, 102 keys, XT, no lock lights, detachable cable) has 4 keys with a wire (ISO enter, numpad enter, numpad +, space) and 3 key with nylon inserts (Numpad 0, right shift, backspace). I don't know much about SSKs, Fs and M122s (but those are probably not that different from 101s/102s regarding stabilising).

[Techno Trousers]

Attached is a quick mockup of the F adapter sleeve I mentioned previously for the stabilizer clips. The 90 degree angles would be closer to 95-100 to retain outward extension against the stabilizer clip

stabilizerClipFSleeve.png

This is perfect.

Are you going to be able to prototype the placement of all of the wire stabilizer points with M and F caps? If you need samples for testing, I'll have to double-check, but I think I have the following wire-stabilized caps I could loan you for fitment and usability checks:

M left shift
M spacebar (duh)
F spacebar
F+M backspace
F ISO enter
M ANSI enter
F+M right shift
F+M numpad 0
M numpad + (on the F-122, this is two separate keys)
F+M numpad enter

[lot_lizard]

Are you going to be able to prototype the placement of all of the wire stabilizer points with M and F caps?

I think I am good there looking at the list. It is sad what I have stored away, but I think most/all are accounted for. If others see something Techno missed (thank you for the list btw), please chime in. Again, the machine shop doing this ahead of time saves someone hacking their way through it later. They are very happy to take our money .

Honestly though, you guys are being cut a REALLY good deal. There is zero chance this happens outside of China (maybe even in) for these prices without them being inspired by our efforts, the idea behind it, and WANTing to participate. The plates and powercoating (two separate local companies) are the predominant costs, and they each threw us a bone. I have confidence in them (friends), and know the deliverable will turn out well.

BTW... I still haven't put together the FEXT. I have test the barrel spacing, but that is about it to date. I will make sure the FEXT gets more a focus when I receive the "final" stainless steel versions in the proper thickness.

[Techno Trousers]

Hopefully someone can lend you that M ISO enter that uses stabs on the right side.

It feels like this project is blessed under your leadership, lizard. In hindsight, it's obvious that what a mechanically oriented switch project like this needed was someone like you on "the inside" who knows metal fabrication and is close with people who do that kind of work for a living.

Maybe I should end up ordering a second SSK drop-in assembly, and keep it mothballed just in case I ever need it!

[lot_lizard]

Appreciate the kind words. It has taken several large leaps by multiple key players over the years to make any of this possible. The real sign of a successful project is that it can be built upon later by others (community collaboration). We are just proving the validity of prior work in a coordinated way.

With appropriate timing, this is the latest version of the DXF plates for both the FSSK and FEXT. Keep in mind they were built with code (OpenSCAD) that produces meshes, so the models are faceted. This isn't an issue for die cutters for metal or foam (all this is needed for), but might appear slightly pixelated when viewing. We will publish more later (the actual OpenSCAD code) when we have a hardened final. We still need to add the stabilizer clips for all of these combinations, and anything else that pops up along the way. In the meantime, if you are interested in seeing the code and are already familiar with OpenSCAD, PM me.

[fohat]

The tab for the space bar wire sits quite a bit farther forward on the Model F than on the Model M.

I will try to measure their lengths tomorrow. The insert concept should still work, I think, but it might need to be more than a simple "U"

[lot_lizard]

The tab for the space bar wire sits quite a bit farther forward on the Model F than on the Model M.

I will try to measure their lengths tomorrow. The insert concept should still work, I think, but it might need to be more than a simple "U"

Agreed... When I measured before, I thought it was ~2mm (which implies the wire moves more laterally on each stroke since the pivot is closer). We could add a third equidistant screw further forward if required (worst case), or make the adapter clip extend past the standard clip (more ideal), and like you pointed out earlier, taper it a bit. It for sure needs work, but seems sound in practice. Help measuring is greatly appreciated.


EDIT: Also, if we can prove the idea and clip placements are sound, we can always refine and produce the clip adapter itself at a later date. I wouldn't think it is on the critical path to have manufactured and in-hand before proceeding with this first phase shipments (though ideal if possible). These would be very cheap to mass produce and ship independently. Still need to prove the idea though. Right now it is little more than a wild hair

[lot_lizard]

Last post for tonight, and I'll stop bombing the thread (I'd still appreciate your thoughts though)

When I met with our metal friends today, I discussed the "engine turning" (Spirit of St. Louis) that I was going to do myself, and they offered up these local guys (cheaper because of reduced transit cost from plant to plant) below as an option, and suggested the pricing was "reasonable". The sheets would come pre-polished with designs, and they could cut and bend without damaging the patterns. I didn't ask what reasonable really meant, but let me know if there is interest and I will. This would only be used for the yellow Chromate backplates (so picture these patterns in gold). It is cheaper to hot zinc dip than powdercoat, so in the end would likely be close to a wash. Otherwise raw would equal slight savings.

http://metalserve.com/pattern-polishers.html

[Vizir]

tagging! very cool!

[Hypersphere]

The engine turning patterns are attractive, and it would be nice to know that they are there, but given that this part would normally be hidden from view, I would be fine with the raw metal.

[lot_lizard]

So I have been on the phone with with one of the founders of Fastenal today chatting about our efforts of non-profit goodness, and he is wanting to help. They will be providing all our fasteners in stainless steel, along with our o-rings for the tension screws, at 50% of retail. For those that don't know, they originated locally to me, but are traded publicly (Nasdaq), and are no small firm (our project is TINY compared to their typical discount tier).

At this point, we will be well below my original (pessimistic on purpose) estimates. I will be very public with the actual cost of each part when we are officially ready to group buy, but plan on a mandatory ~5 dollar donation that will be presented to my local volunteer fire department (amounting to all the proceeds). Several of these local firms are inspired by our story... pricing our project accordingly, and I would like to show that we gave back (keep phase 2 in mind... politician at work). The original 150 estimate is looking closer to 130 for the SSK (without switches, and already factoring in the donation), and ~136 for the FEXT. I still need to harden the numbers for the PCB and controller after I$'s revisions, and the soldering cost so we can just use header pins with ribbon cable connectors, but these numbers are seeming very realistic at this point.

[alh84001]

That's great to hear. And it's always lovely to see someone from a big firm keep an eye out for hobby projects and finding ways to help them. Or us, in this instance

[Techno Trousers]

Such fantastic news! I've heard plenty about Fastenal, and I'm on the left coast. At those prices, I'll buy an extra unit for sure. Donations to a volunteer fire department are a great touch. I have so much respect for volunteers like that.

[elecplus]

That is quite an accomplishment! Kudos to you!

[pyrelink]

That's actually quite impressive. More than happy to add a donation for volunteer firefighters. I am friends with a bunch at my local station. Great work again Lot.

[lot_lizard]

Thanks guys... Appreciate. I had a cheesy post a bit ago that I deleted, but the moral is everyone should really be thanking Cindy. We wouldn't be proceeding without her efforts in finding the switch "haul". Otherwise we would be several additional months away.

There was an update to the fasteners I should note. Before I was using M2 zinc plated flat head screws countersunk/tapped for everywhere but the top screw locations (which are still M3 with fender washers and a nylon lock nut). We are replacing all the M2 with M2.5. That pushes the LIMITS (M3 are just too big), but the M2.5 are actually a little cheaper, and add ~1.2mm to the overall circumference of the screws. M3 would be ideal for multiple reasons, but is just not a possibility around the edges since our tolerances are so tight. The M2's I am using on the prototype are working well, but we want the largest diameter possible in the long run.

Also, stainless screws are actually stronger than untempered steel. Fasteners this small are rarely tempered/hardened since they tend to warp significantly (learned that today). I knew they were grade 2 (unhardened), but when I asked if we could get grade 4+, he explained why they just don't exist. Stainless also had the upside of being more attractive imo (but again, these are hidden with the case on).

[Techno Trousers]

I don't quite understand all the implications of the screw choice, but it sounds like we're getting larger, stronger screws than we thought we'd have so that's a good thing!

I was curious too about the o-rings for the center tension screws. Are those going to be placed between the plates, or...? How do they help improve the solidity of the sandwich?

[lot_lizard]

I don't quite understand all the implications of the screw choice, but it sounds like we're getting larger, stronger screws than we thought we'd have so that's a good thing!

I was curious too about the o-rings for the center tension screws. Are those going to be placed between the plates, or...? How do they help improve the solidity of the sandwich?

Larger diameter screw shank equals a better "bite". We would take the largest possible diameter since the threads (cuts into the screw sleeve) themselves are deeper the wider the diameter of the screw. Our tolerances are REALLY tight though, so I started with what we knew would work, and expanded where the opportunity arose. It did, but not by much. What I previously had worked well, but any chance for improvement should be captured.


The o-rings on the tension screws are really Phosphorglow's brainchild with his M bolt mod of my recent board, but the concept is simplisticly brilliant. Adding the o-ring around the portion of the screw that holds the bottom plate (the countersink) reduces any vibration transfer between the plates. Without, If you tighten these tension screws too much, you are actually making the sound louder since the screw carries vibrations between the plates then (becomes a true "standoff", which is not ideal). The o-ring eliminates any direct top to bottom plate contact (other than the edges, which need to be highly compressed regardless). Dialing in a specific board's tension becomes a much easier process when you are also compressing a separation layer (the o-ring).

These are vague words unless you can picture the action. Let me know if a diagram would help. In the end, it will make world of difference.

[pyrelink]

I would appreciate a diagram if not too much effort for you. I am currently imagining these o-rings functioning like washers on the counter sunk side of the plate? Making the compression a little less stressful on the plates and the screws, and insulating the screw from the other side of the plate. Makes a lot of sense actually. What are the size of the o-rings you are planning, and did Phosphorglow actually use an o-ring on every single one of the screws/bolts on your bolt modded board?

[fohat]

Adding the o-ring around the portion of the screw that holds the bottom plate (the countersink) reduces any vibration transfer between the plates.

How is this not moot with a nice fresh mat separating the plates and bedding the barrels?

I can see the point on the Model M, I think I even tried it once.

[lot_lizard]

I would appreciate a diagram if not too much effort for you. I am currently imagining these o-rings functioning like washers on the counter sunk side of the plate? Making the compression a little less stressful on the plates and the screws, and insulating the screw from the other side of the plate. Makes a lot of sense actually. What are the size of the o-rings you are planning, and did Phosphorglow actually use an o-ring on every single one of the screws/bolts on your bolt modded board?

The previous picture of the screw with o-ring was taken from Phosphorglow's photo archive of the recent build. The airline board in the photos is mine, and was used on it. Some crude mock-ups below, but everything should be to scale

As we would be today

tensionScrewNoORing.png (3.07 KiB) Viewed 4613 times

Addition of the o-ring

tensionScrewORing.png (4.84 KiB) Viewed 4613 times

Under compression

tensionScrewORingCompressed.png (5.91 KiB) Viewed 4613 times

How is this not moot with a nice fresh mat separating the plates and bedding the barrels?

I can see the point on the Model M, I think I even tried it once.

Totally agree that this could be moot, but worth the experiment I think. We are going to have 14 tension screws in the SSK, and 18 in the full size, so spread across that many instances, I could see mild audible benefit (if any), but the bigger advantage will be that it would allow you more easily dial in the tension that you would desire (I believe that was Phosphorglow's reason over an vibration reduction). So one full twist of the screw is actually compressing less than a full thread pitch because you are technically compressing the o-ring instead of the plate itself with the screw head as tension increases more slowly. Again, until we can test, this is all academic. The full set of o-rings for the board would cost in the neighborhood of a 40 cents (purchasing in bulk) if we decided we liked it.

[fohat]

I don't really have any experience with a "screw" mod because I have always done "nut & bolt" mods on both Model F and Model M, and I try to put washers in wherever I can (which is usually almost everywhere except for the bottom rows of the M). I use pan head screws in the shortest length that I can get away with, usually M2-8mm for the short ones and maybe M2-10mm for the longer ones on an M and M3-12mm on the Model F (because there in a lot more clearance in that case).

Both keyboards have a resilient layer somewhere in the sandwich, and I have always assumed that this provided the tension and cushion that I was looking for.

[lot_lizard]

I don't really have any experience with a "screw" mod because I have always done "nut & bolt" mods on both Model F and Model M, and I try to put washers in wherever I can (which is usually almost everywhere except for the bottom rows of the M).

Thank you for mentioning that. We should include some washers and nuts in case you wanted to "lock" the tension in place. Even though these are tapped and would hold nicely, I could see the screw shifting slightly over time (moving the board around, etc), and a nut on top of the top plate tightened against the tapped surface would secure any tension bolt you deemed critical from ever changing.


The screws just make assembly easier since these parts are so tiny, so with tapping and a nut, we should get the best of both worlds. This will be a little different than an M mod because we will be compressing the plates at the same time, and having the extra hand (no nut initially) would help. Plan on all this stainless hardware (even with these extras we are talking about) still being less than 5 dollars a board (probably closer to 3-4).

[fohat]

a nut on top of the top plate tightened against the tapped surface would secure any tension bolt you deemed critical from ever changing.

I have always run my screws from the top (face) down with nuts on the back (bottom) side. Not that looks are all that important, but screw heads around the keys look cleaner than nuts. Of course with a "screw" mod there is little or nothing visible from the top anyway.

Ms have some clearance issues near the front (spacebar) side but Fs sit up a little higher.

And there is the issue of interfering with the downward travel of the keys, so you shouldn't let anything protrude more than a couple of mm off the front plate on the key side.

[lot_lizard]

I have always run my screws from the top (face) down with nuts on the back (bottom) side. Not that looks are all that important, but screw heads around the keys look cleaner than nuts. Of course with a "screw" mod there is little or nothing visible from the top anyway.

The appearance would be the unfortunate trade off between the ease of screw assembly and the benefit of bolt locking, since the unit would need to be assembled upside down. Not sure a way around it. Originally I wanted to use something called floating PEM fasteners (stronger than plate taps), but the front lip of the top plate offset just doesn't allow for it. Rethinking it now though as I type, we could certainly use them for the tension screws in the middle of the board. I will check back into the pricing there. It is a fascinating little device that is hydraulically pressed into the plate for a very secure bond.



See the link below for an animation of the bonding process:
http://www.pemnet.com/popup/animations/ ... ating.html

[Techno Trousers]

So if the floating PEM fasteners are used in the middle, would they be pressed into the bottom of the top plate, so the screw is caught coming up from below? Our do they need to be at the hole where the screw enters?