I'm still relatively new to the forum and spend most of my time here either drooling at the boards of others, or reading up on how to mod, solder, adapt and so on. So be advised that I might have a lot of questions that probably are answered in other threads.
Anyway, so far I've acquired a few various boards – but the post here today is about my Keytronic KB 5150. I'd like to convert it to USB, but my main issue is with the switches – they're awfully scratchy and squeaky. I'd really like to replace the switches and re-use the caps in order to have a clicky, tactile board with that lovely vintage look. (Is this even feasible in the first place?)
The board
IMG_2653.JPG (1.46 MiB) Viewed 2931 times
What I'm curious to know is what's easier for me to do:
- Stick with the original PCB and try to convert the board to USB? (and ditch new switches.. )
Or
- add new switches, wire them myself and make some adapters for the keycaps?
Also, is realistic of me to think I'll be able to make adapters for the keycaps? I have access to a 3D printer and necessary software to model tiny objects.
As far as I can tell, the holes in the mounting plate are ever so slightly larger than regular MX switches, but with that said I only test-fitted with a switch I pried off another board, so the tabs that kept it in place might've been bent.
Images:
Spoiler:
PCB
IMG_2655.JPG (2.11 MiB) Viewed 2931 times
Back of the plate with some switches either partly or completely removed.
IMG_2656.JPG (2.11 MiB) Viewed 2931 times
Keycap and slider. Note how tall the stem inside the cap is.
IMG_20180403_114142362.jpg (238.49 KiB) Viewed 2931 times
Kailh brown switch test fitted. Note the narrow slit. I'm hoping this is only b/c of some bent retainers on the switch itself..
IMG_20180403_114432367.jpg (430.35 KiB) Viewed 2931 times
Keycap closeup. If this was to be converted to say MX I'm guessing that the stem inside has to be trimmed down.
IMG_2657.JPG (1.06 MiB) Viewed 2931 times
Any suggestions are highly appreciated. If there's anything critical I've missed then let me know. I'm new to all of this but have a really strong drive to get this all done and to learn, learn, learn.
Thanks!
Posted: 04 Apr 2018, 11:19
by Anakey
Replacing the switches could work as long as the switches themselves were hot glued into the plate, however then you would need to manually handwire the matrix. As for the keycaps, making some adapters could be possible, however you would need to first find someone to design the adapters for you , then find a good enough printer to print them at the tolerance you would need. Then there is an other problem of the extra height of the end result once you make the adapters the keycaps will sit higher then on the original board and so will probably end up outside the case anyway and will also probably wobble more because of the extra height. My best suggestion would be to leave the keyboard in it's present state for now or if you are really wanting it to work, then the traces can be mapped and then converted for USB, however there is little that could be done to improve the switches as they are aside from replacing the foam that has probably disintegrated by now. Sorry if the above seems so negative, it is just that it is going to be hard to achieve what you want to do by using such an old layout and with non standard sized/stemmed caps. If you really want to try converting/ handwiring etc even building a custom keyboard from scratch those are great goals and should be appreciated that you want to learn these things as you can then build a keyboard in whatever layout you want with keycaps wherever you want. I have recently handwired my own custom board so i can help if you were to go down that route.
The other option if you really want that layout is to get a IBM Model F XT which is the original keyboard that keytronic and many others copied, that has capacitative buckling springs which are both clicky and tactile and it can also be fairly easily converted to USB.
Posted: 05 Apr 2018, 16:26
by snufflecat
Spoiler:
Anakey wrote: Replacing the switches could work as long as the switches themselves were hot glued into the plate, however then you would need to manually handwire the matrix. As for the keycaps, making some adapters could be possible, however you would need to first find someone to design the adapters for you , then find a good enough printer to print them at the tolerance you would need. Then there is an other problem of the extra height of the end result once you make the adapters the keycaps will sit higher then on the original board and so will probably end up outside the case anyway and will also probably wobble more because of the extra height. My best suggestion would be to leave the keyboard in it's present state for now or if you are really wanting it to work, then the traces can be mapped and then converted for USB, however there is little that could be done to improve the switches as they are aside from replacing the foam that has probably disintegrated by now. Sorry if the above seems so negative, it is just that it is going to be hard to achieve what you want to do by using such an old layout and with non standard sized/stemmed caps. If you really want to try converting/ handwiring etc even building a custom keyboard from scratch those are great goals and should be appreciated that you want to learn these things as you can then build a keyboard in whatever layout you want with keycaps wherever you want. I have recently handwired my own custom board so i can help if you were to go down that route.
The other option if you really want that layout is to get a IBM Model F XT which is the original keyboard that keytronic and many others copied, that has capacitative buckling springs which are both clicky and tactile and it can also be fairly easily converted to USB.
Thanks for replying Anakey. I've been working in Fusion 360 and tried to print some prototypes on the 3D-printer I have access to via work. My findings are in accordance to what you've been saying:
Adapter, top and bottom. Total height is about 8mm, which corresponds to the barrel of the original slider inside the old switches.
D42ADC83-5D0B-4CCD-91D3-86AE7D12309C.JPG (450.91 KiB) Viewed 2868 times
Adapter test-fitted on a Kailh brown-switch. The print result wasn't the best, so the adapter is quite fragile, but it does the job as proof of concept.
E14132B2-5D9E-4DB6-A447-FF72B396DFC0.JPG (460.04 KiB) Viewed 2868 times
In order to try to reduce the height of the switch + adapter + keycap I decided to file down the original stem inside the keycap(s). For all I know I need to file away even more, but this is a start.
967BEB31-0EAC-470F-B18C-E36994424191.JPG (455.52 KiB) Viewed 2868 times
Keycaps with adapters. Notice the difference in height.
EC5536F7-80AA-4FE4-BE66-23F1BA02F488.JPG (469.59 KiB) Viewed 2868 times
Switch + adapter + keycap. It's massive.
F50E96C5-7C07-4396-9282-5F4DE2E55553.JPG (410.12 KiB) Viewed 2868 times
Height difference of the "new setup" compared to the original height. I can't have it be this tall.
IMG_2664.JPG (1.34 MiB) Viewed 2868 times
I'm thinking of ways to decrease the height to the point where it'd be indistinguishable from the original. One idea is to try to 3D-print a custom slider that matches the MX-ones, but with a custom "barrel top" type to fit the old keycaps. With that said I'm also playing with the idea that the stem inside the keycaps might have to be filed down even more - but not too much - so that it doesn't become unstable. I guess that ideally I'm looking at a combination of these two ideas, but the former is definitely the most challenging part so far.
Anyway, more testing needs to be done. In the end I'm thinking of maybe having the adapters/stems printed at Shapeways, if I can find a material that's not too expensive and that won't break easily.
Posted: 05 Apr 2018, 17:17
by Anakey
good work so far with the prototype, i think shapeways might be the best bet unless you can find someone with a higher resolution printer. I have just had a thought, but have you considered Alps switches rather then cherry ones? Given that they have a different slider design it would be easier i think to get them to work with the keycaps rather then using an adapter for MX ones.
Posted: 06 Apr 2018, 08:50
by snufflecat
Anakey wrote: good work so far with the prototype, i think shapeways might be the best bet unless you can find someone with a higher resolution printer. I have just had a thought, but have you considered Alps switches rather then cherry ones? Given that they have a different slider design it would be easier i think to get them to work with the keycaps rather then using an adapter for MX ones.
Alps switches is a good suggestion, but as far as I can tell the housing is more rectangle-shaped than square, making the fit in the mounting plate harder. From what I can tell the housing on Matias switches are the same dimensions, so no luck there I'm afraid, unless I either get a new plate laser cut or something.
(Bothced) Alps switch vs. Cherry test fitted in the old plate of the board.
Screen Shot 2018-04-06 at 08.45.45.png (2.66 MiB) Viewed 2838 times
I mainly picked Cherry since it's more easily available (I have a board already in the process of de-soldering), and so the total cost of the board then will be lower. Some quick measurements also tells me that the … "outer border", in lack of a better description, is ever so slightly larger on MX vs. Alps, so I don't know if it matters, but instinctively this tells me there's more "headroom" for a possible adapter to slide in and out of the housing.
Screen Shot 2018-04-06 at 08.32.08.png (614.33 KiB) Viewed 2838 times
Posted: 06 Apr 2018, 09:45
by Anakey
The reason why I suggested Alps is;
Firstly the Alps stem is designed to have the keycap stem going into it, this would then make it easier to mount the keycaps as you could even i think drill out the Alps stem to make the keycap stem fit, whereas on a cherry style switch it is the switch stem that goes inside a keycap.
Secondly the Alps slider is of a fairly basic design and would be easier to even 3d print if you were to go down that route as Alps get the click/tactility from the leafspring. Cherry stems however are rather complex especially as you want clicky switches where the click and tactility comes from the stem itself so a cherry stem will be highly complex to design and expensive to make to the tolerances that would be required.
Posted: 06 Apr 2018, 14:26
by snufflecat
Anakey wrote: The reason why I suggested Alps is;
Firstly the Alps stem is designed to have the keycap stem going into it, this would then make it easier to mount the keycaps as you could even i think drill out the Alps stem to make the keycap stem fit, whereas on a cherry style switch it is the switch stem that goes inside a keycap.
Secondly the Alps slider is of a fairly basic design and would be easier to even 3d print if you were to go down that route as Alps get the click/tactility from the leafspring. Cherry stems however are rather complex especially as you want clicky switches where the click and tactility comes from the stem itself so a cherry stem will be highly complex to design and expensive to make to the tolerances that would be required.
Thanks! Your insights are really valuable. Do you think it would be possible to just file the switch housings on Alps switches down a bit so they would fit in the pre-existing mounting holes, or would this damage the switch too much?
Posted: 06 Apr 2018, 14:36
by Anakey
hmm looking at your picture again i think the alps switches might fit the mounting holes as they are, the plastic clips at the bottom are there just to hold the switch in place, if these are squeezed together perhaps using tweezers then i think the switch should fit through the existing plate, otherwise the plastic can be cut but you would then have to glue the switches into the plate as there would be nothing to hold the switches in place.
)