deskthority

@lot_lizard - You (we) should start thinking on creating a new thread with this. In my experience, marketing of the project is key to success. Your work is hidden behind the manual.
I do not have any problem to continue with this thread, but a good title will give you some more visibility. What do you think ?

idollar wrote: @lot_lizard - You (we) should start thinking on creating a new thread

...
Totally agree. I have felt like this belongs outside of the manual for a while now too, plus it has been making the manual messy with harebrained ideas vs focusing on installation. I'll make one later today (add a link here) and consolidate so our vision is clear

EDIT: Adding "said link" above:
workshop-f7/remodeling-the-model-m-t13796.html

lot_lizard wrote:

idollar wrote: @lot_lizard - You (we) should start thinking on creating a new thread

...
Totally agree. I have felt like this belongs outside of the manual for a while now too, plus it has been making the manual messy with harebrained ideas vs focusing on installation. I'll make one later today (add a link here) and consolidate so our vision is clear

Let me propose that you start with a post in which you describe the background, the progress and the intended target.
I would also propose that you post after a couple of entries as a placeholder for future information. It is handy to have these placeholders to organise the information in a single place (at the beginning) of the thread. My experience with other threads is that looking at the many pages that are created later with the discussions discourage people to read it. Instead a single set of post with the summarise ease the understanding of the content.

... just ideas that you do not need to follow

I have a working FEXT prototype without any problem !!!

As it should be, I am typing this post with the first of the FEXT, based on the PCB v1.00b with the XMIT keys (which I have not tested as I do not feel like breaking the case that I have)

I will post this in various threads in which this PCB is discussed.

After some more grounding every key except the "T" work. I used also a shorter cable.

hypx, out of curiosity, what is your xwhatsit voltage threshold level, with your setup now?

FYI - the first post has been updated with the source files of the FEXT v1.0b

Such as amazing job guy... really. Looking at the dip always fascinates me. Just let us know when you can what you did for the LEDs. That one has me stumped on the membrane ribbon cable

would it be enough for an meanwhile case if I just place it on an 'curved wood-stand'?
Or is it netter to fix it in some way to the backplate of it? Did someone build/tried this earlier?

Just wanted to say that I'm giving my FSSK another go in day to day use. Since last time (when I complained about some grounding issues) I tried grounding the Space bar and wrapping some aluminum foil around the front edge of the plate assembly to try to get rid of any excess static charge that was building up from typing with a neoprene wrist rest.

So far, so good. I wrote this entire post on the FSSK with no ghost keys and no missed keys. IIRC my voltage threshold level is set to something like 140.

I went to a local keyboard meetup last Friday and brought the SSK with me. I had to explain what the FSSK was and why it was special. Most folks there didn't really know what a Model F was, but for those who did, they were floored that this was even possible.

Thanks again idollar for doing such a phenomenal job with this FSSK PCB.

Is there anything that prevents using a teensy instead of the xwhatsit controller? I couldn't come up with anything that could prove otherwise...it's the same atmel controller and using soarers controller it should work (if my bigfoot does).

The bigfoot is different because Soarer's controller is using the original IBM capsense chips to read the capacitive switches. You don't have these here.

If you wanted to use a teensy you would need additional hardware, like voltage comparators similar to the ones that xwhatsit used on his controller. You would have to adapt e.g xwhatsit's software to use that.

If DMA is successful in getting his Cypress-based controller to work, that might be a very nice alternative.

I see...I read through that thread, exceptional work by DMA. Sometimes it would be cool and useful if some old IBM keyboard engineers could lend us a hand...it's just so complicated and time consuming to merge old and new tech in a successful way especially considering it's "just a hobby" for majority of us

PS. Where did Soarer go, again?

any left over pcb's?

Here it is - Fake-SSK in all it's glorious incompleteness


There are a couple of issues:
- when cutting the case I was so into it, that when I tested the assembly and it didn't fit, I decided to cut the case a bit more and it didn't occur to me that I should widen the assembly hole (stupid) - that's why it's narrower on the right side than it should be; hopefully that won't affect the typing
- no isolation on the back plate
- no grounding
- one column is constantly registering - that is down to a bad solder joint or grounding
- bottom row is not registering - i did a reassembly and it did work immediately after, but it's not working again for some reason

And of course, half the springs are missing. I'll wait with fixing the issues when the springs arrive, which should be in a week or two, and then do reassembly. Hopefully, there'll be some more pics.

Hey i$, I'm working on soldering the FSSK PCB to the xwhatsit controller, however the amount of wires from both ends changed:


Suddenly on the xwhatsit, there are 11 wires on the left side and 19 on the right side, and 2 ground wires merged somewhere between all of that. Problem is, I'm not sure what changed-- do you think you could elaborate? I'm guessing you merged the grounding wires in the midst of all of that somewhere.
EDIT: I'm going to assume that pins 11 and 12 (on the xwhatsit) are the ones that don't connect to the FSSK PCB.

hello micrex22,

It should be a problem with the quality of the pictures because the cable that I used was 30 wires.
In any case, it is not important if you use the same cable type. The important bit is that the rows and columns are separated by ground.

Please check this post again. I think that the pictures explain better what I mean. If you still have questions, do not hesitate to ask again.

alh84001 wrote: Here it is - Fake-SSK in all it's glorious incompleteness

Aaaaand, it's complete. Well 98% complete. Two issues remaining/emerged:
- I'm still missing some 4 springs, and have a couple of bad ones, so some F keys are not working - it will have to wait till MFSSK ships so I get some extra springs
- all keys register at 140 threshold when pressed except "C"; no keys register when not pressed until 134 threshold; "C" starts registering when pressed at 127, at which point a bunch of other keys are registering when not pressed; I'll try reassembling everything, and testing the PCB pads by themselves, but it will have to wait for the next weekend

shreebles wrote: Amazing stuff, I$. Thank you for the tutorials, your idea and all the work you did on this project.

After postponing it for so long, it was finally FSSK building time today

The attachment DSC_0905.jpg is no longer available

Following your advice from the other thread, I did not use the mat.

I have the IDE ribbon cable ready, all it needs now is a controller.

Here I am a year later, and finally got my Xwhatsit. It was complicated, but Ellipse delivered in the end.
I got really far on the weekend, and though it took a while to get there, almost all keys work now!
I was really happy, almost a bit surprised I got it to work so well on the first day.

- first, I tried the "elegant" solution, reusing the mat BEHIND the PCB as isolation between it and the plate. The result was a mess, almost all keys registered simultaneously or not at all, depending on threshold

- second try was isolating the plate with gaffer's tape exactly as idollar showed in his manual. This was a huge success, the difference between "doesn't work at all" and "kinda works"

- last big success was adding the grounding screw and wire between the PCB, plate and controller as can be seen in the picture. I had kinda forgotten about this grounding the first time and adding this improved the clarity of the signal massively. Now, at about 127 voltage threshold, most keys register fine. I can get 99% keys to function by lowering the threshold to 124, but that adds too much noise so I end up with way too many wrong actuations.
The thick cord of little cable strands on the right does nothing to improve grounding. I had kept them as part of the IDE ribbon to stay as close to idollar's manual as possible, but it makes no difference whether I connect it to the grounding pin on the controller, or not.

Now, I'd be really thankful for some comments or advice from people more experienced than me, as I do not know how to proceed to make all keys work.

S, F9 and LCTRL start working at 124 threshold, but that introduces too many incorrect keypresses.
At 127, those above do not work, but most other keys do. However, the escape key seems to fire randomly. Pressing Tilde always actuates Escape as well, but not the other way around.
One or two keys never actuate, but I hope that changing flippers on those keys could fix it.

As for the other keys, any kind of advice is welcome!

Either way, this was an amazing project so far, and even if I can't get it to work 100%, it's really cool to have Model F tech in a Model M. Kudos and and many thanks to idollar and others who helped make this a reality!

Use Kapton tape for isolation. It's thinner and more consistent. That's what I did. Also I had to ground my Space bar.

Thanks for your input XMIT!

Never heard of Kapton tape. Is that it?

Space bar already grounded. At least I hope so, used the little grounding cable that came with my european model M. Looks like this

There are three holes on the controller. Are all three of them grounding pins? Do I need to increase grounding to the controller somehow? As it stands the thick cable is useless, unfortunately...

Yep, that's Kapton. Really any polyimide tape will work. Super thin, heat tolerant, and a really good insulator.

Yeah, your likely Greenock M would have an ABS Space bar and already be grounded.

I perhaps over-did it but I just had one extra ground wire connecting the metal plate to one of the grounding holes.

This is all a game of consistency and killing any little voltage differences. If bolt modding make sure all the bolts are at about the same tension, and that the board is the same distance from all the flippers!

FWIW, this is one of the hardest keyboard projects I've ever done. Granted, I haven't restored a Beamspring or a Selectric. It takes patience and multiple tries to get this right.

IT WORKS!! ALL KEYS WORK! This is so amazing

Thank you all very much, especially the man himself idollar!!
Detailed report coming soon!

shreebles wrote:IT WORKS!! ALL KEYS WORK! This is so amazing

Thank you all very much, especially the man himself idollar!!
Detailed report coming soon!

Great, congrats!!
(and also thanks again that you were so enthusiastic about the FSSK at our meeting, you drove me to build one )

shreebles wrote: IT WORKS!! ALL KEYS WORK! This is so amazing

Thank you all very much, especially the man himself idollar!!
Detailed report coming soon!

Hooray! I hope I helped here, at least a little bit.

Nice work shreebles!
That reminds me, I *still* have to fix my "C" key which requires lower threshold than all the other keys. Sigh.

XMIT wrote: Hooray! I hope I helped here, at least a little bit.

Definitely, you gave good advice and I'm using Kapton now.

So, how did this go? Both good and bad, in the end.

Kapton tape works well. It's really thin, though not the easiest to apply. After applying the Kapton, making sure to go around the corners to avoid the PCB getting contact with the plate at all, I had to make sure to remove the tape where grounding was applied, i.e. at the top where the grounding screw is and at the bottom for the grounded space bar.
Kapton tape improved the treshold at which I was able to use the board without errors... by 1. Not much, but it sure didn't hurt. I also found that key recognition improved a bit over time as the PCB settled and conformed better with the rounding of the plate. Possibly grounding was also improved by using two wires, one on the PCB and one on the plate, which were connected with the screw.
What also really helped was adding the rubber mat back to the assembly, thus increasing thickness and overall tightness of the whole plate/PCB/plastic sandwich. I added the mat to the back of the PCB where it would provide only electrical, but not acoustic, isolation.

After that,I was able to find a threshold at which all keys registered, which is fantastic!
Typing feel was mostly uniform, but there were some outliers with a heavier or scratchy feel, possibly due to the barrel plate not being 100% perfect.

However, as I mentioned a few pages back the sound of this board is ridiculous. I did not expect it to be so loud, it is so loud that it distracts me massively while typing. That said, I have been using quiet boards in the last few weeks for use at work so my sense of volume is currently skewed towards quieter keyboards.

So, being in possession of a green prototype PCB from idollar's first run, I decided to try the option of including the mat in front of the PCB, between the PCB and flippers, similar to the original model M design. I hoped to reduce sound by this, and, in some way, succeeded.

I had to lower the electrical treshold significantly (by around 10 or 15) to make all keys work, and at that treshold, there were many incorrect keystrokes registered. Most of them were around the edge of the board. Currently the controller is set to ignore all F-keys and arrow keys, as I have frequent misfires on those switches. I'm currently using the board as a 60% with layers to achieve F and arrow functions. That way, the main alphas all work and register correctly, but the utility of the Tenkeyless sized board is much reduced.
Sound is also much reduced, but not in such a way that I would call it a quiet board.

Unfortunately, typing feel is much worse with the mat, heavier and less smooth overall.
I think I will put the mat behind the PCB again this weekend, just to get the full functionality of this board back, and then decide what to do with it.

My recommendation is to use the mat if your assembly is too loose, behind the PCB. Using it between flippers and PCB is not recommended with the final production red PCBs anyway, but I guess if someone were to try they would have to have a near perfect bolt-modded barrel plate, with perfect uniform pressure across the board, to achieve full keystroke recognition even on the edges of the PCB. This was a problem with my screw-modded version.

I've started floss-modding the springs, and while it will not reduce the DEAFENING volume of the keyboard itself, it will at least get rid of the ping.
In the end, this was a really cool, and fun project. You may lose some sleep troubleshooting the beast but when it works it makes you so happy.

Still, in the end the Model F hype left me scratching my head. I cannot see myself using this, or something like it, daily, as I get really distracted by the high volume.

Either way thanks to all people involved for making this happen!

I'm typing thi on my FSSK, and I'll leave th text as it is so yo ucan see the isue I'm havin. I have a vltage threshold rang between 135 and 141 wher al the keys work when pressing tem one by one, but when I start typing quickly letters start droping, quite often when there are two same letters in sequence (at least that's true for ll pp and maybe some ohes). I actually don't hink my back plate is insulated, (but I assembld it so long ago that I don't rememberright now. Is that what cold be causing it, or is there somthng else? I'll also hav to chck if I added groundin as ewl. Also, order of letters is sometms wrong.

`1234567890-=qwertyuiop[]\asdfghjkl;'zxcvbnm,./

Thi was pressing alphas one at a time and they all seem tobe working. I don't know if there are some key combos that are actualy moeraffected than others like in th word 'raly' or if that's just the rsult of how I tpe those words.


And yes, the board is loud, but I actualy don't mind. It's not distracting tome, and it actually drowns out the background noie, It's definitely not office friendly thouh.

shreebles wrote: IT WORKS!! ALL KEYS WORK! This is so amazing

Thank you all very much, especially the man himself idollar!!
Detailed report coming soon!

Thanks

To those that have wondered where I was .... I cound not been active lately. Too many things to do.

shreebles wrote: IT WORKS!! ALL KEYS WORK! This is so amazing

Thank you all very much, especially the man himself idollar!!
Detailed report coming soon!

I just purchased some Kapton tape. I have to give it a try.