Page 298 of 330 FirstFirst ... 288294295296297298299300301302308 ... LastLast
Results 4,456 to 4,470 of 4943

Click here to go to the first staff post in this thread.
Thread: DIY Head Tracker For A Tenner

  1. #4456
    They're in stock, normally sent out within two working days - rest depends on where you are of course

  2. #4457
    Click image for larger version. 

Name:	Screen Shot 2015-11-20 at 23.38.44.png 
Views:	58 
Size:	405.6 KB 
ID:	76837

    EDTracker-Mac Alpha 3 is available for testing.

    This version has all the basic functionality I planned adding for the first release version.

    You can inspect the output, change the orientation, alter the scaling and smoothing, recentre, recalculate gyro bias, and perform manual magnetometer calibration.

    Automatic calibration and flashing of firmware are not (yet?) supported.

    Bug reports to my PM, please! :-)

    How to manually calibrate:

    Fire up the program, select your serial port and hit open. I've had some issues connecting the EDTracker, but it's infrequent now. If it says connected but there's no data populating the Data From Serial column, try hitting the 'Hello' button. Failing that, close and re-open.

    Once the serial is connected the program will start collecting magnetometer data and plotting it. Take the EDTracker to its working position, hit clear, then start rotating it through as many orientations as you can to trace out an ellipsoid on the plot. When satisfied, hit pause to stop additional (potentially errant if you now put down the EDTracker) points being recorded.

    Use the offset, scaling and rotation controls to position the yellow ball of dots closely around the ideal sphere (turquoise). If you have an extant setup, hitting 'Current Values' will transfer those to the prototype settings to save you some time. 'Reset Values' will take you back to identity (ie. default settings).

    Once happy with the alignment of the yellow and blue balls, hit 'Send Calibration Data to EDTracker'. Select your orientation by clicking 'Mounting Orientation' until it describes your mounting position.

    Don the EDTracker and hit 'Recentre' then, sitting very still, hit 'Autocalibrate Bias'.

    Once that's all done, try moving around and seeing how it responds on the HID graphs. Scaling and Smoothing can be applied now.

    If you're seeing yaw being very sensitive and then springing back, try fiddling with the 'Blue' offset setting (up or down) and re-sending the calibration data until it improves.


    The major problem I've encountered so far has been memory leaks with the ARC system in the current version of Xcode. I've completed replaced the matrix logic from the previous version with pure C versions (managing my own memory), but there's still a medium-paced leak that's caused by me updating the joystick position sliders(!)

    I'll need to figure that one out before the release version, but unless there are any serious problems reported with this alpha, I'll move onto beta which will be a complete refactor.

    Once at release I'll publish the source and pass the whole shebang to Rob and the team.

    Cheers,

    BW.
    Attached Files Attached Files

  3. #4458
    Thanks brumster
    I'm in Scotland
    Product looks fantastic
    looking forward to trying it out

    Cmdr Simpleye

    DBX/Serenity's Call/Near Bubble Exploration------Asp Explorer/Navigator of Solitude/Deep Space Exploration
    Vulture/Tenacity/Fast Killer
    Fer-de-Lance/Ax Wielder/Anti-Thargoid
    Python/The Letter of Marque/Privateer
    Type 9/Dreadnought Buffalo/ Slow Killer Tank
    Anaconda/LongStrider/Battle Wagon

  4. #4459
    Originally Posted by brumster View Post (Source)
    Hi Hoff, I doubt it's your soldering, they usually work or they don't. Are the PCBs (well, the MPU board in particular) flat to the plane of your head and it's no flapping about or loose? The blobs for you magnetometer calibration look very tightly packed together, I would turn down the sensitivity slider before calibrating it again - we do recommend that you move the slider just to the point where the mag readings number starts to increment, then back it off just to the point where it stops again (this is with the device still) - then do your mag calibration at that sensitivity level.

    Maybe, for some reason, the mag you have is more sensitive than most... although it'd be the first case I've heard of it :-S !!
    Thanks for the help brumster, im gonna skip on desolder the unit then. All the parts very firmly soldered to the PCB, so no flapping or anything, the only thing that is loose is the unit itself until i screw the lid on the box i got from your website (hammond box). When the lid is on, there is nothing moving. I have tried multiple reflashes and recalibrations now, but will try some more tomorrow, been going at it for days now.

    For me it seems like the MPU board has some serious defect on the sensetivity levels of yaw, anyone know what https://hobbycomponents.com/ policy is on units like this? Do they just ship me a new one if i give them some pictrues/videos etc. Or do i need to return the whole kit ?

    Uploaded this video to youtube that shows how the device reacts to different movement, up and down works fine, but to the sides is just not working at all. Kinda sad, these EDtrackers seems really accurate if they just work.
    Here is the video:



    Gratz on the new site, looks good. Might just have to order a EDtracker pro if i cant get this to work.

  5. #4460
    I know you start to doubt anything at this point but let me assure you it is not the arduino or the circuit board if it goes like this. 99% guess it is the calibration and 1% wrong firmware.

    As i wrote earlier it may be a magnetic field from computer, monitor, head set, anything

  6. #4461
    Just went to check the EDtracker.co.uk site as the DNS had not propagated to me last night. When I do it tries to make me download a shockwave file. I do not have flash installed, so something is misconfigured.

    Click image for larger version. 

Name:	edtrackerproweberror.JPG 
Views:	38 
Size:	90.4 KB 
ID:	76919

    Other than that, WOOT! Can't wait to get the updates.

    "You're doing it wrong". No, we are having Fun OUR way.
    ~-~ The A-Team Site ~-~

  7. #4462
    Shockwave?! Let us not talk about that ancient 23rd century stuff.

  8. #4463
    Originally Posted by Cataclysm274 View Post (Source)
    Shockwave?! Let us not talk about that ancient 23rd century stuff.
    Technically it is trying to show me a Youtube Video, but instead, since I don't have flash, it says D/L it instead of "You need Flash".

    But if it did not try to make me download it, I could still see the video, as I use NOFLASH addon that automatically converts flash requests to Youtube to the HTML5 version as I never use flash since its poo

    "You're doing it wrong". No, we are having Fun OUR way.
    ~-~ The A-Team Site ~-~

  9. #4464
    Totally off topic but all platforms these days support HTML5 video players. No need to use/push flash from a website any more. In case that your browser doesn't do HTML5, i seriously advise you to upgrade. Flash is history. Shockwave even more so.

    http://thenextweb.com/apps/2015/09/0...nks-to-google/

  10. #4465
    Your link in the latest news section contains a comma: http://www.edtracker.co.uk,/

    just a heads up

  11. #4466
    Badwolf, your OSX calibration tool is fantastic!

    What is lacking now on the OSX side is a way to get the EDTracker talk to Opentrack. Not sure if Stanislaw will come up with a Joystick Input Tracker for it. Could the EDTrack Arduino Sketch be modified to output Hatire data that Opentrack on OSX understands?

  12. #4467
    Just giving you guys a little update on my problem with my 9250 board, incase anyone ever gets this problem and decide to look it up. Will update the youtube video aswell later.

    Had a chat with Brumster, and he looked on some of the readings in the control application, and we pretty much concluded that the device had to be defective in some way, cause the readings were way off.
    This happens every now and then with all electronic devices. So this shouldnt really put anyone who is intrested in the DIY kit or the pro version off buying it.

    Im getting it replaced with a new MPU, and hopefully everything will work as normal when i get that soldered in, will make an update post again when that happens.

    Shout out to everyone for contributing with some great tips!

  13. #4468
    Originally Posted by ElEsido View Post (Source)
    Badwolf, your OSX calibration tool is fantastic!

    What is lacking now on the OSX side is a way to get the EDTracker talk to Opentrack. Not sure if Stanislaw will come up with a Joystick Input Tracker for it. Could the EDTrack Arduino Sketch be modified to output Hatire data that Opentrack on OSX understands?
    Thanks for the feedback!

    I looked at Opentrack briefly (I say briefly it too me hours to get the dependencies together and get it compiling!) and it's very much a cut-down version of the windows version. I suspect expanding it to read joystick inputs, as does its windows counterpart, would be approach I'd suggest. It's then pretty generic going forwards.

    Regards,

    BW.

  14. #4469
    Joystick input is already planned on the issue tracker for Linux and OSX. For OSX this requires some basic Objective C++ knowledge. It's gonna be done eventually but I have limited time and will for feature I won't really use.

    Originally Posted by badwolf77 View Post (Source)
    Thanks for the feedback!

    I looked at Opentrack briefly (I say briefly it too me hours to get the dependencies together and get it compiling!)
    Are there issues with the official release?
    opentrack :: cross-platform open-source head tracking software
    Common issues :: read before posting opentrack issues

  15. #4470
    Originally Posted by sthalik View Post (Source)
    Joystick input is already planned on the issue tracker for Linux and OSX. For OSX this requires some basic Objective C++ knowledge. It's gonna be done eventually but I have limited time and will for feature I won't really use.
    FWIW, the HID routines that I used are all pure C, if I remember rightly. I can let you have the relevant code once I've refactored for beta if it'd help?

    Are there issues with the official release?
    This was several months ago so bear with my memory, but there wasn't a binary release and the mac instructions were sketchy if they existed at all. It certainly wasn't a case of ./configure ; make.

    Annoying, I only deleted its folder a couple of weeks back, else I'd have a look and see where I struggled. Sorry.

    BW.

Page 298 of 330 FirstFirst ... 288294295296297298299300301302308 ... LastLast