Author Topic: Re-Create the DataHand - Thumb cluster under development. Project 75% done.  (Read 338446 times)

0 Members and 1 Guest are viewing this topic.

Offline wolfv

  • Posts: 255
Re: Re-Create the DataHand - Thumb cluster under development. Project 75% done.
« Reply #850 on: Thu, 03 December 2020, 21:51:15 »
Has anyone explored using the conductivity of the magnets as the switch?

The magnet-contacts would pit after many cycles.  Here are some work-a-rounds:
https://en.wikipedia.org/wiki/Reed_switch
https://en.wikipedia.org/wiki/Magnetic_switch
https://en.wikipedia.org/wiki/Hall-effect_sensor

Offline tinyenormous

  • Posts: 26
Re: Re-Create the DataHand - Thumb cluster under development. Project 75% done.
« Reply #851 on: Thu, 03 December 2020, 22:03:43 »
awesome. it's always great to be able to see behind the curtain!

I agree that it seems like your specific-fit version might not be viable for mass consumption. I did have a related thought though: It's hard for most people (ie me) to translate manipulated real-world elements into 3d space. Have you put any thought into designing your individual key clusters with that step in mind? Or laying out a recommended method for making the prototype->final files? Do you think the end goal for most users would be a more permanent setup?

My initial thought would be something like tracing the foot position on the plate and measuring the leg heights. Tracing the foot position could give you something scannable (and you could get 2d x-y positioning from it), measuring the height of the cluster plate above the metal plate at each leg could give you the 3d position.

OR another 'permanent setup' could even be a 3dp plate that replaces the steel one (or a drilled steel plate). New legs could be used with it and threaded into precisely located holes.

gah, it's too late to be daydreaming I suppose.

ps I spent some time looking at your py scripts. That is wild. It looks like a python version of openSCAD. Is there any way to import those models into the gui version of fusion other than stls? I was so surprised by `import adsk.core, adsk.fusion`!

Offline JesusFreke

  • Posts: 71
Re: Re-Create the DataHand - Thumb cluster under development. Project 75% done.
« Reply #852 on: Thu, 03 December 2020, 23:10:08 »
I agree that it seems like your specific-fit version might not be viable for mass consumption. I did have a related thought though: It's hard for most people (ie me) to translate manipulated real-world elements into 3d space. Have you put any thought into designing your individual key clusters with that step in mind? Or laying out a recommended method for making the prototype->final files? Do you think the end goal for most users would be a more permanent setup?

...

Yeah, that's actually a bit related to the alternate mounting mechanism I'm thinking about/working on. I actually tried to do almost exactly what you describe (tracing the feet, measuring the heights, etc.) when designing the static mount for myself. I found that in practice, I wasn't able to get a very good level of precision doing that. I did use the results of that as the "first pass" of the locations for the static mount, but it took a good probably 5-10 iterations of tweaking the positions after that to really dial it in.

My current project involves basically a big block that sits behind the cluster, with a rod sticking out the front that the cluster will be mounted to. The block has a few gearing thingies (I'm pretty sure that's the correct technical word for them! :D) that give it an adjustable angle in the "pitch" and "roll" axis,  and will (hopefully) have some sort of angle indicators so you can read off the 2 angles somewhat precisely. And then add in a z axis adjustment, and you've got 3 axis of adjustability. And then the mounts themselves can be positioned freely on a steel plate, held down with magnets, so you get x, y and yaw axis, for a full 6 axis of adjustability.

There's still a fair bit of work to go on it. I'm pretty happy with the adjustable pitch axis, and I'm currently working on the yaw axis. I'm still scratching my head on the best way to do the Z axis, haha :). Both rotational axis use worm gears, so they're (ideally) firmly held in the chosen angle and can't back-drive the gear, or otherwise wiggle around. (The yaw axis is still a bit wiggly, but definitely making progress in that area!).

But yeah, the long-term goal/idea would be to have someone use this mounting system temporarily, and let them adjust everything to their liking, and then take the angles and positions of the mounts and translate that into the static mount.

I'm thinking maybe some sort of grid contact paper on top of the steel plate to indicate where they are in the x/y/yaw axis.

257016-0257018-1


ps I spent some time looking at your py scripts. That is wild. It looks like a python version of openSCAD. Is there any way to import those models into the gui version of fusion other than stls? I was so surprised by `import adsk.core, adsk.fusion`!

Thanks! That was exactly my goal with it :D. It's called fscad. "fusion scad"? "FuSion cad"? something like that. Fusion itself provides a scripting api, but it's extremely verbose to do anything with it. So fscad is a layer on top of fusion's api, with the goal of having an openscad-like design experience, backed by the power of python and fusion :)

If you want to get set up actually developing with it, I have some setup instructions here. That's how I always run the scripts.

But if you just want to run one of the scripts manually, I think you should be able to install the fscad addin in fusion and then add the part folder in the lalboard project (e.g. <lalboard_root>/parts/cluster for the main finger cluster) as a script in fusion (Tools->Add-Ins->My Scripts->'+' button) and run it from there.

I should really add some documentation about that. Documentation has never been my strong suit. haha :)

The whole fscad thing was a bit of a parallel project to the lalboard design. I was designing the lalboard using fscad, while simultaneously developing fscad itself. I would add new functionality to fscad as I needed it for a particular part of the lalboard design.









Offline JesusFreke

  • Posts: 71
Re: Re-Create the DataHand - Thumb cluster under development. Project 75% done.
« Reply #853 on: Thu, 03 December 2020, 23:20:09 »
A somewhat better view of the "exploded" parts.

257020-0

Offline JesusFreke

  • Posts: 71
I think I've mostly scrapped that idea. I actually got a reasonably working module for that... but it's just too many parts and too complicated. It was a fun project though! Definitely learned a lot about designing gears in the process.

I think this thread is getting a bit.. long and unwieldy, so I went ahead and created a project page for this on hackaday.io. I'll mostly post updates there now. Feel free to swing by over there and leave a comment or follow the project :)