Beamspring Factory process - How to

User avatar
DMA

13 Jul 2019, 04:33

Needle-nose pliers are da shit.

As for the bending tool - plastic is likely too soft for that. Metal will probably deform it.
Also this being spring steel - to give it a shape you must bend it _beyond_ that shape. And if it's not - it might have trouble holding the shape for long.

I like the idea of pliers though. There's some real love there.
You'll fuck up first 20 and probably cut yourself with the sharp edges couple times, but then you'll have a wonderful half a day of bending the metal to your will, and as a result you'll have a set of springs for the best keyboard you'll ever have.

After 2 years I still find it strangely satisfying to type on the keyboard which has parts _made_ by me.
Doesn't happen every day, but it comes across my mind every now and then (usually when people like "..and I thought _my_ keyboard is clicky") and puts a smile on my face.
Every time.
But software engineering is not for everyone, and flashing the firmware in a part you received in the mail just doesn't count (although, I admit, it _is_ a matter of perspective).

And bending springs is a thing that requires skill, yet is not beyond reach of the person reasonably willing to learn it.

User avatar
SneakyRobb
THINK

20 Jul 2019, 01:13

Hi,

I have due to a stencil manufacturer error only about 50 flysprings in these packs. I tried for 0.003 inch thick steel this time. I think it will allow for a nicer* feel. We will see.

I will be begin bending these and combining them with my carbon 3D printed feet. I still don't know if I like the screw mount or the press mount better. I don't think it matters too much for the capacitive foot. The beamspring mount is much more critical.
Spoiler:
20190719_185329.jpg
20190719_185329.jpg (2.67 MiB) Viewed 2981 times
I have now received... Hopefully... please.... please?... my 4th pcb revision for the 60% beamspring keyboard. I think this one will work. I ordered it to the specs and requirements that DMA laid out for me. I think it will finally 100% work.
Spoiler:
20190719_185449.jpg
20190719_185449.jpg (2.84 MiB) Viewed 2981 times
20190719_185522.jpg
20190719_185522.jpg (3.02 MiB) Viewed 2981 times
20190719_185538.jpg
20190719_185538.jpg (2.8 MiB) Viewed 2981 times

After this pcb, I will be begin working on my end game. The Full Sized model M style beamspring keyboard with 100% new construction parts.

I have procured silicone molding materials and will begin casting full beamspring housings soon. I will also update my 3d Models gallery in this thread.

For stabilizers I have now decided on a policy of "only space bar." The beamspring is so robust and side to side stable that basically the only key that needs stabilizers is the spacebar.

This crude drawing is how the rest of my stabilizers will work. I will either cast these parts or 3d print them.
Spoiler:
mx stem crude drawing.png
mx stem crude drawing.png (4.3 KiB) Viewed 2981 times

Will update on my flyspring bending adventures soon.

User avatar
DMA

20 Jul 2019, 04:48

I don't think you need such stabilizers - the switch is under the key center everywhere.
The stabilizer is there so it pulls down the other side of the key when you press off-center. To achieve this, it must be held down at the other end.
Suppose you press the left end of your spacebar. Downforce on the left is converted into stabilizer rotation, which twists the stabilizer stem, and tries to convert that rotation back to downforce. But it needs an anchor there - otherwise the stem will just raise towards the key, not press down on it.

Your scheme has no anchor, so it won't help much there. The only benefit is a bit less bending of the keycap socket - not sure how much of a problem is that. For shitty keycaps it's a big problem I guess.

User avatar
SneakyRobb
THINK

21 Jul 2019, 03:57

DMA wrote:
20 Jul 2019, 04:48
I don't think you need such stabilizers - the switch is under the key center everywhere.
The stabilizer is there so it pulls down the other side of the key when you press off-center. To achieve this, it must be held down at the other end.
Suppose you press the left end of your spacebar. Downforce on the left is converted into stabilizer rotation, which twists the stabilizer stem, and tries to convert that rotation back to downforce. But it needs an anchor there - otherwise the stem will just raise towards the key, not press down on it.

Your scheme has no anchor, so it won't help much there. The only benefit is a bit less bending of the keycap socket - not sure how much of a problem is that. For shitty keycaps it's a big problem I guess.
Hi this does make sense. The main problem is that for original beamspring switches, the metal stamped stem that the key sits on is very tall, very thick and very sturdy steel. The keycap socket holds onto that stem with thick walls that really brace against rotation. You would only ever break your keycap socket if you used excessive force to rock the keycap sideways when removing it.

The sort of rotating "left down/right down" motion that a stabilizer would correct for basically doesn't happen because of this. It is so strong that the only key on the original with a stabilizer is the spacebar. Even then, you can honestly remove the spacebar stabilizer and they keyboard is still mostly useable.

For cherry mx mount, the stems being plastic, and in this case 3d printed by me, they are in fact much less durable. This is just due to the 3d printing method I have used where the layers go across. So rotational force, can break the stem off while inside the keycap socket.

So the stabilizers I am printing are not actually stabilizers, they are stress relievers. They are designed to relieve stress on the cross of the stem, not the keycap. So that I don't have more keys that end up like


Like this.
Spoiler:
20190720_213830a.jpg
20190720_213830a.jpg (3.14 MiB) Viewed 2916 times

User avatar
DMA

21 Jul 2019, 07:51

"keycap support structures" :)

User avatar
SneakyRobb
THINK

22 Jul 2019, 06:30

DMA wrote:
21 Jul 2019, 07:51
"keycap support structures" :)
Hi, yes very close to it!

I will assemble the 60% now. and hopefully post next day with everything together.

I will try to update more on the more regular everyone sorry for recent slowness.

Flyspring bending.
So for the beamspring flysprings. The 3d printed aparatus works just fine. ALL it takes to make beamspring flysprings is the metal and the 3d printer object to shape them on. The real key thing, is to get them square. This can be done by eye and just requires perpendicular angles. Once you get that. You can align them to perfection. The final shaping here was slightly off.

I think ill just make the bending tool which is just a quick 10 minute 3d print, have alignment slots.

Pretty okay/10
Spoiler:
20190721_201131.jpg
20190721_201131.jpg (1.85 MiB) Viewed 2855 times
For some reason after a lot of back and forth with JLCPCB, I didn't see it at first, but all the flysprings I just ordered were delivered as 0.15mm which is about 5 mi. So they are too thick :/ I wanted 3 mill which is about 0.08mm. Most unfortunate.

So hopefully I will get the correct thickness stencils soon. I don't think JLCPCB is the best long term partner...
Spoiler:
beamspring holder.JPG
beamspring holder.JPG (110.01 KiB) Viewed 2855 times
-Not original.
Finally one last thing that I have been trying to experiment with. This is something I sort of came across by happenstance but that I think could be big. When I was cutting beamsprings by hand with my hole punch I had a lot of issues getting the center holes just right. So frustrated I put the flat hole-less spring in the flyplate. The feeling? The best. Why does the beamspring need this hole in the middle anyway? Alignment?

If I could have one issue with IBM Beamsprings. This is not an issue you would ever really come across unless you were experimenting with them. Beamsprings are known as perhaps the best "Clickly" switch. They have this massive 20mm long plate spring. If you look at this spring though. It has a giant plastic rivet in the middle with a huge square hole. This hole/rivet almost cuts it in-half. You would never guess this, because they are known as the #1 switch. But that hole takes them from absolute best switch, to best switch.

What if that hole wasn't there? Maybe the spring can just slide into the stem? Instead of a 20mmx3mm spring with a huge hole in the middle for riveting or for me, screws. We just get, a captive spring of full length, with no interuptions? I will now try to also find this out.
Spoiler:
beamspring holder.JPG
beamspring holder.JPG (110.01 KiB) Viewed 2855 times
This project started as a recreation. I intend to finish that, but what if. What if we can take the #1 switch, and make it #1+..?

User avatar
SneakyRobb
THINK

13 Dec 2019, 16:10

Hi, long time no see. I am still alive. I am now trying to figure out the best way to affix beamsprings to the stems. My current idea is to use resin casting to make stems with a small hole in the bottom that can be filled with more resin.

My 3d printer is okay but not the best resolution, so I have ordered the stem piece from a higher quality printer service so that I can cast it and make more.

Here is that design. It has a raised area slightly smaller than the original size cutout so it can fit on without rotating. It also has some interior texture so the resin or other can hold it in place.
stem boi.JPG
stem boi.JPG (21.45 KiB) Viewed 2419 times
Aside from that I have finally got a supply of new beamsprings!!!!!!!!!!!!!!!!!!!!!! Has taken forever to get a quote that wasn't in the thousands of dollars. They had a hard time cutting them so there were many failures, but they seem okay. Some are a little rough, but they should still work.

Will get some better photos later. Might be January before I post again but, the factory is back open.

Source material
material cutout.jpg
material cutout.jpg (323.06 KiB) Viewed 2419 times
Original Size ones.
original square hole.jpg
original square hole.jpg (288.08 KiB) Viewed 2419 times
Circular hole ones.
circular holes.jpg
circular holes.jpg (433.32 KiB) Viewed 2419 times

Fkazim

13 Dec 2019, 16:47

Wow amazing work glad to see everything is on track.

User avatar
SneakyRobb
THINK

13 Dec 2019, 23:58

Fkazim wrote:
13 Dec 2019, 16:47
Wow amazing work glad to see everything is on track.
Hey thanks.

Okay I am home. I have handled these beamsprings. They are 0.0015" thick vs the original 0.001" thick. I thought this would be enough, it isn't.

They still crease too much and are soft. The original ones are much harder and much more rigid. I got special shim spring steel but still it is no where near as strong. I am honestly going to get a blow torch and try to quench them.

Will report back.

:\

User avatar
DMA

14 Dec 2019, 09:02

Quenching might make them brittle. But you have dozens and it doesn't hurt to lose a couple in the name of !!SCIENCE!!

Keep us posted!

User avatar
SneakyRobb
THINK

15 Dec 2019, 00:11

DMA wrote:
14 Dec 2019, 09:02
Quenching might make them brittle. But you have dozens and it doesn't hurt to lose a couple in the name of !!SCIENCE!!

Keep us posted!
Hi yes agree.

I am now trying to consider how to proceed. The "flyspring" part which holds onto the capacitive foot seems fine being made out of the AISI 1095 carbon steel/shim/spring steel that I am using.

The beamsprings themselves are the current problem. I just can't seem to manufacture them to be tough enough they work for a while, but crease. Versus the originals, the originals are much much harder.

How did IBM make them so thin, yet so extremely hard. I must conclude that the metal stock I am using is incorrect.

The two options now are these: I will temper the ones I have.
I will try to find a better steel to use.

I will explore both of these options although the second seems better.

User avatar
SneakyRobb
THINK

07 Feb 2020, 01:55

Hi, okay.

I got some shapeways sliders to experiment attachment methods with. So I am now trying to figure out how to best attach beamsprings to these various sliders. I was thinking that I would put the beams onto the sliders with an alignment square shape. Unfortunately the square I made wasn't high enough to align the beamsprings. I will order them next time with a higher alignment square.

I would then after alignment put some resin in. This kind of worked, but maybe Ill design it with a plug, or even with a peg that I will melt/rivet them together with. Must experiment.

beamstem.jpg
beamstem.jpg (44.27 KiB) Viewed 2196 times
alignment tab not tall enough.jpg
alignment tab not tall enough.jpg (204.5 KiB) Viewed 2196 times

The metal beamsprings themselves are not my final version. I still need to get proper spring steel cut, but it is expensive to get metal finely cut like this so I will wait. This high carbon steel does still work well for now surprisingly! Even the tiny resolution they are quite durable.

I got 3 types of plastic from shapeways from basic to very smooth clear. The basic one was actually perfectly fine and usable in terms of smoothness. There is no reason to get a higher quality plastic from shapeways. Not sure if ill print more sliders like these myself or order them.


All the stems from left to right
1 Original
2 Super smooth high quality clear plastic
3 normal plastic with finishing
4 most basic plastic
4 stems.jpg
4 stems.jpg (244.92 KiB) Viewed 2196 times
bottom holes.jpg
bottom holes.jpg (211.01 KiB) Viewed 2196 times

The single barrel I got from shapeways, was actually pretty bad and poorly printed. There should be 2 holes you can use to reseat the beamspring. Where are they shapeways? Given that they actually did really well with the stems I am a bit sad. As well there is this weird lip inside the barrel. Very strange. Doesn't really work that well with any original or new slider.
barrel 1.jpg
barrel 1.jpg (4.55 MiB) Viewed 2196 times
barrel 2.jpg
barrel 2.jpg (1.95 MiB) Viewed 2196 times

Here I mix resin for the next part.
lets mix.jpg
lets mix.jpg (254.58 KiB) Viewed 2196 times

I put the resin onto the stems and hope it goes into the hole. This kind of worked, but I will improve it. 2 had too much resin, 1 had not enough.
too much resin on all.jpg
too much resin on all.jpg (221.33 KiB) Viewed 2196 times
too much.jpg
too much.jpg (117.91 KiB) Viewed 2196 times
The one without enough resin immediately broke off.
too little resin.jpg
too little resin.jpg (152.46 KiB) Viewed 2196 times

The other 2 worked great.
stems up.jpg
stems up.jpg (188.37 KiB) Viewed 2196 times
stems down.jpg
stems down.jpg (182 KiB) Viewed 2196 times

Then one broke. Sad.
second casualty.jpg
second casualty.jpg (186.58 KiB) Viewed 2196 times

Otherwise I am pleased with the situation. I will try to refine this attachment method.
Here is a video of a switch that aside from the metal slider and return spring. Is 100% new production.
https://www.youtube.com/watch?v=z_G94hj ... e=youtu.be

User avatar
lucar

07 Feb 2020, 10:10

Unbelievable work!! I'm impressed, what about springs, did You find any clue about how to reproduce them?

Luca

User avatar
XMIT
[ XMIT ]

07 Feb 2020, 14:55

I have a friend who does elemental analysis of metals for a living. If you have an OG beam spring that you're willing to sacrifice for science, let me know and I'll ask for an analysis.

User avatar
wobbled

07 Feb 2020, 15:01

Yeh I could analyse your Beamsprings for you guys as well, I wont be able to tell you anything, but always worth a try.
Send em over!

kmnov2017

07 Feb 2020, 15:17

A single beamspring switch costs 9 USD. Cheap enough for a sacrifice.

User avatar
SneakyRobb
THINK

07 Feb 2020, 16:37

lucar wrote:
07 Feb 2020, 10:10
Unbelievable work!! I'm impressed, what about springs, did You find any clue about how to reproduce them?

Luca
Hi,

The external coil springs are probably the last unsolved element at this time.

I will eventually probably buy some of
these
https://us.misumi-ec.com/vona2/detail/1 ... wyg_area_2

or these
https://www.thespringstore.com/pc020-50 ... -n-in.html

These springs are both... roughly about where they should be. I have found that my original beamspring coil springs have a slight hard time returning up after each press. Not exactly sure why. These ones are somewhat heavier which I don't necessarily mind.

I haven't got them yet, because it costs a lot to buy them and then ship them. I will at some point though!

XMIT wrote:
07 Feb 2020, 14:55
I have a friend who does elemental analysis of metals for a living. If you have an OG beam spring that you're willing to sacrifice for science, let me know and I'll ask for an analysis.
I will keep this in mind but at this point I am confident the spring steel I have will work.
https://www.kbctools.ca/itemdetail/1-957X-K1 I have several pieces of this feeler gauge metal. I hand made 1 spring out of it and it worked good.

I am getting some of this metal cut soon like the previous beamsprings I had cut. The previous beamsprings, as I mentioned are made of just plain carbon steel, and they actually work pretty well all things considered. Most issues I had with them were due to previous screw method of attaching them to sliders creating creases/folds/tears.

When I was able to attach them, they did click.


This is what I had in mine as a next version of the slider bottom for attachment. It's a peg that will have some resin on it, instead of just trying to put some resin in. Hopefully this will make a sturdier tighter connection.

stem alignment peg.JPG
stem alignment peg.JPG (37.46 KiB) Viewed 2067 times

User avatar
ZedTheMan

07 Feb 2020, 16:54

What if you used acetone to bind the peg and slider?

User avatar
lucar

07 Feb 2020, 17:28

"The external coil springs are probably the last unsolved element at this time. "

Check Aliexpress, https://www.aliexpress.com/item/3296025 ... 4c4d90iFjC

Those are havier than the originals but better than none.

Luca

User avatar
SneakyRobb
THINK

07 Feb 2020, 17:57

lucar wrote:
07 Feb 2020, 17:28
"The external coil springs are probably the last unsolved element at this time. "

Check Aliexpress, https://www.aliexpress.com/item/3296025 ... 4c4d90iFjC

Those are havier than the originals but better than none.

Luca
Hrmm, that actually looks like it could work size wize... I'll risk it. Ordered. Thanks for the recommend!

The 0.5mm wire is actually slightly narrower as well. I think the inside diameter should still clear the middle. Spring cavity is...from memory 13.3mm at the outer, and 10.4mm on the inside..? So this 12mm OD spring with 0.5 mm wire, should be 11mm along the inside part. So it should fit as far as I can tell. Away from main pc right now.
inside diameter spring cavity.JPG
inside diameter spring cavity.JPG (105.74 KiB) Viewed 2028 times
ZedTheMan wrote:
07 Feb 2020, 16:54
What if you used acetone to bind the peg and slider?
Hi,

That is also a good idea. I got these from shapeways for fun because I had a gift certificate. I will also try printing in ABS for that reason.


As well my milling operation has been very very slow to develop but I can now reliably cut holes. So stay tuned for that!
20200202_144040.jpg
20200202_144040.jpg (2.59 MiB) Viewed 2028 times

User avatar
SneakyRobb
THINK

01 Mar 2020, 00:56

So those return springs havent arrive yet. Obviously many issues with ordering from abroad right now.

One thing that is long overdue and I am just bad at updating is the capacitive foot assembly with flyspring. I am bending these en masse for assembly. Pretty simple operation, might take a while. I use the 3d printed piece block to get the general shape and then the loop pliers to secure them into the final shape.
20200229_181528.jpg
20200229_181528.jpg (4.64 MiB) Viewed 1825 times



Another thing I've done is order some 8.5mm teflon tubing from China. I was wondering how I would get the sides of the slider smooth if I print or cast them. Maybe sand or polish. I finally decided to get some teflon tubling of the same diameter. 8.5mm. I will slide it onto the sliders.

Will update again soon.

User avatar
SneakyRobb
THINK

14 Mar 2020, 22:27

Hi,

So the springs lucar recommended finally arrived.

They have many less coils than the original ones but are taller so will be compressed. Will upload some more images later of an original switch assembled.

They seem to offer about the same resistance but I don't have a thing to measure.



They are the correct diameter and do fit in the switch. They came in a giant ball so it was hard to get them apart.

I am building up the nerve to disassemble an original switch because I always somehow slice my hand open doing so.

On a reproduction switch, they do work, but I want to make sure on an original switch that they could work.

I hope to report back soon they do, fingers crossed.


Hopefully this month the teflon tubing will arrive so I can affix it to ABS printed sliders.

Robb


side by side.jpg
side by side.jpg (912.16 KiB) Viewed 1707 times
old.jpg
old.jpg (829.79 KiB) Viewed 1707 times
new.jpg
new.jpg (916.67 KiB) Viewed 1707 times
box.jpg
box.jpg (1.34 MiB) Viewed 1707 times

User avatar
SneakyRobb
THINK

15 Mar 2020, 00:50

Hi,

Okay. I can confirm that every other piece being original. The springs user lucar recommended, work.

A video ( the tape method is responsible for noise differences. Also my wife was listening to some music in the background...)
https://youtu.be/3N6w2Tx058s

Another video where I remove the tape on the original so you can hear it held against table.
https://www.youtube.com/watch?v=R_Ef86WL1gk


So, thank you lucar, springs WORK.





20200314_192239.jpg
20200314_192239.jpg (1.12 MiB) Viewed 1681 times
20200314_192501.jpg
20200314_192501.jpg (1.11 MiB) Viewed 1681 times
20200314_192505.jpg
20200314_192505.jpg (1.08 MiB) Viewed 1681 times
20200314_192613.jpg
20200314_192613.jpg (943.2 KiB) Viewed 1681 times
JPEG_20200314_192538.jpg
JPEG_20200314_192538.jpg (1.15 MiB) Viewed 1681 times
Attachments
JPEG_20200314_192256.jpg
JPEG_20200314_192256.jpg (1.06 MiB) Viewed 1681 times

User avatar
DMA

16 Mar 2020, 00:05

Super, man!

Ball of springs - just get hold of one and screw it out of the ball.

What's with all the copper?

User avatar
SneakyRobb
THINK

16 Mar 2020, 14:13

DMA wrote:
16 Mar 2020, 00:05
Super, man!

Ball of springs - just get hold of one and screw it out of the ball.

What's with all the copper?

Hi thanks.

Yes it was a process. I thought it was a fake box when it arrived it was so small and light. Felt empty and heard no metal on metal.

I have a lot of copper tape and I enjoy how it looks! I've also used it to fix mistakes on my Cap-sense PCB's before.

User avatar
lucar

17 Mar 2020, 22:47

SneakyRobb wrote:
15 Mar 2020, 00:50

So, thank you lucar, springs WORK.
Thanks, I know they work but they are stiffer and less fluid then the original ones in my opinion. I consider these springs the best affordable alternative if you have nothing else. I also reproduced the metal stems and made some samples but covid-19 trapped all us and the local laser shop is closed. I also managed to duplicate the hinge and spring of the plastic front door that holds the user manual of the 3278 and I'm waiting for a tool to reproduce the space bar metal stabilizer and more parts. Before releasing anything, I need to test and assemble all parts and check everything.
It will take time. Actually I'm busy preparing live YouTube sessions from my workbench for the Italian mechanical keyboard interest group ClickClackClub to keep guys distracted from the horror story here in Italy.

Luca

User avatar
SneakyRobb
THINK

18 Mar 2020, 16:53

Hi, I think you are correct. I do notice that the newer ones are stiffer. Regardless I appreciate the recommendation.

That is cool about the front door area. I've always not liked that part of the board and a replacement would be excellent.

That would be great to watch a livestream. Would be interested in watching it.

I should really try making the dual stem pieces ive mentioned out of metal this time. I think this could make for a sturdier keycap interface for mx keycaps.

I don't think i've posted this image before...?

regardless I will mill some of these in aluminium to see how they work with mx keycaps.
I was thinking of some 1/4 circular pieces that would slide in the plastic stem-tube part to make the metal-stem not rattle a lot. When you pull out the locking metal stem piece, the circle parts would not be held in place anymore and would just fall out.

2 piece and a keycap.JPG
2 piece and a keycap.JPG (80.55 KiB) Viewed 1378 times

NadeemDoesGaming

11 Apr 2020, 04:54

Hi, could you post a clear picture of the wiring between the PCB and the DMA Commonsense controller?

User avatar
SneakyRobb
THINK

12 Apr 2020, 21:20

NadeemDoesGaming wrote:
11 Apr 2020, 04:54
Hi, could you post a clear picture of the wiring between the PCB and the DMA Commonsense controller?
Hi, I have it disassembled right now. But I will reassemble it for you soon.

I basically just used the wiring diagrams from the DMA thread. It is a very basic x and y matrix.

User avatar
SneakyRobb
THINK

22 May 2020, 01:56

Hi sorry about this question. I have just received this new pcb so I will update this tomorrow.

Will finish this project soon.

Beamspring model m is soon.
IMG_20200521_194034.jpg
IMG_20200521_194034.jpg (2.82 MiB) Viewed 799 times

Post Reply

Return to “Workshop”