Let's create the FSSK/FEXT = DONE !!!!

[idollar]

alh84001 has proposed to check if we could create a PCB for the 73X38xx, 09F4230 boards here -> post300979.html#p300979

I think that it is worth reviewing it and check if the effort would pay off.

Does anyone knows where can one find pictures of the internals of one of these ?
Or even better, where to get one at a decent price ?

[Scarpia]

There's one on recycledgoods for around $60 if I remember correctly.

[alh84001]

Quoting wcass here

IIRC, the non-F unsavers are NOT Model M unsavers.
Trying to convert it to an F would be like trying to convert a Logitech to an F.

https://geekhack.org/index.php?topic=7377.0

True. They were not even built by IBM. However, and forgive me if I'm talking out of my a$$ here, but they seem to share the same broad build-style:
1. curved plates
2. top plastic plate
3. back metal plate

I've never seen one live, let alone handled it, so I don't know the feel of these plates, their thickness and so on, so I will let someone else chime in on this.

Now, I realize that FSSK only requires PCB and controller (and as I'm in the middle of getting stuff for the XTant, I sometimes get these two kind of mixed up in my head), but still, just as a thought experiment, would it be possible to convert non-F unsavers to F a-la XTant? Would anything else be needed except top plate and PCB? Could a M122 top plate be adapted for use here?

[alh84001]

There's one on recycledgoods for around $60 if I remember correctly.

http://www.ebay.com/itm/IBM-73X3832-Key ... OSwhcJWQ58~
http://www.ebay.com/itm/IBM-09F4230-TYP ... SwAuNW5ukt

Quite expensive.

[Redmaus]

I have one. Whatever measurements you need I can provide.

I got it so I might be able to use it as an extra case for the unsaver, but I think it's too small.

I still think I am gonna cut my OG case, but if I find an alternative I will use that instead.

[fohat]

I have a Model F "Unsaver" but the sticker on the back says "Model M" and it is dated 1996.

My guess is that it was serviced or refurbished at that time, stickered with a "for-IBM-by-Lexmark" sticker, and resold or sent back to its regular owner.

[Redmaus]

You have an Unsaver Fohat????

Wow nice! You have the regular version with cable, I think mine is the CAD one. I have the CAD Unsaver and the rubberdome one discussed above.

Have you converted it yet?

[fohat]

Have you converted it yet?

I did the ANSI mod and added 3 bolts to firm up the action but I left the original controller and cable in place.

After installing a few internal Teensies, I usually just plug them into an outboard converter these days.

[wcass]

I have one. Whatever measurements you need I can provide.

I got it so I might be able to use it as an extra case for the unsaver, but I think it's too small.

I still think I am gonna cut my OG case, but if I find an alternative I will use that instead.

These 3 would help:

Thickness of the plate assembly (bottom plate - top plate inclusive)
height of switch + cap (bottom plate - top of cap inclusive)
front "lip" of switch (distance from front edge of spacebar cap to front edge of top plate)

[Redmaus]

Should I be using calipers or something to measure?

[idollar]

I have one. Whatever measurements you need I can provide.

I got it so I might be able to use it as an extra case for the unsaver, but I think it's too small.

I still think I am gonna cut my OG case, but if I find an alternative I will use that instead.

These 3 would help:

Thickness of the plate assembly (bottom plate - top plate inclusive)
height of switch + cap (bottom plate - top of cap inclusive)
front "lip" of switch (distance from front edge of spacebar cap to front edge of top plate)

And any information about the hole for the switch
We could try to take all the switches away and replace them with barrels ....

[wcass]

Should I be using calipers or something to measure?

Yeah, that would be best. If you don't already have one, consider getting a cheap one (under $5). We don't need high accuracy.

[idollar]

I have decided to open the FSSK v1.00 users manual: FSSK v1.00 - Installation & Users Manual

[Hypersphere]

Thanks for posting the FSSK v 1.00 User's Manual! This is greatly appreciated! Nice work.

For now, a couple of things occurred to me:

1. It seems like there ought to be a way to get a rubber mat to work in the FSSK, but I have not tried this yet.

2. Is it necessary to heat-bend the PCB, or could it be installed onto the curved plate without pre-bending?

[need]

Thanks for posting the FSSK v 1.00 User's Manual! This is greatly appreciated! Nice work.

For now, a couple of things occurred to me:

1. It seems like there ought to be a way to get a rubber mat to work in the FSSK, but I have not tried this yet.

2. Is it necessary to heat-bend the PCB, or could it be installed onto the curved plate without pre-bending?

from my memory someone did it without bending the PCB

[idollar]

Thanks for posting the FSSK v 1.00 User's Manual! This is greatly appreciated! Nice work.

For now, a couple of things occurred to me:

1. It seems like there ought to be a way to get a rubber mat to work in the FSSK, but I have not tried this yet.

My experience is that with the rubber mat between the flippers and the PCB one will get problems in registering the keys.
This is due to the uneven distribution of pressure across the board

2. Is it necessary to heat-bend the PCB, or could it be installed onto the curved plate without pre-bending?

I have not tried to mount an FSSK without bending the PCB.
The force that the PCB would do, would be to high for the plastic barrel plate. I would not recommend to test the FSSK if the PCB is not curved first.

[idollar]

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 where the PCB is discussed.

[Halvar]

That's brilliant news! I'm in the middle of building mine, but can't do any more today. I was confident that you would have no errors in the design, but with the capacitive sensing, you never know what happens. So it's good to know you built it with success.

[idollar]

That's brilliant news! I'm in the middle of building mine, but can't do any more today. I was confident that you would have no errors in the design, but with the capacitive sensing, you never know what happens. So it's good to know you built it with success.

Thank you Halvar, It is very kind of you. This PCB was redesign to accommodate what I call the XMIT Keys. These are between the ALT and the Control Keys. To use them, one has to cut the case and somehow insert a barrel in the middle.
I think I will not do it. My boards are too good to break them

I am typing with it. Still adjusting the threshold. The autocalibration reports 132. I am using 130 without problems.

[tentator]

as said i$ in the other thread, i will then first start converting my unicomp to model f with your pcb for FEXT. If you already have one I could happily test for you the xmit keys since the unicomp already has the barrels for those keys there
I can receive post in germany so to reduce shipping and time, just let me know.
I think I'll do this first and then at a later point think after learning this process how to do, if at all, something like an FUNSAVER (nice name eh? ) starting from an M122 and chopping the numpad (Actually what could be done is to reuse the case from a rubber dome unsaver and cut only the barrel plate of the m122, the pcb should then be already of correct size).. yes still dreaming, but let's do it one step at a time

tent:wq

[alh84001]

I see I'm not alone in thinking about the rubberdome unsaver. @tentator keep in mind that they are extremely hard to find in Europe

[Abby]

It feels so cool, I look forward to your continuous update.

[wyatt8740]

Here are the very final files that I have sent to "pcbway".
I just hope that they used this version that I submitted via email after the order was made, I added some instructions in the bottom silk layer to clarify that the pads were only in the connector.

The zip contains the deltacad, diptrace and final gerber files.

...where is the zip?

[idollar]

Here are the very final files that I have sent to "pcbway".
I just hope that they used this version that I submitted via email after the order was made, I added some instructions in the bottom silk layer to clarify that the pads were only in the connector.

The zip contains the deltacad, diptrace and final gerber files.

...where is the zip?

You can find it here: viewtopic.php?f=7&t=13551&hilit=fssk

[idollar]

Hello guys,

Long time since I have not come around. I am really busy with little time to hobbies
Today I reordered my desktop resulting in a broken FSSK ! The USB cable was two short so I opened it to replace it with the longer version that I needed. The connector in the xwhatsit controller got lose and … BROKE ! This is the FSSK prototype. Yes, the very first keyboard which demonstrated that we could use F flippers in an M model.

I tried to re-solder it without success. In my desperation, I decided to cut the USB cable, search for the cables and solder them directly in the components. I am posting this in case that you suffer from the same problem one day.

Here is how the cables should be soldered:

Board.jpg (101.57 KiB) Viewed 22247 times

I hope that no-one suffers from the same, but if it is the case, you may try my solution

Cheers

i$

[DMA]

i$ please double-check polarity.
According to https://oshpark.com/shared_projects/eTR858GJ the square pad (with "GND" arrow) is Vbus, and GND is the one directly opposite (top left pin of P2). That USB socket is super-easy to tear off the PCB.

[bahimy]

Hello. Finished with my FEXT recently, and I have a peculiar problem with an xwhatsit that I bought from Ellipse. I connected LED panel to the controller and the problem is that when I activate a lock the actual lock light switches off, and when I deactivate the lock the light turns on. I did not test it with the original firmware (v0.9.0), but with v0.9.1 and v0.9.2 the behaviour is the same.

Did someone faced something similar? I guess if it was a physical connection problem it would not be working at all, but it works only in strange manner, so it should be problem with the firmware. On the other hand there is not a word about such a problem on the Internet, and I do not believe no one uses locklights. Ellipse himself says LEDs should work flawlessly, because it historically did. So, I think the problem is in the firmware, but reality just counters my guess.

Please tell me how did it work for you?

[NathanA]

Sorry for the thread necro...just a (hopefully) quick couple of questions for others specifically who have done (specifically) an FEXT install in the past:

Has anybody installed an FEXT into a Gen 4 Model M chassis (late 90s / Lexmark / Unicomp Classic / the one where the controller itself is what the lock lights are mounted to, over the numpad)? Having not looked at the insides of a Gen 4 enclosure vs. older Ms side-by-side, I am mostly curious about the mounting situation for the replacement (xwhatsit) controller board. Obvs. there's no way to fit at least any of the existing xwhatsit designs in the area over the numpad, but with the controller having been relocated vs. older Ms, is there a convenient place to secure the xwhatsit controller over in the more "traditional" area?

Also, the recommendation seems to be to do a screw mod to your M instead of a bolt mod. While I understand that a screw mod is much easier to do and is pretty much just as good, it's unclear if the reason for specifying screw mod is merely on account of that, or if it's because a bolt mod would actually interfere with installation of the PCB in between the barrel frame and the backplate. So, what if I had an M that had already been bolt-modded? Possible to FEXT it or nah?

Thanks!

[dcopellino]

Yet another model FEXT implementation,
(aka IBM model MotherFucker*)

Assuming that “Let's create the FSSK/FEXT = DONE !!!!” should never be considered a necro but rather an evergreen common thread, I’d like to share here my pic log presenting my personal FEXT implementation according to i$'s original post to whom I give credit and gratitude, along with all the other thread contributors.

20221215_171250.jpg (1.75 MiB) Viewed 16569 times

I don't know exactly why I embarked on this laborious journey that began about a year ago. Irony of fate, I started just a while after I got rid of a couple of extra Model Ms on eBay I had previously bought just because of their affordable price. Even though I’d already given a glimpse to the original i$’s post Let's create the FSSK/FEXT = DONE !!!!, At first, I didn't take it as really feasible. It must have been the awareness gained over years of accumulation and use of IBM models that finally pushed me to consider the thing in a real way. Among all the reasons that led me to assemble a model M/F I’d put in the foreground the need to ideally try to perk up a product like the Model M that is commonly perceived by click-clack lovers as nearly perfect. A feeling that is truly realized only after trying Model Fs following being a fan of Model Ms. Btw, most likely I started this adventure for having some talking during the long pool playtime with the DT user Man from Mars (thanks mate), or I'd rather tell the truth: I liked the idea of playing to be Ellipse in the middle of the QA process on his model F repros.
Ok! now that I've made peace with myself, I can start my pic log, highlighting those moments of the process that I consider by my humble point of view particularly crucial or worthy of a little more attention, just in case some other madman wants to follow me on a similar journey.
Disclaimer: the following pics were taken with my smartphone. So apologise me in advance for the poor quality, bad light exposure, etc.
Furthermore, be aware that I just documented my FEXT implementation, focusing on issues that I encountered on my way and how I troubleshot them, without any further claim.

The FEXT 1.0B PCBs

Spoiler:

thumbnail_20221212_181906.jpg (263.51 KiB) Viewed 16569 times

PCBs (0.6mm thick) arrived very quickly from JLCPCB’s China factory, even though with a manufacturing defect. Keep reading for more details about it.

Trying to bend the PCB

Spoiler:

20220827_111319b.jpg (4.29 MiB) Viewed 16569 times

Even if you too have a Greek column at home, you don’t really need to do this for bending the PCB. Being it thick only 0.6 mm, the screw mod will be sufficient to bend it properly without any problem.

Meanwhile....

Spoiler:

20220901_102252.jpg (3.19 MiB) Viewed 16569 times

Meanwhile, I had some fun by customising an old and quite yellow dust cover

Working on the barrel frame

Spoiler:

20220907_170237.jpg (3.22 MiB) Viewed 16569 times

drilling into the Model M's barrel frame to house the Win key and the Fn key on the right

20221020_150957b.jpg (3.67 MiB) Viewed 16569 times

with Dremel I slightly lowered the thickness of the barrel frame in order to better align the two model F barrels to the others

20220907_164729o.jpg (2.78 MiB) Viewed 16569 times

beware of debris while using the Dremel. they are your enemies

Model M top case trimmings

Spoiler:

20221012_100157.jpg (3.08 MiB) Viewed 16569 times

cutting off the extra pieces from the front part of the Model M case

Soldering the controller I: First “bad” attempts

Spoiler:

20220908_180446.jpg (3.38 MiB) Viewed 16569 times

20220908_180457.jpg (2.77 MiB) Viewed 16569 times

soldering has always been my gripe. I’m not good at all. So I hired someone to do the job for me, but……something went wrong.
Warning: also the choice of wires can be very tricky. Those illustrated in the pic have proven to be completely flawed. They were poor in copper and very rich in heavy plastic. Also the length of the wires was bad and misunderstood by the fellow. So I had to start all over again.
in the pic the SMD universal Model F controller rear part, very kindly donated to me by Darkcruix (thanks again mate)

Soldering the controller II: on the right way.

Spoiler:

20221202_203047.jpg (3.32 MiB) Viewed 16569 times

A new fresh start seems to get things better, but the issues were not over…..

Checking the electrical signal. Capacitance....ooh my gosh!

Spoiler:

keymap.JPG (257.33 KiB) Viewed 16569 times

Since the first capacitance test sessions, the PCB had shown a huge electrical signal problem on all the pads belonging to the first row. At first I didn't want to believe my eyes. Thinking about a problem involving aspects related to capacitance, I asked the DT Arkku user for help, and he was very friendly and kind in providing his support, offering me several pre-complied firmwares, derived from Pandrew's QMK fork which adopt, through his source code reprogramming, significant improvements in terms of performance as well as adaptation to signal tolerance. No luck. The problem was to be found elsewhere.

My big big BUG

Spoiler:

20221202_203055.jpg (3.28 MiB) Viewed 16569 times

Due to some manufacturing defects, partly due to a borderline project design too, a hole interrupted the first row circuit path of the PCB. In the pic, my saint welder made the trick and the miracle of resuming the circuit integrity.

Starting with the assembling

Spoiler:

20221126_105324x.jpg (2.3 MiB) Viewed 16569 times

Having overcome the electrical issues, I dedicated myself to the conception of this system for the assembly of the mechanical parts. A few toothpicks guided the insertion of the perforated pcb onto the metal back plate so that a screw mod can be carried out more easily. Some bi-adhesive strips relieved the initial tension of the screws as they are fastened in.
NB. I also coated the metal plate with Kapton tape in order to insulate and minimize electrostatic charges, as suggested by i$’s original thread.

20221126_105926z.jpg (3.7 MiB) Viewed 16569 times

The presence of toothpicks, inserted into the back plate lying on a story foam sheet, acts as a guide to align the holes of metal plate with the PCB’s and the incoming barrel frame (next documented step).

Springs coming from an old Model XT mix with some from Ellipse

Spoiler:

20221020_150906.jpg (3.41 MiB) Viewed 16569 times

Springs coming from an old Model XT mixed with some from Ellipse

due to thread limitations in the number of uploadable images, the post continues below.....

[dcopellino]

.......continues from the post above

Fastening and unfastening all layers together or tuning a musical instrument

Spoiler:

20221020_150632.jpg (920.1 KiB) Viewed 16561 times

I had to devise this system, instead of the classic pair of books placed on the sides of the metal plate, in order to get the right tension throughout the surface during the process of inserting the screws. The wood sticks (2/3 horizontal and 1 vertical, the big pencil) protruded just enough to prevent the springs from touching the desk, letting the hammers lose their correct position in the barrel frame.
I know, I completely missed the building of the right tool, lost somewhere among DTs posts.

20221022_130847.jpg (1.79 MiB) Viewed 16561 times

I’ve done several MODs on Model Ms, both the bolt mod and the screw mods, even on a Model M122 but this time was different, very different due to the presence of a PCB that changed my references in terms of thickness and curvature tolerance of the parts involved. I had to pay more attention to filing the edges around the holes. Any inaccuracy in the alignment of the parts or excessive screwing force inevitably caused problems in obtaining the correct key feeling and relative click-clack sound. I had to close and re-open the layers at least 6 or 7 times in order to get the perfect combination. It was nothing more or less like tuning a musical instrument, in the true sense of the word.

Making extra room to house the controller into the bottom case. Look at my birth date....

Spoiler:

20221215_155716.jpg (2.94 MiB) Viewed 16561 times

Finally, the case closes. I had to cut the piece of plastic swown in the pic if I wanted the SMD controller with the attached microUSB cable to properly fit beneath the keyboard bed, avoiding to damage the Micro Pro the usb connector (oops).

Keep testing and other extra final steps.

Spoiler:

20221215_162315.jpg (3.07 MiB) Viewed 16561 times

Some final tests with Pandrew’s USB utils before closing the case. On the right side LED lights work even though in a weird way, due to some incompatibilities. It seems that the SMD controller supports natively only the IBM AT leds.

20221215_171542.jpg (2.5 MiB) Viewed 16561 times

all's well that ends well