Project diva homemade controllers

Thank you. No problem for sharing.
I simply found the logos with google image. Future tone logo borders are not clean but it's ok for small size printing.
For the button graphics, i drew it myself with Inkscape. I give you the original .svg vector file and a PNG extract. It is not exactly the same size and font as the cab, but looks great.

https://drive.google.com/drive/fold...ourcekey=0-rN2_nmVpciIu7Zn89lVKmw&usp=sharing

 

I managed to get a prototype touch slider with 16 elements working to the point that it can tell me when fingers start and stop moving, and the same technique should work fine for 32 (or 48, for that matter...) and for any capacitive controller.

Current blockers for my own project:

• I want to know what components the real touch slider is using. Part numbers if possible, or if not, whether it is 32 independent sensor units, or one unit with 32 electrodes, or some combination like my own prototype... If the real part exists, maybe that can be used.
• I want to light up the touch panel like the real thing, but I can't tell from outside the machine whether it's 32 LEDs or 64 LEDs, because there is that fancy plastic stuff blocking the view. Someone with the real thing should be able to easily check. (What I can see is that there is some kind of plastic channel directing the light so that it only lights up immediately in front...)
• The real thing's LEDs light up as you touch the sensors and seem to light up twice as strong if you have two fingers on the same sensor, and I'm wondering how I'm going to do that detail, but considering ignoring it. (Every sensor is analog but I don't really know if I can afford the time to get analog readings for all sensors.)
• I am looking for diagrams of the panel and all the sizes and angles. I do have a lot of photos from the last trip to Japan but every time I find something I'm missing, having to wait until the next trip sucks.
• Figuring out how I'm going to do power. If I can somehow get LED bulbs for the main buttons then there is a question of whether I can scrape through with just the USB power supply. (If I have to use the real bulbs then obviously I'm forced to use external power as well.)

It seems like a long road, and I'm constantly worried that Hori is going to release something usable at some point, but now it's looking more like they won't, which is good for motivation.

 

I can answer some of your questions, you are a few months of work behind where I was before I got busy, but I'm motivated to finish it. Well done on progress so far.

1. The real touch slider is a piece of custom shaped PCAP film, like used on touch screens. Sectioned out to 32 sections, but the sense changes between those points. In the test menu, you'll see the threshold increase and decrease between these parts as you slide across. As a result the touch controller and PCB and all that is likely all endemic to Sega's designs, and I'm fairly convinced this same sensor is used for Chunithm for the main slider for that game. You won't be able to buy this even if you had the part number. And I've looked, no one makes a 24" x 3" PCAP film, but would for the low NRE cost of $8k from one source. In my opinion, 32 is overkill. I got mine working with 10, so there is that

2. Its a piece of acrylic side lit with standard spaced RGB Leds. The Sega board has them precisely mounted, but you can (and I have) achieve the same effect with side lighting with an RGB LED strip. Harder to assemble, trying to figure that out, but it looks good when you do it.

3. I would ignore the intensity part of the lighting of the white. My touch sensor can track my movements across the elements by turning on the white LED, but that is constant on or off (a sperate LED in my RGB lights, instead of the white being the combo of R, G, and B). It works fine and looks like the arcade to me.

4. Look a few posts back, I put up my measurements from the actual manual from the game.

5. Power will be complex for mine, I imagine you'll want to follow my method there. I need to power the LED strip, the touch control board, and the lights for the buttons. The LED strip is very bright and has its own wall wart, a USB charger can be used to power the other boards, lights from the buttons might work at 5V but maybe a seperate 12V supply should be part of this, or something that generates 5V from the 12V mains. Many ways to skin this cat.

I wouldn't worry about Hori creating a slider based controller. Its expensive, and hard to mass produce. The margins would be too small I would think for them, and the demand is not there. I mean it is, but not at their level.

Once I make more progress I'll stop being lazy and upload more pics to share. If you need help with the LED code, I'm happy to share that, since I just hacked what I found from others to work for me, and likely it will work with yours.

 

I'm actually new to this and trying to work on bulding for the controller. I really REALLY want to build one with the REAL slider for that Arcade nostalgic feel. :) so I wanna build one based from the ACTUAL Arcade button layout [with PS4 modifications. HHeres a blueprint preview [W.I.P] and I do have one question:

• Is it reall possible to make a real fuctional working touch slider for a controller? Like is there a material to mimick the clear touch slider? I felt the arcade cabinet slider and feel that I can see this being possible and know it should require LOADS of programming "00011000" of a functional sliding, I know I can use DS4 touch chipset but is too small compare to arcade.
  
  I know making this a LONG LONG project, but Im willing to put in the work and the money to this controller :)

 

At the rate I'm going, I'll be able to sell the touch slider module for controller designs, and in its final form it will be like the arcade one in terms of feel and look. Early tests say yes, its possible, all it is is capactive touch technology through a thin (1/8") piece of acrylic, side lit with LEDs.

Build your own, or follow my progress, either way its going to get done. Like I said a few posts back, I literally went from parts in hand to fully working slider within 1 week, so anything is possible. Proper motivation produces results, you can do it I'm sure. Your blueprint sketch is EXACTLY the same idea I had for my first build btw, great minds think alike

This thread is by far the most comprehensive combination of ideas for controllers I've found on the net, so its the best place to lurk. Also, its not "loads" of programming, LED control plus the slide logic algorithm takes up 120 lines of code for my arduino sketch.

 

thats a good possibily of mimicking the slider.

Oh wow! I just lookat at ur design xD Only difference is the four smaller buttons [WIP might change to analog, the smaller ones is suppose to be the directional buttons for Up, Down, Left, and Right], Im working on the blueprint for the slider, thats the only part I havent done yet.

I love looking for more inspiration and helpful ideas and share controllers. :) Community controllers designs look so beautiful, especial when some uploads with the plate design.

 

This post is really informative. It helped me a lot on designing my own controller, mainly on the look and feel part and some information on the implementation of the official controller (especially the drawing of the official controller posted by @colonelmasako. Thank you very much)

As a return, I would like to share my implementation and some of my other ideas on the sliding controller that might help others decide which types of slider to use/implement.

My implementation uses a SoftPot, which is a potentiometer that acts as a resistive touch slider/position sensor. It can be bought fairly easily online from major electronics distributors for a relatively affordable price. I chose the 500mm version mainly because it matches the size of the original PCT slider. Multi-touch (2 point) support is implemented by measuring the resistance from both sides of the potentiometer instead of using it normally as a single voltage divider (thanks to a YouTube video that makes me realize that it is possible). It can offer sensing at very high resolution (way beyond the official 32 positions, though I highly doubt that official sliders using other techniques, e.g. interpolation, to improve the sensing resolution looks like that the official slider uses 32 touchpads at 10 bit resolution so it is 15 bit, not only 32 positions) depends on the ADC you are using. One drawback of my implementation is that it requires some force to operate the SoftPot, much like the old resistive touch screens. However the force required is relatively low (0.7N to 1N on normal room temperature according to the datasheet) and unlikely affect my gameplay because I tend to use more force than necessary even on the official PCT slider to make sure good contact. Also, the friction between the finger and the SoftPot will become pretty large if too much force was applied, therefore wearing a pair of gloves is highly recommended during the gameplay. Again I do this personally when I play on the official arcade controller (and other music games such as maimai) to prevent potential injuries and reduce frictions.

Some other ideas that I abandoned but other people may find it useful:

• 32 channel PCT (or other capacitive sensing) slider. This was the most "stock" way of implementing the slider. Making a simple one could be easy and inexpensive (use, e.g. Microchip's QTouch/QMatrix PCT controller or even some micro's built-in touch sensing core). However it requires you to fabricate your own PCB/FPCB and it could be very costly if you repeatedly make bad design decisions (especially if you are not already familiar with capacitive sensing technologies), thus high R&D cost. Also a really long (500mm) capacitive slider may became really noisy due to the extended length of sensing lines.
• FTIR (Frustrated Total Internal Reflection) (Not Future Tone something ;P). Basically using a infrared-illuminated acrylic sheet as a touch panel and an infrared camera (e.g. Wiimote blob tracking camera) on the back to detect light blobs corresponding to your touch input. This could be working well when correctly implemented but since it is camera based, it requires a long distance between camera and the touch panel (50cm-1m depends on the camera's FOV) and thus not suitable for a controller-only setup (however if you are building a full-sized cabinet it may worth a try).
• FTIR, but using >32 high sensitivity photodiodes instead of a camera as the sensor. This might work but may requires lots of tuning on the sensitivity of the photodiode and the design of the cell that holds them. Also it uses a lot of ADC pins or requires one or more analog multiplexers.
• Other possible ideas that are available on NUI Group Wiki. They are generally helpful (although I didn't use any of them, it was a good brainstorming process)

 

Introducing BBCon (The Breadboard Controller), the (probably) smallest Project Diva controller over the world XDD (jk

And no it is not GIMX based. It is completely independent and can run without a PC (It does use the same technique that GIMX uses to emulate the DS4 though).

Now almost all the hard parts were done. I just need to actually design and make the case, buy all the buttons and asemble them together :D

BTW: this build does not include a slider, but adding it is trivial (I just don't want to stick the softpot on my table).

 

By the way @colonelmasako you said that SEGA did something ridiculous on the official implementation of slider. What exactly did they do? Did they split the slider into zones using multiple touch controllers? Or it was just because they added 32 electrodes to the slider?

 

The parameters of Sanwa OBSA-SP-200G spring (I can't find them anywhere so I decided to post it just in case that someone wants to find an alternative supplier for these springs. Although I already searched major hardware suppliers such as McMaster-Carr and Fastenal and no perfect match was found. Seems that sprins of this size are all way too hard for buttons (500gf/cm to kgf/cm range, where the hardest OBSA-SP was ~400gf/cm))

Code:

Length l=26.5mm=1.043"
Compressed length lcomp=7mm=0.276"
Outer diameter od=14.4mm=0.567"
Wire diameter wd=0.7mm=0.0276"
Inner diameter id=13mm=0.512"
Spring constant k=175gf/cm=0.98lbf/in (not 200. Sanwa is lying)

 

Back from lurking for way too long to finish this project of mine. There is a lot of work to go but big things happened recently on this.









@Doctopus , yeah Sega created a custom piece of PCAP with 32 sections of sense with various sense zones between those 32 sensors. In the test menu, you can see the sense value increase and decrease in increments of 0 to 1023 between these 32 sections. WAY overkill for what this game needs. Really cool, but ridiculously detailed.

 

Great case. I'm still trying to design my own. (I suck at this :()

Wait, they put 32 touchpads on the slider? Or is it just 32 electrodes on a single touchpad (the slider itself)?

 

So I kind of made large strides this weekend, here is the video of it working (enough to make the LEDs do something anyway)



Secrets of my slider design revealed, kind of just winged this but it works.







@Doctopus ,thanks. Tthe official slider appears to be a 2.5" x 21" or so piece of PCAP film behind a thick piece of smooth acrylic. The same type used in touch displays, which contains on the order of millions of sensors on tablets and phones. In the test menu, it sections this out to 32 areas along its length. As you move slowly from section to section, the sense threshold goes from 0 to 1023 as you cross each one. So really you have resolution of 32 x 1024 = 32,768. Like I said, WAY overkill. Makes it so that you can do things with even the smallest movements of your fingers I guess, but for this game play kind of crazy. My design uses 10 and so far its working pretty well from what I can tell.

 

As far as my "research" goes, they don't. There are typically only hundreds of electrodes (discrete sensors) in a smart phone or tablet, and achieved "millions of pixels" precision through interpolation (yes, the same technique used on scaling photos). But I have to admit that I didn't see anything that scales 1024 times on one dimension only through interpolation (phones and tablets did that because they are actually 2D sensors and each dimension only scaled up by ~100x maximum. 32 synchronized touchpads on a single panel sounds doable and matches what the test menu says (i.e.32*1024 positions), but still crazy). That's one of the reason why I'm curious about how they implemented the slider.

The other reason, as I mentioned earlier, is the noise. You are basically build an antenna and attract all kinds of noises from the environment if the traces for the electrodes are too long (even if the touch controller is on the center of the slider, the longest trace will still be 250mm minimum. FYI one of the largest vendor of capacitive touch controller, Microchip, suggests that the traces of electrodes for their touch controller should not exceed 150mm or signal quality will start to degrade, result in poor performance). Again if they built 32 touchpads this shouldn't be a problem anymore because they can place the controller near the electrodes (that's also where the multiple controllers idea poped up, but I only ever think about 4 controllers and believe that's already a lot so this still sounds crazy lol)

So yeah if someone have a photo of the internals it would be awesome

BTW @colonelmasako I also have some questions about your implementation. What are you using to drive the touch sensor (Arduino? external controller? any third-party libraries?)? Is your sensor glove-safe (i.e. penetrates non-touchscreen-friendly gloves) like the official one?

 

Did more things. Doing so found many issues to resolve but its built now, and works ok. Slider needs more work.





@Doctopus , my implementation is pretty simple. I'm using the Bare Conductive arduino board to interface to painted touch sensors. I took an existing project from them and modified it to act as a multi element slider design. As is, I doubt it is senstive enough to penetrate the acrylic to detect gloves. I haven't actually tried that on the official one, if that does work I'm even more impressed Sega.

My code, which I'll eventually share once its really solid, documents all of the things needed to make it do the slider and light control. I'm taking advantage of funky tricks to do multi-threading in a processor that can't multi-thread. The logic I used to differentiate between normal slides and double slides in any combo is insane too. But it works

 

Here's an action video of it. After this I'm probably done spamming the thread with my photos, sorry folks. I'll of course answer questions about it as it comes. Its not 100%, more like 95%, where I want it, for slides.



I have to do more work to make double slides in one direction work. Any other combo works very consistently now. Keeps you busy I guess :p

 

That touch sliders looks amazing.

I've been lurking this thread for months, but haven't made an account until now. I'm making my own controller, I've been waiting a few weeks for the buttons to ship but they should be here in the next day or so.

It's going to be a laser cut acrylic case with a Yowane Haku theme. Ultimately I'll do artwork on the inside with frosted vinyl graphics and some commissioned artwork for the main buttons. I have my own laser cutter, so hopefully I can prototype out some different things. I've been using the Hori FT Mini controller for months waiting for this, so super excited.

I wanted to see if you guys had any input; I'm going to be using the Hori FT Controller as the base controller, most of it really straight forward. But what I'm unsure is the placement of the analog sticks. I see a lot of people prefer buttons (usually four) to work inplace of the joy sticks for slides.

Would you guys think it would be best to mount the joysticks below the buttons, above the buttons or higher up (Where the touchpad on an arcade machine would be) part of me thinks having it where the touch pad would be makes sense for that crossover between arcade/homepad muscle memory but I really have no idea. I'm not sure if that would be too clumsy to reliable hit those analog sticks and move between the buttons vs. how the touchpad actually works at the arcade which is so easy. What have you guys found in your experience works best for placement?

Edit:

Got my buttons early and put together a quick prototype to try different analog positions and stuff.

It's great having the giant buttons. I notice that a poster prior said they dampened the buttons because they were going down too far? I feel like I have this issue too and would like to know what they did exactly to fix it.

 

Yes i have dampened my buttons. I assume that you have the same kind of generics as mine and not sanwa.

I felt that the button had too much room for movement but also was not touching the bottom (the black part) when press so the button was not stable during the hold notes. All the weight was on the switch.
To fix that, I added some foam stickers that you can find in home supplies stores. I sticked 4 of those under each plunger. Now the plunger with the foam moves just enough to activate the switch and makes stable contact with the black part underneath. I don't remerber what thickness i used, the best is to buy different types and try.

Also, consider buying the 200g sanwa springs, it makes a HUGE difference.

I hope it is understadable. I don't have picture to show you right now.