SHINSANWASWITCH - blog of creating DIY-analog of OBSA-LHSXF-LN

Hello friends.

I hope the moderators will forgive me, I just do not know where this can be published, but a little later, I hope here will be a real tutorial.

I got the idea to build a new controller, but only to make it better, more convenient and more reliable, especially in maintenance, than the one that I have now, because now I have optical switches, but they are canopy, on snot, working good, but they needed veeeery long time of calibration and that time when I needed to disassemble buttons - I cannot assemble them again for a long time, because I lost order, to which button which switch (sorry for the pun). So I understand that I need to think, to do something with it. I really want to use the original Sanwa OBSA-LHSXF-LN, the only problem is their rather high price, especially for me. In this all, perhaps, it will also be your personal interest, since if I succeed, then you can also use all these developments in order to install it in your controller.

What exactly do I want to do?

I want to completely repeat (clone) the OBSA-LHSXF-LN by 3D printing the outer casing and the inner PCB of the device, since I am sure that the board inside is elementary and there will definitely not be much difficulty in repeating it, the outer casing will be more difficult.


But I have a few fundamental questions that accumulated about this case:

1. Some time ago I wanted to buy one OBSA-LHSXF-LN on Rakuten, but did not have time and the store closed. Where I can buy it now? Who knows other places?
2. Does OBSA-LHSXF-LN fit into an ordinary Chinese button? (like this on photos bellow) In theory, it can be installed and work, but again, may be anyone have experience? I need more accurate information on this issue until I have my own OBSA-LHSXF-LN. (Sorry, I do not know why, site do not want to upload my photoes, so here it is on my google photoes here and here.)
3. If anyone already has OBSA-LHSXF-LN and you could very much help me if you disassemble it and show me what it has inside, in order to approximately understand its contents and components. Can someone help with this? More detailed view of external case parts.
4. It feels like there was another question, but I forgot which one .. :)))) But if I remember, I will add here.

If, nevertheless, the above questions will be solved, if I get my hands on one OBSA-LHSXF-LN and if I can repeat it - materials for 3D-printing and gerber-files of the PCB that is inside, for the production of the PCB in some Chinese factory (or maybe even for self-production at home, everything will depend on the complexity of the board, but I do not think it wil be difficult) will be published by me here and you can use them. Source-files will of course also be available. It seems to me that this is a completely feasible task.

What do you think?

UPD: SHINSANWASWITCH completed and available on my Github, make it and use it :)
https://github.com/steelpuxnastik/SHINSANWASWITCH

 

Thank you very much! One question is closed, somehow.
Can I ask you to show more detailed the outer case components?
Daaaaamn, it's a way more complex, than I thought.. So in real, i thought before, it's somehow uses only lever arm of button itself for crossing the optocoupler, but in fact, arm of button only acts on another lever arm inside button that crossing the optocoupler, am I right?
But still I do not understand, can it be used with ordinary chinese button, as I have? Need to see it assembled both.
Pleasant, it can be made as single layer PCB, so it can be made at home, if you have enough skills. But again, it needs to find components, especially this variant of optocouper, but I think, it's doable task.
And now I am even more convinced that I need to get my hands on it - needs to find place where to buy it!

 

You are quite right, yesterday I came to a similar conclusion about it (and because of problems which you voiced above) and I realized that no need in one-in-one copying of it, because we can use only a conception of it (we see how it works) and design everything to fit regular chinese button (the main thing is how it mounts here and how mounts LED inside it, similar to ordinary switch fix, like on photo)

also with using only common components (because I see that I cannot find anywhere any somehow similar in size optocoupler to original LHSXF, so result product will be larger then LHSXF, because of optocoupler, but in any case it's an opportunity to choose better or best optocoupler). The most difficult moment for me is to understand how to made similar leverage system, that will affect the optocoupler and is it worth to be repeated at all? But in any case, this formulation of problem is greatly simplifies the matter, that it becomes clear where to think.

 

I have some progress in this business and little success. I found some 3D-models for SolidWorks of arcade buttons, after little editing I printed on 3D-printer switch holder to understand, does it fits the button. And it fits! But may be needs some very little editing more to fit easier.


But that means that only think I need to do next is to make PCB of switch and make design of it's internal lever mechanism, it will be pretty easy, much easier then in OBSA-LHSXF-LN. Button will push lever like in normal switch (red on the photo above) and it will cross the optocoupler. About optocoupler: I found ultrasmall, and may be the same optocoupler used in OBSA-LHSXF-LN, so it ideally will fit, I hope.

In very short term I hope I'll make PCB of it and will know dimensions of PCB and in accordance with that will make outer and inner design of switch. I hope it will be success :))) The only difficulty now is that I do not really know how to work in SolidWorks and I need some time to learn this system.

 

Finally I made 3D-model of RPI-221 optocoupler and received some experience in SolidWorks, that was pretty solid work :)) Two days for 10 minutes work :D


And after that completed PCB model for understanding dimensions of PCB for making outer and inner case design.
Connectors looks gigantic, but in my case there will be connectors without those "ears", and as I can see there will be enough space to connect LED, so I think it will be good.




I think for LED will be used this kind of socket, it will be glued inside the hole for LED and it will be needed only to cut legs of LED to shorter length and put inside.

I hope next part will be very soon, because I'm very interested :))

 

Some huge update!
With the great help of @Doctopus were made several fundamental changes in principle schematic that theoretically will make my "shinsanwaswitch" (it's the name of this thing :))) just like original switch. @Doctopus helped to analyze elements of original switch, told me what would fit in it and in result I've got this

The main fundamental change here is usage of n-channel mosfet that unlike scheme before (it's the same scheme that I use now in my controller) it works like standard mechanical switch: when button not pressed - at the output of switch has Hi-Z (not connected), but when pressed - minus, but before it was minus when not pressed and Hi-Z when pressed and it needs to invert signal programmatically. That means more easy implementation.
I hope it will work just like I described :)))
And PCB that turned out in result.



This board was exported to solidworks and finally I managed to do case and mechanism of switch that will be compatible with chinese button.


At this variant it will use M2 screws with countersunk head, 2mm outer diameter spring with 10mm length (I hope 0,2mm wire diameter will cause no inconvenience, because I have hope it will be very light in terms of pressing force and with 200g sanwa spring it will be the same as original configuration).

The LED will be mounted, as I mentioned in previous message inside BLS-socket connector, that will be glued inside the hole, the length of LED's connectors must be cut to make shorter.

I think in near days I'll print this elements to see, if I made mistakes and will correct it and if everything right in PCB - will make order of PCBs and after that I'll upload everything that you can make the same

 

Another huge fundamental changes in this project, because I try not to stay still and the thought process is moving somehow. Now with the great huge help of @Doctopus was noticed that we have here mistakes in principle schematics, but we finally managed to find out what exactly parts used in original OBSA-LHSXF-LN. And what is it:

Was figured out that Q1 element is not n-channel mosfet, but NPN transistor, and in result changes with resistors. Also, I understand that PC1 RPI-221 is too small and not this optocoupler were used and originally were used RPI-352, that much bigger and it's much better to use (I do not know why I didn't not understand before, that on board it uses special mounting holes, but better later then never). And another thing that universalizes using different LEDs with different input voltages: instead of resistor I think to use D1 NSI45020AT1G is a constant current regulator and LED driver 45 V, 20 mA ± 10%, 460 mW package that gives possibility to supply LED in switch from 5V to 45V input voltage with stable 20mA output. that in case if you'll use 12V or 5V LED-strip or something like that gives you possibility to not think about making special 12V line to supply LEDs in buttons, can be used 5V or 12V without thinking. Or if 20mA is a lot for you - you can use NSI45015WT1G with 15mA output or lower. the only thing you need to think over - what exactly LED you will use and what brightness you'll need (because, relatively speaking, more LED consumption - more brightness) and if you need less brightness - you need less output. In any case, standard LEDs from chinese buttons uses 20mA, so I suggest using this.
I made some tests too to make sure I'm not mistaken. I took the LED from chinese button (fortunately or unfortunately, I have a lot of them, because I bought year ago a lot of different buttons and there was even some extra spare LEDs for buttons, so I took one, removed resistor from there (there was a little shock for me, that it's already has resistor, because I additionally soldered resistors for LEDs because thought there is nothing but LED), increased missing contact length, soldered to NSI45020AT1G wires, because it's an SOD123 package SMD component and cannot be used with breadboard and I get that everything works perfectly:





And it's shines with fixed brightness on any input voltage.
I tested this with resistor on LED too and with resistor on 5V it's shines a slightly less, but not really tangibly, what is naturally. So I'm sure that the decision is at least not wrong.

PCB somehow begins to look closer to original:




So I hope in case of electronics it will work well.

In case of body also some changes, because @Doctopus and my friends that has 3D-printers said me what changes to make, to make it better, but I still haven't printed body in any variant, so I do not know how it works in real.
But:

Huge work for me was attempt to fill all extra cavities (there were a loooot inside) and making one solid cut from up to down for BLS-02 LED-holder as to say and wires from it to PCB. And except one little nuance I managed to do that, so in result it will be easier to print (I hope so) and it will be much stronger.

Also solid fills of this elements to make them more stronger.

On cap of body this bosses to fix lever inside body to move only up-down (about upper), the bottom too + it will hold the spring.
I hope next week I'll receive springs for it and someone (at least, I'll print it on order for money in some company) will print me it and finally assemble body to test how everything works.

 

Finally!
Finally I got printed prototype of ShinSanwaSwitch's body to understand am I doing right or not and I can say that mostly I'm doing right because mostly everything works as expected. You can see it down in video.
This printed variant is very raw because at this moment no need in high quality, and that's the reason why some details needed to finalize with sandpaper: I'm talking about lever, because it had roughness and the other thing that I (may be) mistaken with diameter of cylinder, because I have spring with 4mm outer diameter and I barely managed to pull spring on cylinder, later I finalized it with sandpaper and it got easier, but the spring itself needed with larger diameter, may be 5 or 6mm. Now I have spring with 0.2mm wire, 4mm diameter and 10mm length and it has ~50g when fully pressed to 2mm length (and on that length that is used to compress it's may be somewhere ~10-15g) and I REALLY LIKE how easy it's going, how easily it can be pressed, like feather and @Doctopus suggested to get spring with wider wire to get compression rate like in OBSA-LHSXF-LN (it has 70g if I understand right) but I think it's way too many and to get real effect - it needs to be as easy as possible. So may be with 5 or 6mm spring it will be the best.
The other noticed thing that is needs to be changed is the top (where is plans to glue BLS-02 connector to mount there LED), because with this moment it's blocks button to be pressed inside, so it needs to be round, or may be just smaller height, so then I finalized it with sandpaper - it become better.
So next step is to test it with already assembled PCB to make final changes because I have fear with it will be problems too.
But we'll see it soon. I hope to continue to regularly supply you with this very important info *sarcasm* :))

 

Hi again!
After first printing of ShinSanwaSwitch body I found some problems of it and brought in some changes. I think that it was one of the last changes, because changes are really massive in many senses and solves many problems, the only left is to test it with real PCB inside, because all parts for it I already received and I need only to order PCBs in one of the china's factory.

First of all this: this will give able to print body laying on it's back with much less supports then it's needed before and it gives other possibility.

This gives possibility to compensate that fact that PCB in places of soldering points will be higher then the board itself, so in this way we give possibility to put these protrusions there and keep board at the fixed required height. Also were changed holes on PCB to bigger, because past printing shows that 3D-printer can have problems with printing such small elements, so better to make it bigger as possible, but their height will not be very big and only gives possibility for PCB to cling on them and this gives possibility (I hope so) to have no problems with PH2.0 connection of LED. Was problems with printed width of the wall that red arrow shows on and with it all walls in the entire body was changed from 1mm to 1,5mm width that presumably will make it stronger. As result dimensions of entire body were changed to bigger, because there was need to move PCB, but I think that this will not be the big problem. Were corrected the screw holed diameter, because there was somehow 2,4mm. that for M2 screw that I use was on the verge of clutch, so I changed it to normal 2mm and I think with inaccuracies of printing there will be exactly what is needed. The only think I need to add here is, may be, some hooks for grabbing the PCB so it can stand still without other help, but probably I'll use double-sided tape that will be glued to the surface marked with the two horizontal stripes and most likely will never get PCB out of there again :)

Also was changed the height of element where will be mounded BLS-02 for LED, because when I've got printed variant - I had problems with it's mounting inside the button and button was not pressed because of great height of that element and needed to work with sandpaper to cut a little extra height of it and after that there was no problems. But BLS-02 element will be need to glue like this or higher (that will be determined experimentally or just I measure height of same element on chinese LED holder and say how much there will be needed to achieve the similar effect). And the most important moment that dimensions of hole itself were changed to bigger, because at printed variant I had no possibility to install BLS-02 inside, because hole was small for it. Have hope now will be much better.



In result cap were changed to because of dimensions changes, screw holes were moved to a fixed from the edges distance, corrected supports for a lever and some others.

So this were finalizing changes that strengthening and eliminating noticed deficiencies.
I hope later will be only better :))
Likes please :))))

 

Finally ordered PCBs of shinsanwaswitch, even more, I tried to make panels and if everything will be right - I'll got a loooooooooot of PCBs, so if everything will go right and if everything will work as planned (and everything goes to this) - maybe I can try to sell ready assembled switches for much cheaper, because as you can see, with shipping it's pretty big investment. It's just possibility, not an offer, but may be if who interested - you can take note about this. And made order on last on this moment version of body for shinsanwaswitch on quality prototype 3D-printing, will receive it maybe in Monday of Tuesday and show you.
Also will make new slider developed (still in development rather, I hope to contribute to this development too) by @Doctopus for new controller, so everything is going to be interested :)))

 

Finally got printed last at this moment version of SHINSANWASWITCH body and I pretty like it, except that moments that I needed to finish something with sandpaper, but I'll add some more tolerance to dense moments and may be I'll ask about some more print accuracy and then it will be the best.



As I can understand, PCBs I'll receive 1st December, why? I mean, how FedEx know when they arrive it to me at that long time distance? But I hope that I'll receive it earlier :)
Also, I hope that I'll receive proper springs for it, I hope it's not lost somewhere, because there is no news from it in 16 days :( because spring that I have now is cannot be used :(

 

Finally! It's finally ready and working!
The only things left is to receive proper springs and final calibration of parts, because there is some little nuances left.
And here is some proofs :)


I'll calibrate lever, because as you can see in video and on this photo that lever is too near to optocoupler break line and it's reacts too early. Also it needs to make little bigger window for a 5-pin connector. But PCB fits case PERFECTLY, I'm amazed :)


PCBs, but not all..:)




Also I tried order panels and in result I received 180 SHINSANWASWITCH PCBs
So if you interested - you can order some from me :))))



And some other receivements :)
For new slider by @Doctopus , I hope it will be finalized in the nearest future and new controller at all.

 

Just bought the 3D-printer and will try to, at first, assemble it, and then to print something for SHINSANWASWITCH, because I have some changes there and will try to bring the body to ideal. So watch the next news, I hope, very soon :))

And by the way, I've already got some messages from interested in SHINSANWASWITCH people, so if you are interested too - you can mail me here :)

 

Yes, I going to upload all data about switch when I'll be able to describe all the nuances qualitatively, and there are some of them. PCB can be made at home easily, if you have all instruments and knowledge how to do it, especially if such a feature as a PCB thickness of 1mm will be observed, same with the body of switch, but I still haven't assemble my 3d-printer, so cannot say something about it here. The only thing where you almost certainly need to use aliexpress is to buy RPI-352 optocoupler and OD 6mm springs. The other elements, for sure, you can find locally in radio parts stores.

 

I finally assembled my first new 3D-printer and trying to print SHINSANWASWITCH body. I hope everything will be good :))
And it's pretty good! More info very soon :)

 

I think I finally made final calibrations of all elements of SHINSANWASWITCH, now I like how it works, I also printed full set of switches, so I think in days I will test it in my controller and literally I will endanger the performance of my controller, because it is designed so that if I put there other switch then I have now - it will ruin that system I have now, but to ruin that unattended system I made SHINSANWASWITCH, so I hope everything will be good :)
I just need now to solder other 3 boards and assemble it to my controller, May be I'll make video about it if I'll able to :)

 

I thought I made final calibrations some days ago, but that was an EPIC FAIL when I installed it to my controller yesterday. The main mistake I think was that I tested everything on small button, but when I installed it to 100mm button I found that button not even reach the lever of switch even when pressed and may be because of some modifications (button has little spacer inside to reduce press noise and may be in result it decrease button travel length) it has some differences with that small button and in result lever again needed to be modificated. And I modificated it and you can see it on video how it works :)
But in result I have little doubts about what problems will have other users with theirs buttons so I do not really know for now, but I think I'll find some answer about it.
Tomorrow I think I'll install other three switches to my controller and test it, and I hope that this variant of lever that I tested on video will work on other three buttons :)