FSSK v1.00 - Installation & Users Manual

lot_lizard · 26 Apr 2016, 22:27

And now we have a barrel... I will send the flipper and the barrel off to be printed. Phosphorglow has graciously offered to help me test these designs, and play an integral part on designing the next pieces (the top plate, foam, and bottom plate). Since he genuinely has more experience on the inside of an M than anyone I know... I am happy to have the help

Making progress

Spoiler:

EDIT: The flipper and barrel have been ordered for printing. The estimated delivery date will be 5/9. Unfortunately I will be out of town that week, but will report on the function once I return. In the meantime, I will get with Phosphorglow and see if we can't make progress on the top/bottom plates. This is the unfortunate part about not having a 3D printer in house

Phenix · 26 Apr 2016, 22:32

That looks really good! Hope a GB for the case will happen

idollar · 27 Apr 2016, 06:09

I have posted the design (deltacad and diptrace) and the final Gerber files that I used to order the PCBs to the first post.

shreebles · 27 Apr 2016, 10:40

Phenix wrote: That looks really good! Hope a GB for the case will happen

I think that would take significant time and effort, and lot_lizard suggested to use a normal SSK case first.

lot_lizard wrote: With that, you would have a full F replacement drop into an SSK case (worry about that replacement later).

Phenix · 27 Apr 2016, 10:51

Probably. Badly i dont have an SSK for now. Also need to buy another Model M, for the metal plate..

typos due to my touch screen are possible

shreebles · 27 Apr 2016, 11:33

I might have one or another for sale soon. I wrote something about them in the Price Check thread: help-f53/price-check-how-much-is-my-worth-t16-2730.html

Phenix · 27 Apr 2016, 11:36

Thanks.

typos due to my touch screen are possible

lot_lizard · 27 Apr 2016, 16:30

So last night, I took apart an early and late SSK, a BigFoot, a F122, and a Kishsaver. All have the same 18 gauge steel back plate (using a various grades of steel). So the findings would prove that the thicker backplate idea would be a bit of a boondoggle, and would suggest that the F's lack of resonance is directly tied to the thicker independent barrels, metal top plate, and the foam sandwiched between the two. The flipper, pcb, and rear blanket would likely have some impact, but very little by comparison.

Moral, for the benefit added (really only weight of the overall board), the thicker back plate's extra cost trade off is not desirable. 18 gauge IT IS!!!

XMIT · 27 Apr 2016, 16:54

lot_lizard, I have easy access to a 3D printer here. Shoot me a PM and I'll see what I can do. I can get access to a MakerBot and a really nice (and expensive!) resin printer pretty easily.

As for resonance: the usual quieting mods (grease on the springs, floss in the springs) would help greatly, I imagine.

lot_lizard · 27 Apr 2016, 17:44

I sent a PM. I am very much in favor of shortening the design window where we can

I am insanely jealous of the high res 3D printer access. My friend that did my replacement Kishsaver footbar sold his (I should have bought it). It produced nice results. The following link are some pics (plans above if you scroll up) from the days when I was lost, confused, and wondered the streets of GH

https://geekhack.org/index.php?topic=52 ... msg2097034

XMIT · 27 Apr 2016, 19:36

Oh, neat. I got the PM. We'll see if I get some 3D printer access soon. It's easy enough, just need to head over to the place.

lot_lizard · 28 Apr 2016, 00:10

The backplate... The existing PCB presents an issue in regards to hole locations. We didn't want to use any of the holes in the middle of the PCB anyway, but I did want to use the top and bottom rows to secure the top and bottom plates together. The bottom row works well, but the top row will eat into the base of the barrels. We could notch barrels, but want to avoid that extra cost in production. I had planned on taking advantage of the 5.5mm space between the assembly and outer shell to add tabs that would EVENTUALLY secure the bottom plate of the INNER assembly to the top of the OUTER case (like the 4704 models) if/when we decide to construct a new outer shell. We are going to utilize those tabs to secure the top and bottom plate together (with a spacer sandwiched). This has the added benefit of keeping debris out, and can avoid foam decay as quickly like the previous F. Along with these new side tabs, we will leverage the ground plate screw location along the top of the PCB. This should be an elegant compromise, and leaves the PCB as it exists today.

If someone else could validate that the radius axis for the arc of the plate is ~300mm, I would appreciate since this part is the only "hard" math, and is critical

Spoiler:

lot_lizard · 28 Apr 2016, 23:30

The start of the top plate with revised back plate. I ran into issues attempting to use the new PCB holes when trying to merge the top and the bottom. Essentially, the tolerances for the underlying barrels left no ideal options. So instead, I am floating the PC with an exception of two pass through points at the top. The model F only uses two rivets to hold the PCB in place (middle), and now that has been shifted that to the top. This allowed me to provide some barrel tolerances under either end (top/bottom) of the top plate that just weren't there before, and am very pleased about.

Also, I added the tabs to the bottom plate previously (see last post) for attaching the bottom plate to the top of the outer shell if/when we decided to make a case. I was able to take advantage of those tabs in holding the inner assembly together. We will have spacers on each side to allow tight torsion of the bolts, and also act as a better debris shield than the previous model F (just foam). It can be another plastic part we injection mold along with the barrels and flippers.

I still need to clean up the rounded corners of the top plate, and add the barrel cutouts themselves, but I am very confident of the fit as this point as a drop-in replacement

Spoiler:

lot_lizard · 29 Apr 2016, 19:06

My last post for a while since I am taking over this thread, and that is not my intention. I have everything wrapped up. The barrels have been cut out of the top plate and "test fitted", the PCB was tested for fit, and I ended up modifying the tabs on the sides to account for everything. For this design, a hole will need to be drilled in the PCB at the top right corner (opposite of the ground screw). The only remaining pieces would be the foam, the spring for the flipper, and will likely have a blanket made to separate the PCB from the backplate even though it looks unnecessary from i$'s posts.

I will clean up the sharp corners, double check everything, and start checking on fabrication costs. At minimum, I will have a prototype made for myself

Spoiler:

shreebles · 29 Apr 2016, 19:07

Holy shit. I don't have time to write a long reply, but you have my respect for coming up with all this.
Your posts definitely deserve more feedback.

Basically, you have created the files to make a drop-in replacement for an SSK case, correct? Without making any modifications to the original SSK?

lot_lizard · 29 Apr 2016, 19:13

shreebles wrote: Basically, you have created the files to make a drop-in replacement for an SSK case, correct? Without making any modifications to the original SSK?

Correct... Also, the new side spacers will let us have different mounting options on new cases later (think longer screws coming up from the bottom that screw into the top of the case from below... like the 4704). We have options now

Khers · 29 Apr 2016, 19:27

I'm liking this development a lot. One question though; wouldn't it be good to have some screws in the middle of the plate? From my own experience I know this is perhaps the most difficult area to get right when rebuilding an F. IIRC Fohat always screw mods his F keyboards for this reason.

Or perhaps I'm too ignorant about manufacturing and it's possible to make everything fit perfect when making this from scratch?

lot_lizard · 29 Apr 2016, 19:44

Khers wrote: wouldn't it be good to have some screws in the middle of the plate?

It's very possible to strategically add some holes in the middle (and not hard to line them up), but not in actual middle of the board without modifying a few of the barrel bases. This was the real reason I added the spacers on the sides to attach the plates, even though the original F didn't have. I didn't show it, but the top plate will have a SLIGHTLY shorter arc radius (think more bend) than the spacer or the bottom plate. This will put more down pressure on the middle than anywhere else in the bend UNTIL THE BEND IS CONSISTENTLY FLATTENED TO THE DESIRED ARC.

We are actually substantially more secure than the original F's with this new approach because of the extra rigidity that bonding the plates with a spacer will provide

Khers · 29 Apr 2016, 20:03

But wouldn't different pressures yield different key feel across the rows which would be unfortunate. I'm not sure whether anyone has actually measured the forces involved in keeping an F together, but my bet would be on trying to achieve a homogenous (or at least as homogenous as possible) pressure across the board. On the Model F, I think I'm right in saying that IBM went for something similar to what you are proposing, with different radii on the two plates, but I think it needs to be tuned to the foam and barrels.

I do agree however on that the screws represent a significant improvement on the original which was probably never intended to be taken apart by the end user, and it shows.

lot_lizard · 29 Apr 2016, 20:50

I'm about to sound like a massive nerd... So I apologize in advance:)

Anytime you have a CONSISTENT arc of a uniform material, and it is bent against the arc (flattened) against a consistent opposing arc, the load becomes uniform. So the middle and edges of the arc would have the same pressures. In our case, the spacer (significantly rigid) and steel back plate would have the desired radial bend, and top plate (more elastic due to the addition of the spacer) would uniformly bend to desired arc. I'm actually in a meeting at the moment, but can Google some examples later. It has to do with elasticity and ellipticals/arcs.... It is actually important that they aren't the same arc radius to assure the pressures are consistent.

I edited my previous post with BOLD CAPS about the additional pressure part. I should have been more specific

Khers · 29 Apr 2016, 21:48

I guess I'll have to apologise for my stupidity then; it was a long time ago that I studied mechanics and I have found approximately zero use for it since. Somewhere in the back of my mind, however, I find support for what you write.

No need to google anything, I'm quite capable.

lot_lizard · 29 Apr 2016, 21:56

Not at all... This is the dialogue I was hoping for. If you only knew how many things I had to go back and rethink putting this together

I really do welcome things I might have missed, or ways it could be improved. I only want to pay for the protoype once

Khers · 29 Apr 2016, 22:09

Ok, looking at this on a screen larger than my phone, I can only say that I'm impressed! The one thing I find lacking is space bar stabilizer mounts in the top plate, both for the wire and the extra barrel (though I'm not certain the extra barrel is strictly necessary).

lot_lizard · 29 Apr 2016, 22:58

Thanks... And appreciate you really looking it over. I'm glad you brought the spacebar up. I had added the spacebar stabilizer barrel (after the pics were captured), but haven't added the spacebar stabilizer bar clips yet. Do we want to do what Ellipse did (metal clip sized with a wider gap for the M spring bar)? Or do we have other thoughts...

Ellipse metal clip: https://www.modelfkeyboards.com/wp-cont ... 2/tabs.jpg

I am assuming they are spot welding these into the front of the top plate. The only downsides to that approach are metal on metal (so it will need lithium grease or the like), and that the clip base is raised higher using the gauge metal he is. I'm not sure it really has any impact, but you are changing the profile height of the stabilizer bar somewhat. It will be a while before I see the F77 that I ordered in the first run to really understand what they did there. If the clip base was somehow slipped under the top plate through some cut, then it would impact the barrel base from the row above (doubtful they did that). The following is a horrible image shot from my phone of the Kishsaver clip (pressed plastic)

Spoiler:

EDIT: Stay tuned... I have an idea I think I like better

lot_lizard · 30 Apr 2016, 05:01

So I think this is the most elegant solution. It will require an additional mold for a barrel (which adds some cost), but eliminates manual effort to attach a stabilizer clip for the spacerbar spring... which would be a MUCH larger cost since we don't currently require welding of any sort. It eliminates all metal on metal contact, which is never good unless lubricated on some frequency (there is a reason IBM avoided it in the first place). It also makes the replacement cost and effort minimal should "events" ever arise since we will have PLENTY of extras produced on a group buy, and barrel replacement is easy.

The spacebar stabilzer stops (full travel) at the next row of barrels. The improvement requires a unique barrel used as both of the stabilizers (2 times total per board). It is substantially more stable than the original pressed plastic inserts in girth by approx 4x on the part that typically breaks, handling stresses much better.

The one downside (which I completely feel is mitigated)... it reduces the spacebar stabilizer width by one half barrel on each side. The previous approach would have been in-between the barrels in either direction. This should have MINIMAL impact on the stability of the bar.

Let me know your thoughts... NO seriously... I want feedback

Spoiler:

Techno Trousers · 30 Apr 2016, 06:56

I really like that idea for the space-bar stabilizer. But if it turns out to be too costly to produce a second mold for those barrels, I've heard good things about the use of neoprene washers over two barrels to act as the stabilizer.

lot_lizard · 30 Apr 2016, 22:24

SO... I have decided to have the prototype printed for fit. I priced it out, and I can be out the door for under $150 for the top/bottom plates, two spacers, and two spacebar clip barrels. In 2+ weeks (would ship 5/13), we will have something to physically look at.

I made a few small tweaks like screw selection and placement, massaging the spacebar clip barrels, other odds and ends... but little else was altered after double checking everything. I am certainly up changing the barrel alignment lock to another design (like the 4704, an X pattern like the model M integrated barrel, dream up our own...), but that can be accomplished pretty easy at any point with zero impact on what I am really trying to test with this print. I have plenty of the XT/Bigfoot barrels laying around, so that was the temporary choice to keep the costs down during design.

Exciting times!!!

shreebles · 30 Apr 2016, 22:35

Exciting indeed... But we would need to provide our own F barrels, correct?

Phenix · 30 Apr 2016, 22:35

really excited to see your results!
But one question (I really do not remember all facts, sorry. Please just say if I am thinking right

):

So this is basically for a barrelframe/bottom plate? (So far I don't have an M to chop off the plate) so I just need to fit in the pcb and the flippers and Im more or less ready, or
(So 150$ for the insards out of metal, alright?)

lot_lizard · 30 Apr 2016, 22:54

shreebles wrote: But we would need to provide our own F barrels, correct?

This really is just a test to prove the concept, but we will be making our own barrel, flippers with springs, etc., on a full blown production run at some point (we officially have the plans for everything but the foam at this point). The goal of this is a full F replacement that can be dropped into the SSK. I personally think we should do the barrel alignment lock differently, but didn't want to pay for the "new way" x 84 barrels just to test the assembly.

Phenix wrote:(So 150$ for the insards out of metal, alright?)

I should have been more specific... This is for resin, but high DPI... not metal though. Think true prototype to prove fit. It will be a fully functional prototype though, just not out of the same material as our end goal. I personally think the metal cost on a larger order will be significantly less than 150 (printing is expensive on large runs)