Beamspring Factory process - How to

SneakyRobb

16 May 2019, 16:24

Hi,

I am making a thread to document my construction of new beamspring keyboards/components. I will collect designs, files and processes here. I wish to be able to make all the components of beamspring keyboards because why not. Thus I will do this.

I have worked with wcass and DMA bouncing ideas off of them as well as them helping in many ways. Obviously they have both been a big help and I couldn't have gotten this far without them.

Main imgur collection here. I find imgur difficult to use. So duplicates etc will occur.
https://imgur.com/gallery/x0B4aZV
https://imgur.com/a/mrZbist

So far I have managed to produce a cherry mx conversion for my 3278 into a 60% Ansi Keyboard.
uVy1nQq.jpg
uVy1nQq.jpg (209.28 KiB) Viewed 560 times


(I have DSA "ice" keycaps now to replace the space and modifiers I borrowed from a quckfire. I know they look bad)0

Most keys do not actually need the stabilizers as beamsprings almost never bind. I have been adjusting the 3d printed stems in order to get the stabilizers to work well. https://imgur.com/bBQUrR5 https://imgur.com/aCJ01Hi

Costar stabilizer bent bars hit the return spring and are not the best.

PCB/Metal cut gerbers
These are my first pcbs. As such they do not follow any particular good design rules for pcbs. They should 'work' but they aren't elegant, or really well thought out. I can't guarantee they will work, so if you order one from a pcb maker I suggest you double check it. Obviously I have uploaded them so you can change them etc. I have opted so far to use 2 layer PCBs because they are easier to make. They use the DMA commonsense controller on an external board because it is easier to test with an external controller. These are works in progress.

From what Ive found and learned from wcass.
https://imgur.com/Yy11gH3
https://imgur.com/IwOjOw8

This style of PCB is possibly the "first" beamspring style pcb. It has many vias which were expensive to produce. The sensing pads are both on the front of the pcb, the "rows" as they were travel using vias and connected to a trace on the back. There is also a massive amount of grounding copper rows on the back.

I theorize that the "front and back" style of PCBs that are seen in later Beamspring and then Model F keyboards are a cost saving measure. As such I opted to produce a PCB which is more akin to the original and hopefully has very strong signalling.

I made an MPCNC machine to make new metal plates. I have only cut some balsa wood ones so far as I just made the machine. I will eventually attempt to use aluminium.

For the plates I have found that having a very thin 3d printed card with the "tabs" to align the keys, makes it easier to machine the plate as you need no tabs on the plate.

ansi board - 60% board
Uses mostly original keys except for space modifiers etc using cherry mx stems for those.
https://imgur.com/uVy1nQq

https://imgur.com/eh1pliL
https://imgur.com/XhMJYPZ

https://imgur.com/gWW6Otx
https://imgur.com/PCRSdSG
Robb 60% pcb and cut.zip
(30.98 KiB) Downloaded 3 times

Model M style
I haven't actually ordered this made yet but will be doing so soon.

https://imgur.com/M79FRCu
https://imgur.com/V1yFxuf
https://imgur.com/Zst6hOx
https://imgur.com/rW9FZSH
Full size board wip.zip
(133.86 KiB) Downloaded 3 times
Conversions so far have been straight forward. As you need a plate and a pcb along with the DMA controller. You can then use original switches/caps etc.

I wish to be able to DIY/DIWHY produce all components of beamsprings so I can fill out my Model M variant.
For the production of new components I am less developed.

Return Spring
Currently I am working to perfect a 3-piece 3d printed plastic stick screw. I have music wire of 0.025 inch thickness.
I have been able to wrap the wire around the stick to make new return springs. This is somewhat time consuming but it works.
The stick has 3 lengthwise pieces. The 2 outer ones have the screw guide and the middle one doesn't. I will upload this file soon.

What you do is wrap the wire around and then remove the middle stick to remove the other pieces, and there you go new return spring.

Here is a basic idea of the mechanism. It is a guide to get the winding right, and then it disassembles.
https://imgur.com/a/gWKzwPS

I will update on the progress of these springs this weekend.


*This system didn't work at all. I have been able to make some decent springs by using a makeshift brake out of 2 metal plates, a bolt/nut and wood with music wire. Must continue developing.

Foot
For the switch plate foot, black carbon 3d printed plastic seems to work somewhat. It is obviously capacitive. It doesn't matter what you make the barrel assembly out of.
I will be uploading some OBJ/STL files later tonight. The main consideration for this part is whether or not it should have a stem on it. The reason this is a question is that the original flyplate was press fit onto the foot with a metal backward fork preventing removal. I would rather use a small screw as if I am going to make flyplates, I would rather just punch out one hole in the middle with maybe 2 position post circles.

The stem can be seen in this gallery which will upload fully soon.
https://imgur.com/gallery/Jx1lnxF

Flyplate
This piece will have to be stamped somehow. I believe the easiest way would be to order a circuluar disc puncher. As the metal is 0.035 inches thick. I would then bend it on a fixture that is 3d printed.

I have been in contact with this etsy seller of custom punches. They have quoted me between 150-350 usd. I am attempting to make one of these out of aluminium currently. Even if the aluminium punch only lasts a few punches, I think it would still be cheaper to keep remaking these.
https://www.etsy.com/ca/listing/6248648 ... rket&frs=1

Beamspring
For the beamspring itself, the steel is 0.015 inches thick spring steel and you can cut it with scissors. (Don't ask me how I know this.)

The beamspring itself "should" be reasonably easy to make.......... with a press as it is just a shape with a hole in the middle. Just like the flyplate, I am currently trying to source/perfect my punch press in order to make larger quantities of new springs.

Barrel Assembly
I will upload an obj/stl of this a bit later. It is a pretty straight-forward geometric object that can be 3d printed.


Keycaps
I have done a lot of research about the beamspring keycaps. They are cool. I believe that were you to try to make a production beamspring board today, the largest single expense would be the keycaps and it would quite likely take up a massive majority of the cost. As such I am not attempting to do this at this time aside from experiments.


Basic work in progress 3d models for plastic parts.
https://www.thingiverse.com/thing:3641754

I will continue to update this post in the future with files etc. Let me know if you have any questions or input.
Last edited by SneakyRobb on 21 May 2019, 04:33, edited 12 times in total.

John Doe

16 May 2019, 16:34

Splendid job. Hope to see more process going on.

By the way, if there are extra original beam spring caps to sell, PM me.😬😬😬

User avatar
Monster-Toys

16 May 2019, 18:23

Very cool project! (thx for making this thread :) )

"The beamspring itself is reasonably easy to make with a press as it is just a shape with a hole in the middle. I am currently trying to perfect my punch press in order to make larger quantities of new springs."

Especially this sounds great. :) (count me in for the first ordering round^^).

Did you think of converting to something else then capacitive sensing? I try to figure out ways to convert my switches to "normal" switches. CS is (at least in my beamspring) quite a diva and disassamble the beamspring every other day and clean it to avoid severel "hickups" is annoying.
I have several ideas for it but did not try them yet (the easiest beeing to tape a thin selfmade foam and foil pad to the flyplate. It requires another plate between switches and pcb, tho).
Doing it with custom housings could be more "elegant" (conductive plate on top of the flyplate and contacts where the flyplate rests when the switch is pressed (the two cylinders in the original) + two legs for through hole soldering).
But maybe i am just crazy. :oops:

SneakyRobb

16 May 2019, 18:53

Monster-Toys wrote:
16 May 2019, 18:23
Very cool project! (thx for making this thread :) )

"The beamspring itself is reasonably easy to make with a press as it is just a shape with a hole in the middle. I am currently trying to perfect my punch press in order to make larger quantities of new springs."

Especially this sounds great. :) (count me in for the first ordering round^^).

Did you think of converting to something else then capacitive sensing? I try to figure out ways to convert my switches to "normal" switches. CS is (at least in my beamspring) quite a diva and disassamble the beamspring every other day and clean it to avoid severel "hickups" is annoying.
I have several ideas for it but did not try them yet (the easiest beeing to tape a thin selfmade foam and foil pad to the flyplate. It requires another plate between switches and pcb, tho).
Doing it with custom housings could be more "elegant" (conductive plate on top of the flyplate and contacts where the flyplate rests when the switch is pressed (the two cylinders in the original) + two legs for through hole soldering).
But maybe i am just crazy. :oops:
Hi thanks. Again I must credit heavily wcass and DMA as well as red for his videos. Shoulders of giants.

I believe in other projects as well as the DMA commonsense thread people have looked into other sensing methods, pad and foot sizes and the such. You can have some variation I believe, but I am not sure. I am just working in the physical rearranging of the keys/pcb as well as production of switch pieces. There are a bunch of threads that I have seen that I have used as reference. So my board wasn't developed in a vacuum by a longshot.

There is the fext project for converting the model m to use model f components as well.

I don't see why you couldn't make a foil and foam system. It would require the work though. I have not had a problem with the CS system itself. I had my 3278 setup with no dust shield for well over 2 years with no malfunction both with both xwhatsit and dma. So I would strongly suggest trying to figure out why your system is having the issues that it is.

There is definitely room for elegance here. Wcass designed a cool case for the "HHKB" variant he designed.

My sandwhich design was more to just get a working system up and running using the commonsense. It is quite... rough.

DMA

19 May 2019, 23:05

Sorry for hijacking the thread, SneakyRobb.
Monster-Toys wrote:
16 May 2019, 18:23
CS is (at least in my beamspring) quite a diva and disassamble the beamspring every other day and clean it to avoid severel "hickups" is annoying.
What are your symptoms? Keys not registering, repeats, something else?

What are your thresholds? Must be about quarter to halfway between "not pressed" and "pressed" levels.

-- check/fix thresholds before going below --

Do you have connection to the metal parts of the keyboard to the ground?
If you do - please PM me photos of the kit and all wires between it and the keyboard visible.

SneakyRobb

20 May 2019, 04:31

DMA wrote:
19 May 2019, 23:05
Sorry for hijacking the thread, SneakyRobb.
Monster-Toys wrote:
16 May 2019, 18:23
CS is (at least in my beamspring) quite a diva and disassamble the beamspring every other day and clean it to avoid severel "hickups" is annoying.
What are your symptoms? Keys not registering, repeats, something else?

What are your thresholds? Must be about quarter to halfway between "not pressed" and "pressed" levels.

-- check/fix thresholds before going below --

Do you have connection to the metal parts of the keyboard to the ground?
If you do - please PM me photos of the kit and all wires between it and the keyboard visible.
Hi, No worries.

Very half baked update.

I will make this process easier in the near future. This is my tinkering update.

Spring progress. This is for the return springs. Not the beamsprings.
https://imgur.com/gallery/fMzgf8r

So I spent today trying to make return springs. I had some success.. but will need more work.

I had a concept of a 3d printed piece of plastic and it basically didn't work Which is good.

For this you need to use music wire. And a long 1/2 inch hex bolt. A piece of wood and some other hardware.

0.025inch thick music wire. You can use any brand/source>
Example https://www.amazon.com/PRECISION-METALS ... way&sr=8-5

A piece of wood to attach everything to.


This is basically what I modeled my process after.
https://www.youtube.com/watch?v=lhQBz6q ... e=youtu.be

I managed to make some smaller springs and will be able to soon make full size springs. I forgot to get a 1/2 nut so made some smaller ones. Most of them were average/poor. I showed the best one.

Basically. you need a piece of wood as a base. You will use 2 wingnuts on bolts to hold 2 plates of metal together. You sandwhich the music wire between the pieces of metal to compress it and not allow it to back up. It lets you keep up the tension.

You thread the wire through the nut and screw it down on the bolt. You then rotate the bolt and the wire takes on the spring shape. Then when done cut the wire and unthread it.

As you can see in the images I managed to make a smaller spring. It has decent spring characteristics and the newer 1/2 inch ones will function perfectly fine.

SneakyRobb

20 May 2019, 22:20

Return springs part 2.

Alright so I think this would be easier with a lathe. I do not have a lathe. So for now I will attempt to make them at home. The 1/2inch bolt I used made springs that were too big. I have read in the machinists handbook that I will need to use a smaller diameter screw in order to make springs of 1/2 inch size.

Rolling spring wire onto a 1/2 inch bolt will make a spring that is closer to 3/4 inch across.

I will also have to shave down some of the bolt in order to thread them on with the 3 and 4 bands that start the original screws.

This can be seen in this photo album.

https://imgur.com/gallery/NXQMfXR

I hope to be able to make useable springs in the next 1-3 weeks. The setup is still quite rough but I am making some progress as you can see.

SneakyRobb

20 May 2019, 23:09

Reorganized the original post and added some files.

SneakyRobb

21 May 2019, 02:32

John Doe wrote:
16 May 2019, 16:34
Splendid job. Hope to see more process going on.

By the way, if there are extra original beam spring caps to sell, PM me.😬😬😬

Thanks for your words. I am a bit covetous of my keycaps.. perhaps in the future any other element of the board I could provide you or provide the knowledge to be constructed. !!

User avatar
Monster-Toys

21 May 2019, 09:25

SneakyRobb wrote:
16 May 2019, 16:24
Foot
For the switch plate foot, black carbon 3d printed plastic seems to work somewhat. It is obviously capacitive. It doesn't matter what you make the barrel assembly out of.
I will be uploading some OBJ/STL files later tonight. The main consideration for this part is whether or not it should have a stem on it. The reason this is a question is that the original flyplate was press fit onto the foot with a metal backward fork preventing removal. I would rather use a small screw as if I am going to make flyplates, I would rather just punch out one hole in the middle with maybe 2 position post circles.
I am stil not sure if i will try that on my own (but i am very tempted and startet to do 3D models and prints for housing already anyway und beeing trained in CAD ... it is tempting^^).
My idea for the flyplate was to 3D print it (as i do not want to go capacitive i will use normal PLA) with 2 stems, put the flyplate spring on them and melt them down a bit (pressing a piece of hot flat metal onto it) to get sthe spring stuck. That should be simple and effective.
SneakyRobb wrote:
16 May 2019, 16:24
Keycaps
I have done a lot of research about the beamspring keycaps. They are cool. I believe that were you to try to make a production beamspring board today, the largest single expense would be the keycaps and it would quite likely take up a massive majority of the cost. As such I am not attempting to do this at this time aside from experiments.
One can make quite good 3d printed keycaps (0.2mm nozzle and quite some postprocessing - so this would not be suitable for mass production because both, printing and postprocessing will take a long time but for a hobby project ...). I am not sure if 0.2mm nozzle would be fine enough for nice legends but i would imagine with some fiddling and a dual extruder printer (i do not own one but that might be a good reason to get one) it might be doable.

edit: with my single extruder printer i would print legends as indentation and fill them up later. Am working on a test keycap at the moment.

SneakyRobb

Yesterday, 17:18

Spoiler:
Monster-Toys wrote:
21 May 2019, 09:25
SneakyRobb wrote:
16 May 2019, 16:24
Foot
For the switch plate foot, black carbon 3d printed plastic seems to work somewhat. It is obviously capacitive. It doesn't matter what you make the barrel assembly out of.
I will be uploading some OBJ/STL files later tonight. The main consideration for this part is whether or not it should have a stem on it. The reason this is a question is that the original flyplate was press fit onto the foot with a metal backward fork preventing removal. I would rather use a small screw as if I am going to make flyplates, I would rather just punch out one hole in the middle with maybe 2 position post circles.
I am stil not sure if i will try that on my own (but i am very tempted and startet to do 3D models and prints for housing already anyway und beeing trained in CAD ... it is tempting^^).
My idea for the flyplate was to 3D print it (as i do not want to go capacitive i will use normal PLA) with 2 stems, put the flyplate spring on them and melt them down a bit (pressing a piece of hot flat metal onto it) to get sthe spring stuck. That should be simple and effective.
SneakyRobb wrote:
16 May 2019, 16:24
Keycaps
I have done a lot of research about the beamspring keycaps. They are cool. I believe that were you to try to make a production beamspring board today, the largest single expense would be the keycaps and it would quite likely take up a massive majority of the cost. As such I am not attempting to do this at this time aside from experiments.
One can make quite good 3d printed keycaps (0.2mm nozzle and quite some postprocessing - so this would not be suitable for mass production because both, printing and postprocessing will take a long time but for a hobby project ...). I am not sure if 0.2mm nozzle would be fine enough for nice legends but i would imagine with some fiddling and a dual extruder printer (i do not own one but that might be a good reason to get one) it might be doable.

edit: with my single extruder printer i would print legends as indentation and fill them up later. Am working on a test keycap at the moment.
Hi,

That flyplate design is interesting. I had considered a similar sort of method where I would print a hole into it, and use a 3dprint pen to shot the last bit of plastic in it to seal it, a rivet of plastic melted could also work. I mostly again just want to avoid having to stamp out the complicated shape in the middle of the flyplate.

I have 3d models for all pieces I shared above, they don't work that well and I am in the process of tweaking them.

In terms of keycaps I think you could probably print quite good keycaps, especially with one of those (resin?) style printers. I only have a .2mm nozzle right now. For those legends I if you are thinking of filling them In I wonder if printing the keycap minus legends and then maybe using the 3dpen might work as well. Perhaps you could laquer them first. The cavity doesnt necessarily need to be filled with PLA, it could be filled with another sort of material, sanded to be smooth and then sealed in.

As well I wonder if the sort of resin custom cap process could work. Like those "klacks/brobots etc" it seems entirely possible to make a new beamspring cap.

Regardless I am not quite at the stage of keycap production or have it on the radar, especially with easily printed cherry mx stems I designed and the existence of custom community made keycaps. I have also seen other forum members print their own cherry mx stems.

User avatar
Monster-Toys

Yesterday, 18:32

I printed a blank test keycap with my FDM* printer (0.2 mm nozzle). It took around 3 hours. WIth some post processing the key is very smooth and shiny and it fits perfectly on the beamspring switch. Funny what you said about the 3d pen, i had the same idea^^ (i do not own one tho).
Printing MX switch mounts seems to be more reasonable.

I came up with some ideas and prototypes how to make a switch with as little time and money investment as possible and had some ideas for a low profile beamspring switch made ... i just need to get 0.015 inches thick spring steel to start making a prototype. Should i make my own thread or just keep posting here, too? (my own thread might be less messy)


* resin printers bring very nice results but the resin is quite expensiv, they are extremly messy and the fumes are quite toxic

SneakyRobb

Yesterday, 18:50

Monster-Toys wrote:
Yesterday, 18:32
I printed a blank test keycap with my FDM* printer (0.2 mm nozzle). It took around 3 hours. WIth some post processing the key is very smooth and shiny and it fits perfectly on the beamspring switch. Funny what you said about the 3d pen, i had the same idea^^ (i do not own one tho).
Printing MX switch mounts seems to be more reasonable.

I came up with some ideas and prototypes how to make a switch with as little time and money investment as possible and had some ideas for a low profile beamspring switch made ... i just need to get 0.015 inches thick spring steel to start making a prototype. Should i make my own thread or just keep posting here, too? (my own thread might be less messy)


* resin printers bring very nice results but the resin is quite expensiv, they are extremly messy and the fumes are quite toxic
Hi,

It would be up to you. You seem to have a different sort of switch method in mind. Lower profile, sensing method and the such. I think a different thread might be a good idea. I am not that organized and you might find it easier to keep your files and everything in order. I think you have the thread that I started posting in before I made this one right?

I have also been experimenting with shortening the switch. It seems like along length of the slider portion you could remove perhaps 5mm of length and shorten the switch in this manner.

I think you could plausibly delete at least the area in the redlines here.
beamspring side.jpg
beamspring side.jpg (282.42 KiB) Viewed 43 times

There is also this product which obviously has shortened the mechanism albeit with different dimensions of spring.
https://kono.store/blogs/keyboards/silo-beam


Something like the ALPS plate spring switch might also be good to investigate. It uses a somewhat similar method of clicking.
If you don't have an ALPS plate spring switch, I can mail you 1 or 2 of them.

Post Reply

Return to “Workshop”