Spiderclaw V3 (aka Ultracortex)

I'm starting this thread to begin the open-source discussion surrounding the design of the Ultracortex, the V3 OpenBCI 3D-printable EEG headset. It will be the first headset design to work with the recently shipped V3 OpenBCI boards. Today I'll be reposting my original post on the V1 and V2 headset designs. This post somehow got lost in the chaos of updating our website, but I'm brushing off the dust and re-posting today. Still need to update the .STL, .OBJ, and Maya file download links.

I hope to use this thread for an ongoing discussion surrounding the V3 design. 

My current R&D involves:
- Following the work done by profbink to replicate his 3D-printable prosthetic hand. I believe that the mechanics of tension, using trimmer line or fishing line and elastics, can be applied to the arms of the V3 headset. I hope that by doing so we can create a uniform tension along each of the "arms", to create a comfortable and secure connection between the headset and the scalp. Currently, the biggest design flaws of the V2 design are: 1) too many moving parts 2) isolated tension at each of the "arm segment" joints, which results in buckling of the arms.

More soon...
«1

Comments

  • Just finished updating the .stl, .obj, and maya files for headsets v1 and v2. They can be found in my latest community page post, along with some instructions on getting started. Reminder: these designs don't work with the latest form factor of the hardware. V3 (coming soon) will!
  • Do you want to design it in a 3D graphics program again or ina  CAD with assemblies and constraints?
  • My plan was to design it in Maya, because it's the program with which I'm most comfortable. Do you recommend something else? I'm up for suggestions. 

    I have reservations with embarking on using another software, just because I've done so much with Maya and feel very comfortable with it. That being said, I'm definitely up for exploring alternatives that better suit an open-source design process.
  • It's a project that would suggest a parametric CAD program able to to assemblies of multiple parts that have constraints referencing other parts.
    (holes matching, faces being coplanar, holes being the same diameter that can be switched between e.g. a metric and an imperial version and parametized for hat-size=head circumfence)
    Problem is, there are so many and of the few people here that have one, hardly two will use the same one.

    FreeCAD comes to mind with it's new sketch module but I haven't used it much yet.

    At the end of the day a 3d graphics program does get the job done but you have limitations.
    e.g. you are limited to polygons in interchangable formats.
    You need to maintain each length of an arm separately even as it's the same arm, just with one parameter being different.
    You have no export for drawings, part list and documentation.
    You don't have physical units (e.g. from the eagle-layout of the PCB). Your base-unit is "1" and you need to tell people if it's 1 mm, 1" or 1 mil.
    That a user can't export the assembly for any custom head circumfence.
    On the other hand polyon nets with no parameters and no constraints are something everyone can edit in dozens of programs.
  • Oops, I thought I'd already responded to this topic, but apparently I'd just saved it as a draft...

    I completely agree with Marcus, a parametric CAD program would be better. Apart from the benefits mentioned, you'd also be able to link the design to (parametric) digital head models and do product fit verification directly in CAD.
  • What are the dimensions of V1? I know it's obolete but I'd like to use it as the basis for a headset I'm working on (and share the results if anyone's interested).

    Cheers,
    Bogdan
  • edited March 2015
    How about modeling it in OnShape, which is a new cloud-based CAD modeling tool that allows for collaborative teamwork on publicly accessible CAD models. The approach is similar to GitHub where unlimited public models are free to use and only private models need to be paid for if a limit is reached. Just read about it today and got access to the public beta version. It seems truly sophisticated featuring assemblies and constraints - not very surprising because it seems to have been built by SolidWorks veterans. Also, it features a built-in version control system including branching and merging of model states. Might really be a cool tool to work on the V3 of the headware to keep it truly collaborative and accessible to everyone?

    Ah, almost forgot to post the link:   http://onshape.com

    http://techcrunch.com/2015/03/09/onshape-launches-mother-of-all-products/ 

    $64 million in seed funding, and offered free to users(!)



  • wjcroftwjcroft Mount Shasta, CA
    I think Conor is overseas at the Brain Tech conference, but looking through recent twitter posts saw this modular flat and flexible V3 design,




    image

    Looks cool, perhaps Conor will add a few comments here...

  • edited April 2015
    Hi everyone,



    Please forgive the radio silence. I was traveling through Israel and London, evangelizing open-source BCI! I'm back in action now and ramping up again on Ultracortex R&D.


    We've been collaborating with Aaron (aka Threeform), a very talented industrial designer based in NYC. We spent a few days drafting and sketching on paper and decided to try out a rigid design, manually modeled for high rez mesh of his head. He precisely followed the 10-20 setup steps to locate (with high accuracy) the basic 10-20 nodes on his own head. See below.

    image


    Then based on the nodes identified, he modeled a rigid model in 2 halves that fit his head with a 5mm tolerance in the z-axis (perpendicular to the surface), as seen below.


    image


    So I know what you're thinking.. How do I get one for my head?! What we're trying to do is find a crew of 50-100 individuals that will pay for the up-front costs of working with industrial designers to manually model headsets based on scans/images/measurements of their heads, similar to the technique that Aaron used for the design above. In the process, we'll take detailed notes on the artistic/engineering modeling process and then replicate the process algorithmically with code. The end goal is to have an open-source platform where anyone can upload images/scans to create a 3D mesh of their head, click a button, and then a printable Ultracortex .stl is rendered specific to the individual. This may be something we Kickstarter or crowdfund on a small scale to get the ball rolling. We mainly need a small group of people who are willing to pay a decent chunk of change to have one of the first manually designed custom headsets to finance the R&D of the future automation process.


    RE mikeschaekermann & OnShape: I think this idea is really interesting. I've never worked with OnShape, but the concept is amazing. At this point, the way I envision the future of Spiderclaw/Ultracortex unfolding is a fission between:

    1) the "one-size-fits-all," flexible design (possibly modeled and iterated on in OnShape)

    2) this new "your-size-fits-you" approach, which I think has more potential in the long run


    Also, I'd like to note that the nodes (for electrodes) that you see in the design above are designed to work with the snap-in 3D-printed electrodes that we've been prototyping. Theoretically people could experiment with different electrode designs (active/passive/wet/dry) by sticking to a specific snap-in form factor.


    image

    image

    image


    Would love to hear your feedback.


    -Conor

  • edited April 2015
    Looks like you guys are doing great work and really have the ball rolling!  A couple comments...

    I think a little more investigation into the metrics of one-size-fits-all or maybe 2-3 standard sizes.  Perhaps it would be good to define what sort of precision we're looking for with electrode placement and also what others think is reasonable.  I think it would make things a LOT simpler if we could avoid the head scans and parametrics in the design.

    Another thought is that 3D printing is not the only way to go.  Some consideration in the design might be taken for CNC construction, and I'd be happy to help with that.  If we can come up with a 2-2.5D type design it might open up production capabilities to a lot more people and construction methods.

    If we do need a custom design for each head then it wouldn't be terribly difficult to do it parametrically based on a single head measurement or hat size. Things get much more complicated with fully 3D designs though.
  • Hello,

    I'm glad to see that the ultracortex is moving forward :) I like the design so far, especially the last sketches with the springs; I've been trying to come up with such mechanism myself (using flexible filament??). The standard snap-in will help a lot to test different electrodes.

    However I'm not entirely convince by the "your-size-fits-you" approach. Maybe in the *very* long run it's the best solution, but I second @JakeStew


    1. It'll be less practical to manufacture -- "scan *accurately* your head and generate the model and print yourself" versus "directly download" + the possibility to do mass production

    2. Do we need that level of accuracy? From what I've seen, medical company typically sell 3 sizes (small / medium / large) and... that's okay, even with up to 64 positions. In our lab, there's few situations where we'd need something else than the "M" cap, and it's only because with small heads it's more difficult for the electrodes to reach the scalp and we need to add more gel. With so little electrodes as we have currently with OpenBCI, knowing the spatial resolution of EEG and everything, I'm not entirely sure the extra effort to scan the head is worth it. Even when accuracy is required, like with motor imagery, in fact you'd really need a scan of your *encephalon*, ie and IRM to get the real data.

    3... maybe you won't be able to share your headset so easily if it has been tuned for your head

    Note that "one-size-fits-all" and  "your-size-fits-you" are not incompatible, the latter can give precious and empirical data for the former -- I don't know if there has been such recollection in the past, most probably. To sum up, while I see the interest, I'm not sure I would back up this first run. I'm more interested in a headset that I could use to make people around me discover EEG and play with OpenBCI than a customized headset. On the other hand, if instead of sizes you propose to shape particular electrodes montages -- ie one for P300, one for motor imagery, one for neurofeedback, one less-sexy-but-with-the-whole-64-positions-available, you'll have my money ;)

    Also the customized headset could be less about anatomy and more about design. If you manage to make a fashion accessory out of a EEG headset, you'll have won the match -- good point to have talented designers with you!

    No matter what, please share your progress, OpenBCI + active circuit + home-made electrodes + customizable headset == full open EEG plateform :)
  • I would just like to add to jfrey's final message.  Share your progress and I'll be better able to consider up-fronting.  Spring is in the air.
  • I don't mean to be rude but this is the worst product design idea I've heard in years. I'm just saying this because I want to help. Can you imagine what it would be like if headphones were like this? Thankfully, they're not---they use designs that are based on common features of all ears. At the very worst, you could have a couple of versions but what you're proposing sounds so ridiculous and unpractical that no one will get one. I'm someone who is actually interested in BCI and I still wouldn't go through the trouble---can you imagine what your average Joe would do?

    Either way, good luck!
  • bryanbeusbryanbeus Spanish Fork, UT
    I like the idea of having an automated process, where I upload my information to a website and the whole thing comes ready-to-go in the mail, with as little need on your end for direct employee involvement.

    I do think that this is such an extremely cutting-edge business/technology model that you may not find an audience to support the costs for a couple more years.

    But when the number of "average consumer" type of person finally shows up, you'd be poised to make a huge difference for people who are interested and/or need the techn, as well as have a profitable business.

    Private message me for help! I'm interested to get involved.

    I finish my first novel for the national market and it is going to press (through Shadow Mountain Publishing) this September. I'll be touring the country to promote my book next year, and I would like to have OpenBCI be a part of my presentation to schools.

    This summer I'm planning on putting together my presentation. So if you can find time to contact me, that would be great so that I can get started in the right direction!
  • On the github repository the Spiderclaw V3 is announced for Spring. As summer is not far away I would like to ask if there are any news of the progress.

    As the openBCI board without a proper headset is pretty useless (as long as you want to measure EEG data) I would like to ask which options I have to get my board working. I have a 3D printer and CAD software over here. Is it  possible to modify the existing design of the V2? Of course I also would consider to purchase an out of the box solution but being a PhD Student tousands of dollars are far beyond my budget.

    As I am planning to use the BCI board as an important aspect of my presentation next month I really need some short term solution.

    Thanks for any advice.
  • nekrodezynfekatornekrodezynfekator Poznan, Poland
    Which electrodes from 10-20 standard V3 can be mounted on Spiderclaw? I am especially interested in electrodes O1 and O2.
  • wjcroftwjcroft Mount Shasta, CA
    @elias , hi. I merged your post into this existing thread. See the thread below for some various electrode 'cap' or band solutions.

    [May 4 post]

    @nekrodezynfekator , my impression is that the Ultracortex will support all the standard 10-20 sites. However since the V3 headset is still under development, best to use an interim solution as suggested above.

    William
  • nekrodezynfekatornekrodezynfekator Poznan, Poland
    So we are not going to know the approximate date of the premiere?
  • wjcroftwjcroft Mount Shasta, CA
    edited May 2015
    Conor and Joel are currently at the NeuroGaming Conference, MakerCon, etc. So may not be watching this thread as frequently. If you are primarily using O1 O2, Have you considered the velcro band and dry sensors mentioned on May 5 post above? Very convenient and available immediately. Also see the $30~ BrainNet 10-20 grid elastic mesh.

  • edited July 2015
    If your print volume is big enough, just print the front2 and back2. The other 4 files are in case you need to split the front and back in half, if your print volume is too small.

    You'll also have to print enough of the electrode mount components to assemble the set (find them here). 

    Lastly, I used these springs in my assembly:
    But I would try these ones, if I were you (the two I used were a tad too weak and too strong):

    Lastly, use these nuts / screws to connect the FRI electrodes to stripped wires (I stripped some ESK and touch-proof adapter cables) and the electrode mount:

    image

    Like this:

    image

    Here's what the headset looks like on the inside, when fully assembled:

    image

    The stiff springs make good stabilizing (but still somewhat springy) structural connections to the scalp, while the soft springs work well to alleviate pressure through the spikey electrodes.The springs at T3 and T4 are the softer springs. The ones at Fp1 and Fp2 and FpZ are the strong springs.

    Good luck!
  • nekrodezynfekatornekrodezynfekator Poznan, Poland
    Image is broken ;(
  • nekrodezynfekatornekrodezynfekator Poznan, Poland
    It is possible to use electrodes from electrodes starter kit?
  • You could try to force it, but this design is not intended to work with them. It is designed to work with electrodes from FRI (listed above), which are cheap, dry, and they work well!
  • edited August 2015
    Uploaded to Shapeways for printing:
    http://shpws.me/JN5R (frame front $37.78) & http://shpws.me/JN5H (frame back $53.85)

    It's reporting potential issues with thickness in some areas however I don't think that's correct. I've contacted Shapeways to find out if printing will be successful. They appear to be the cheaper option over Sculpteo which estimates costs over $420.

    Update: I uploaded to 3D Systems QuickParts to see an estimate and can't get the total cost down below $1000, I'm not sure which process or material was used by Conor but more info might be helpful. Perhaps a group buy somewhere will help reduce costs?
    Regards,
    Corey
  • For reference this is the total cost from Sculpteo with 1 complete frame and 22 (1 extra) sets of small parts.

    image
  • Hey @Coreygo, so far I've printed everything out of our Rep2 printers. It has the precision necessary to get the job done, but I'm not sure if that is the case for all desktop printers. If you end up going with shapeways or another "printing as manufacturing" technique, it will probably be 400+. 3D Hubs might be your best bet if you want to keep the cost down, though you won't get the same quality on parts.

    We are hoping to doing another crowdfunding campaign soon to get the #s up, and hopefully get some economy of scale on a run of 100-500 headsets. Stay posted. :D

    - Conor
  • I will be the first buyer :)
  • Nice, thanks Conor! I'm holding off and getting more quotes from different options. Right now Sculpteo is the cheapest and best third party I've found -- quite crazy considering they're in France. Any chance the crowdfunding campaign will see headware shipped before 2016? If so I'll definitely wait for that.
Sign In or Register to comment.