Re: Flex Dance: open-source dance pad | Hacker News Reader

Re: Flex Dance: open-source dance pad

reflex.dance

113 pointskadoban37 comments

I've owned a Cobalt Flux, Blueshark and two Impact Arcades on top of a host of soft pads. It's a sad state seeing the current offerings of pads. Basically the only company manufactured pad is Stepmania X with long wait lists and limited supply or very DIY solutions like the L-Tek pads or used stuff.

I want to start playing again but I literally can't get any decent pad in Europe, which makes me sad. Almost easier to just buy an older cabinet.

Why do you say the L-Tek is DIY? It's a tad DIY if you want the bar, and also if you are a really high end player you might need to mod it with the arduino hack, but out of the box it's good for most players.

The StepManiaX stage isn't a wait list, it's the opposite with first come first serve and literally sells out in ~20 seconds every time a batch goes on sale. I have one and can attest it's absolutely spot on and worth the hassle though.

Because it uses the older type capacitive sensors and not the pressure sensors that Kyle Ward stole off Brittany's pads that ended up being Rhythm Horizon. It's kinda the industry staple these days as you can adjust the sensitivity to your play style in software without having to touch the hardware.

I'm a pretty high level player (passes in 15 blocks and 3 stars in 12), although I'd be content with any pad that I can just play for fun on that doesn't require modding or constant fixing.

Like the blueshark, the sensitivity in Ltek pads have to be adjusted by adding material between the panels and the sensors and they're pretty prone to either breaking or just losing their sensitivity, so they require basically weekly tune ups and they're never really perfect.

Ive been looking for a pad with a bar and pressure sensors for about 1k USD. I'd buy the Stepmania X pads if the shipping costs weren't absolutely bonkers.

Do you live outside of the US? Within the US I'd argue the smx shipping is quite reasonable. The crate weighs 255 lbs, is custom made and it's delivered by freight. Within the US the total is about $1800-1900. For a perfect, modern arcade quality stage that can be configured anyway you can imagine it's well worth it.

In my case, some of the songs that I used to play in arcades in Asia use all 9 directions. The L-Tek pads only have center + 4 arrows, but none of their products has all 9 of them.

EDIT: Here's pictures of the L-Tek: https://www.vierpfeile.de/phpBB3/viewtopic.php?f=14&t=7709

It also looks like the sensors inside aren't based on piezo crystals and rubber like the arcade machines.

I'm in the same position. I'm looking for pads with all 9 directions for those B4U songs. Let's hope a kind stranger on HN will know of a new and upcoming pad manufacturer...

This is amazing, but.... I just wanted to point out, the latest "dance pads" are basically a giant touch screens so they are far more flexible than just 3x3 buttons.

https://www.youtube.com/watch?v=kksXU42lEBY

love the "no camera" sign right in the middle :P

I feel the gameplay becomes very different with these type of dance pads. The fixed button ones are more like Guitar Hero with focus on speed than actual dancing.

If you want a game that treats the pad as a 3x3 grid that's entire possible with a trackpad.

Wow, that's awesome! I never considered the possibility of such a large input surface... And the way the LEDs light up around the movements is really nicely done.

> love the "no camera" sign right in the middle :P

I'm not sure why this is here, you would think it would be good PR if it went viral (like this video possibly did). Even I now want to try my hand at replicating it.

Someone in the YouTube comments suggested it was asking people not to record strangers without their permission.

It's even funnier considering the machine actually has a camera built in (eg. https://youtu.be/ya_zzEMHCoc). I believe it also has YouTube upload built it but I could be wrong on that.

Many years ago, I used to have a Mayflash TX4000 dance pad [1]. It had a simpler sensor construction, similar to how regular gamepad buttons work. Each panel flexed very slightly upward at rest, and had a foil layer on the bottom, which would make contact when the panel was pressed down. This worked surprisingly well, even for "heel-toe" playing (and didn't seem to deteriorate over a few years of quite heavy use). Though, for really fast songs the "stickiness" would become a problem, and you'd start getting a few errant misses. But below, say, ITG difficulty 11 it was fine.

I remember reading about more expensive pads like the Cobalt Flux using more sophisticated sensors, similar to (or exactly the same as?) what you'd find in an arcade machine. Are the sensors in this design similar to what that pad used?

1: http://www.mayflash.com/Products/DANCEPADS/TX4000.html

The CF design, which I own and have been meaning to use again, is actually surprisingly simple(albeit with good materials) and relies on panel flex like the Mayflash, but it uses layers of clear lexan screwed into the platform rather than a steel construction. Both are a lot different from the DDR arcade design which uses a deeper stage with a well for each panel and a sensor lining each side of each well(so, 16 sensors for a 4-panel stage). On an arcade stage, the player can rely a lot more on bounce since the panels don't exhibit "bottoming out". They perform quite a bit differently in expert play and I never got totally used to the feel of the CF, but I would be happy to use it again as a cardio device.

Cobalt Flux pads actually have the simplest sensors imaginable: on the bottom of the pad is a large steel plate, connected to the control box. Under each arrow is another steel plate, also connected to the control box using a wire and separated from the base plate by 4 velcro patches, one on each corner. Step on an arrow and the plates make electrical contact.

The reason these pads worked so well compared to other home pads is that they used more robust materials: steel instead of aluminum and polycarbonate instead of acrylic. It still had issues though, the wiring and wood screws in particular.

Arcade pads use rubber sensors: press the arrow and the sensor get squished, connecting two carbon tracks inside it. These were notably finicky, at least for the early models. They progressively became more sensitive as they wear out, before they fail and get stuck.

The linked pad uses load cells, which allow to adjust the sensitivity. I don't know how it works in practice but in theory that should be great for high level play.

As for the "TX" series of pad, they come from a long line of questionable quality pads. Some were great, other not so much, but it was mostly due to the build quality and materials rather than sensor design.

I built something like this a while ago, but mine is not nearly as nice. I etched an interlaced finger pattern on a big copper clad board, then put a tented piece of metal overtop of that, then some lexan sheets with a graphic inbetween them on top, so when you step on the arrow, it squishes the metal plate on to the circuit board and closes the circuit. It works remarkably well. And then the whole thing is powered by an Arduino that pretends to be a keyboard: https://github.com/jwalton/arduino-dancepad

I built one too, made out of wood (particle board). Arrows are spring loaded and when you step on it, it bottoms out on a metal bar, which make contact with another metal bar on the bottom side of the arrow. Ugly, noisy, but worked really well.

For the electronics I just took a cheap gamepad and soldered the wires from the sensors to the button traces. Cheaper is usually better here. The gamepad was still functional, so I could use it to select songs and things like that.

We built something similar to that in the early 2000s as well, although we used copper pipe hanging strap for the connectors, a J-PAC from Ultimarc for the PC interface and some plastic tubing for pushing the arrows away from the connectors. It worked pretty well too.

Oh yeah, this sounds familiar. Was it also on the DDR subreddits a while ago?

I meant to ask the first time I saw it -- if I'm understanding this correctly, the entire frame consists of parts that can be either printed or just sourced off-the-shelf? I've personally had "pad build" on my backlog, and have been playing around with modifications of the Cobalt Flux design on paper ever since a failed attempt to put one on risers (apartment noise issues...), but wood- and/or metalworking have generally been barriers to actual implementation.

Maintaining it up until that point is actually also how I DIY'd my first controller (to replace a broken USB adapter), and finally learned to solder.

(It also just became less of an issue between living in a region with a few different cabinets I could drive to and the L-Tek pad releasing around a year or so later.)

Yeah there's two projects I know of that have been making their rounds on reddit and discord. This one just released today, and seems like the more interesting one to me. There's also a 9-panel project called Rhythm Horizon (used to be called Impulse I think?) that's I think going to be sold fully assembled.

I've been looking through the plans, yeah it seems like all the parts so far are either 3d printed or can be bought "off the shelf". Only exception I found so far is seems like you have to solder on a small number of components on a pcb the way they've suggested to order it.

I am still working through the sourcing though, so take with a grain of salt.

Impulse was sold off to Rhythm Horizon and Reflex I think is the spiritual successor to what Impulse was, by the same people.

I've been out of DDR for a while. I still have a Cobalt Flux at my parents home. That seemed to hit the perfect sweet spot of size, weight, and durability, I've literally had it for a decade. How does this compare?

I think those are worth a lot because there isn't a comparable thing in the market anymore. At least there wasn't when I looked a few years ago.

These pads are considered about on par with Cobalt Flux: https://www.maty-taneczne.pl/

I've never used one myself, just going off what the internet says about them. There are lots of reviews of them on YouTube (people either call them "l-tek" or "polish pads")

I can't speak much to it, as I don't play - my girlfriend does, and finding a pad for her was difficult. There's just nothing out there. Couple of dead projects, the L-Tek, and some total junk pads.

I ended up with the L-Tek one you linked. The buttons seemed awfully futzy to me, and she says they're ok but not great - plus her toes hit against the edges of the raised non-button squares and hurt.

Cobalt Flux had 9 active tiles, L-Tek only has 5.

Cannot find a price average for the full build. I’d be so interested in building one…

Pretty cool project, the LEDs are pretty nice!

In case the author or someone else knowledgeable reads this, is that application for the load cells used for sensing are able to handle? The load cells are rated for 50 kg (~110 lbs), and even if there are four load cells, the dynamic load might not be evenly distributed. It's also likely that the 50 kg limit is a static load, not a dynamic one.

The manufacturer's site lists the max load as 150% full scale, so there is a little room there. (And exceeding it probably just makes the cell go out of calibration, which might not be a big deal for a dance pad?)

Yeah I'm curious as to what is the failure mode for these sensors when they're run out of spec. As you point out, if the sensor is just returning an inaccurate it's probably fine as you're only interested in a relative load value (loaded vs unloaded), not an accurate reading.

I don't have actual data with regards to this, but the instantaneous force applied to the load cells would likely exceed their spec in the case of a heavier adult jumping on the arrows or otherwise hitting them off center with more force than placing a static load on it.

Thank you kadoban for posting this :) I've been meaning to research what to buy now that my old Cobalt is breaking and they aren't selling anymore.

While this does look more DYI than I would have hoped for, I'm sure that having a good and open design will lead to more manufacturers entering the market again.

Bought two soft pads, one was DOA. The other was dead after a few months. The hard ones basically don't exist as they cost too much to ship.

Some entreprising industrial designer needs to do a good flat-pack one reconfigurable, ideally with nesting during shipping for higher density, and with multiple levels of upgrades. It would probably make bank on crowdfunding. Maybe on the ultra-budget version, you just buy the shell and slide in a soft pad.

It's best to use foam pads if you're getting started. I wanted to get metal ones one day (as I've worn through a few foam pads). I'm in an apartment right now, so they're boxed up, but I look forward to moving away from the city next year and being able to get them back out again.

It would be cool to build metal pads like this, but I'm super lazy and will probably just buy some.

I prefer foam pads myself because they’re more reliable than the flimsy soft pads but don’t hurt your back to use like the metal pads.

If you're looking for a simpler way to build a solid DDR pad I did this mod to a soft pad a few days ago and it worked well, even for someone terrible at physical construction like me: https://www.angelfire.com/pro2/softpadmod

Yeah I did this exact mod with a madcatz 5 button pump pad and picked up Xbox to USB adapter to play StepF2. There are so few pump pads. I would have loved to have done this with a 9 button pad for ddr/piu but the main issue is that the buttons are different sizes in those games.

One thing I wish I'd done differently is just used MDF instead of plywood, my piece ended up being slightly warped which makes it annoying to use on hardwood, but it plays fine if it's sitting on carpet.

I had a soft pad similar to this - take care that the pad is able to register arbitrary combinations of left/right/up/down simultaneously. Mine was not, so I ended up using it rotated 45˚ with the auxiliary buttons remapped, which avoided the problem. It would be a shame to do all this work and not notice the issue until you start playing with jumps!

Reply on news.ycombinator.com