DIY Head Tracker For A Tenner

looking into making one of these so i ordered a couple of gyro boards :)

although i remembered i may have already had this type of thing in the form of a multiwii based pocket quadcopter (actually i got 2 kicking about) and so i checked them out and they indeed have the MPU 6050 chip on them :) and an atmega 32. nice and i dont have to solder anything up :)

So i unplug the motors and flash the calib sketch from the gui.. Bingo! left it to warm up and then calibrated niceley :) and right now its doing its drift compensation (best time to make a cuppa and make this post ;) )

hopefully it works (will still make one since i may want to use the quads as quads one day XD )
 
Hi Guys just wanted to say thanks for the latest edtracker UI 3.0.4.0

last time I calibrated this thing was back in august last year and needed to recalibrate it due to a new pc, but what a difference the GUI is, its so much simpler and brumster your vids are great keep up the good work, your last video was so simple no impending questions or anything, so all in all very good this device has come such a long way :D
 
Hi guys,

First of all, thanks for the work you guys have done on EDtracker.

Now, I bought a PCB from you and the hobby components bundle. Although I have no soldering experience after lots of swearing I thought I had it all together.

Now I can flash the Calib file from the GUI, but it is not detected after that. When I plug the EDtracker into the USB, on the Arduino an orange LED flashes once and the red LED stays lit, whilst on the MPU board the LED fades up then fades out over about a second and then stays out.

I've checked my soldering, and although far from pretty I can't see any bad connections, anyone have any idea what I've messed up here?
 
Sounds to me like a register is wrapping somewhere. Probably to do with the drift compensation value reaching its limit and wrapping over, but not wrapping cleanly. I wonder if there's a ">" instead of a ">=" somewhere....

If someone can reproduce this easily (ie. very repeatable) then I suspect Rob would be interested in some debug data. I've asked for a debug log to be added into a future revision of the GUI so that it can be turned on, left running, and then a resulting log can be emailed to us to figure out what's going on in situations like this.

Thanks for the input and this happens every night for me so can provide any info needed :)
 
Edit:

I was able to resolve the issue by unplugging the EDT, rebooting my PC, plugging the EDT back in, then re-flashing the 9150 calibration and software.


Pocketmoon and gang, I'm hoping you see this and can help me out. I put together the EdTracker with the 9150, and it was working wonderfully. I disconnected it for a couple of days (needed the USB ports for other stuff, and the usb cord was getting bothersome when not in the game), but now that I have it plugged back in, it's not being detected. Here are the facts so far:

- The light on the MPU is lit up, and the device is getting powered.
- Windows detects it, and it shows up in Computer Management in Windows 7.
- The EdTracker UI 3.0.4.0, upon scanning ports, shows it connected to COM3.
- If I click the "Connect to Tracker" button, the little text box to the left of the "temperature" box says "connected", but the large text box above that says "NOT CONNECTED".
- Even though it says it's connected, I'm getting no readings from it.


Any ideas?

URNF0Vp.png
 
Last edited:
i have some annoying behaviour i've been trying to find a cure for (pre-built 6050 in the black box with the sexeh USB cable). when i look at a UI screen, say for 10 seconds, then look forward again my view is not centred, it's too far to the side of the UI screen i was looking at. this happens either side.

i've let my EDTracker stabilise for 30 mins during the calibration process (EETracker2Calib 2.5.3), those dots are now nicely lined up and the gyro numbers stayed stable with the accel numbers fluctuation by 1, maybe 2. i left it running another 15 mins once calibrated and this behaviour stayed constant so i assumed that was as good as it was going to get.

i took a similar approach to letting the drift compensation run it's calculations (EDTracker2 2.20.7), left it wedged between 2 books. after i saved the value i let it run again to see if it would change. it started out at 0.05 difference then that gradually went down to 0.02 so i called that stable enough. i then pinned it in place with some constant pitch to see if it would drift, which it didn't as i was wondering earlier if the pitch my headset puts it at would mess with the drift compensation.

my temps are annoying, on the bench it's at 27 deg, on my head it's at 32 deg, so i'm aware i cannot get a perfect calibration.

that said i seem to have eliminated the drift issue i had initially (recentre set to strong and GUI minimised), time spent calibrating really does pay off!

so it's just left me with this habit of not going back to centre after i've looked to the side for a while

i'm using opentrack 2.3rc5, joystick as input, EDTracker2 selected and i've had to map yaw to #2, pitch to #3 and roll to #4. i believe these differ from the tutorial but with a bit of messing about on the bench tilting it one way then the other it seemed clear this is the way i needed them assigned. i'm using freetrack with only show trackir. i'm using accela with smoothing to the min and noise reduction at low. it makes things a little rough but i decided to stick with data closer to raw to get a better feeling of what was happening. the spike noise smooths out just enough for it to be playable.

on mappings i map yaw of 40 to 180, pitch of 40 to 180 and the rest are pinned at zero. i have a 27inch screen so get to look about a fair bit. i disabled translation compensation but that didn't seem to make any difference (no surprise there, it's only getting info for 3DoF)

after about 1 hrs flight time bumbling about looking for pirates it was behaving impeccably, but when i'm flying like that i don't use my UI screens for very long. by the time i finally docked the station UI was nice and centred. so it looks like i've got things working well, starting to feel better than my LED tracking (with a bonus i can look up and behind more in my ASP ;) ).

so i'm not sure why it wont centre if i've been looking to the side. i noticed it when coming out of the fittings screen as i'd been looking at the left side of the screen mostly. i notice it after selecting a nav target while in station. it's really bad if i look at my 2nd monitor to read the forums etc, at times i'd get back in game and my ahead view was looking at the UI!

so is this some inherent limitation in the hardware? is there some tweakage i can do? i'm no hardware guru but i have messed with serial components before (they let me play with a frickin laser! i then had to use it to write some distance logging software). so i'm happy to get my hands dirty so to speak.

i've played with both LED tracking and the mobile phone free solution. the EDTracker knocks the socks off my phone, i had to turn it off as soon as i'd landed, but after getting a taste for it i couldn't fly without it! despite having an EDTracker on order i went and got an LED tracker too so i could get an upgrade from the phone. i was still looking forward to the EDTracker because it has a much finer degree of accuracy (LED is limited by the 640 resolution on the sony webcam). now that i have it working it's got a better feel to it than the LED does. i was also a bit surprised to find the LED tracking isn't much better at the return to centre thing, so could it be some limitation in the tracking software? i've tried both opentrack and freetracknoir for the LED, both had issues recentering though mostly after looking at my second monitor. i didn't notice it when looking at the UI screens.

if there is no fix to this then it's not a show stopper, i've got it to a perfectly workable point, just looking to get rid of the remaining niggles.
 
Here is the next question, is there an affordable HMD that this can be coupled with? If not, are there parts that can be purchased that can be used to make something affordable? I am currently looking but not having much luck as of yet.

EDIT 4:19AM 1/9/15 I found some links of interest.

This link is for a 4.3 inch HDMI display. From the looks of things you pretty much have to put the whole thing together yourself.

http://whttp://www.buydisplay.com/d...av-driver-controller-board-tft-module-display

This link is for a Styrofoam kit First Person View kit that comes with a display but the display is AV cables.

http://www.hobbyking.com/hobbyking/store/__53706__Quanum_DIY_FPV_Goggle_Set_with_Monitor_KIT_.html
 
Last edited:
Here is the next question, is there an affordable HMD that this can be coupled with? If not, are there parts that can be purchased that can be used to make something affordable? I am currently looking but not having much luck as of yet.

EDIT 4:19AM 1/9/15 I found some links of interest.

This link is for a 4.3 inch HDMI display. From the looks of things you pretty much have to put the whole thing together yourself.

http://whttp://www.buydisplay.com/d...av-driver-controller-board-tft-module-display

This link is for a Styrofoam kit First Person View kit that comes with a display but the display is AV cables.

http://www.hobbyking.com/hobbyking/store/__53706__Quanum_DIY_FPV_Goggle_Set_with_Monitor_KIT_.html


the 1st link seems to be leading to a dead domain :x

EDIT: i cut the 1st part of the url off ;) http://www.buydisplay.com/default/4...av-driver-controller-board-tft-module-display
 
Last edited:
Pocketmoon and gang, I'm hoping you see this and can help me out. I put together the EdTracker with the 9150, and it was working wonderfully. I disconnected it for a couple of days (needed the USB ports for other stuff, and the usb cord was getting bothersome when not in the game), but now that I have it plugged back in, it's not being detected. Here are the facts so far:

- The light on the MPU is lit up, and the device is getting powered.
- Windows detects it, and it shows up in Computer Management in Windows 7.
- The EdTracker UI 3.0.4.0, upon scanning ports, shows it connected to COM3.
- If I click the "Connect to Tracker" button, the little text box to the left of the "temperature" box says "connected", but the large text box above that says "NOT CONNECTED".
- Even though it says it's connected, I'm getting no readings from it.

Any ideas?

That's a very odd one, if windows is detecting it but the GUI isn't I can only suggest a bit of unplugging/replugging, restarting the GUI, or maybe even a reboot. All I can think is that the .NET serial port libs have the serial port locked and it can't communicate for some reason. I assume if you look in control panel at the joystick that EDTracker is pretending to be, you can see it moving fine?

i have some annoying behaviour i've been trying to find a cure for (pre-built 6050 in the black box with the sexeh USB cable). when i look at a UI screen, say for 10 seconds, then look forward again my view is not centred, it's too far to the side of the UI screen i was looking at. this happens either side.

I think I know this one, I've seen it happen myself when I spend a long time in station menus, but moreso before they added headlook working in the menus. I don't use Opentrack but I can't see that being a contributing factor.

What happens is when you're browsing the station menus you head naturally moves a little bit but you're not so conscious of it like you are in battle. In battle, you tend to control you head movement much more, because you're actively using headlook. In the menus, you rarely hold your head dead-centre and look around the menu with only your eyes - you look more with your head. And I think the problem is, because you're holding your head off-centre for extended periods of time (probably browsing market prices is an example, or the outfitting menu - most of the menu options in ED are on the left, so your head tends to be more to the left) the auto-centering won't correct - it believes you're tracking a target off to the left, for example. If you've got a big drift value (but auto-corrected, of course) then that can creep in over extended periods.

Not really a lot that can be done without compromising auto-centre :( I just re-centre if it happens after extended periods in the station menus. Or use the 9150 mag version, if you can (depends on the hardware in your EDTracker)....
 
I have voice attack commands to "Head Lock" which toggles on/off head look, and issues a F key autocentre to the ED Tracker GUI (I have mine binded to F11).

When I dock, I say "Head Lock" and it turns off head look. Then everything is still. Once I take off I issue "Head Lock" while looking forward and it auto centres and turns it back on.

I also have "Alignment" command to issue the F11 on its own - but I don't need to do that very often. Since I auto centre after any docking EDtracker is stable enough between docking to work fine.
 
I think I know what you're getting at Robb, I'm using a JS key to send the recentre (f1) to the edtracker and toggle headlook (f2) to ED. This works fantastic (I'm not able to get VA to issue a command to two separate programs).

However If I bring Teamspeak 3 into the mix, it interferes. It seems that it interferes with both ED AND EDT or maybe it's interfering with the Joystick (x52pro) and the VJOY(EDT). The toggling headlook doesn't always switch on and off, also my headlook movement is either jerky, doesn't happen, or happens smoothly, so as things stand, I've setup a TS3 server for community, but I'm the only one that can't speak on it. Even just having TS3 open and not connected interferes this way.

Any ideas's anyone?
 
Pocketmoon and gang, I'm hoping you see this and can help me out. I put together the EdTracker with the 9150, and it was working wonderfully. I disconnected it for a couple of days (needed the USB ports for other stuff, and the usb cord was getting bothersome when not in the game), but now that I have it plugged back in, it's not being detected.

If you load the calibration sketch does that respond ? If so then it could be the button is stuck on/shorted. The main sketch reads the button state and recentres when it's down. If it's stuck then the tracker gets stuck in recentre mode and never talks back to the UI or generate joystick output.
 
Hi guys,

First of all, thanks for the work you guys have done on EDtracker.

Now, I bought a PCB from you and the hobby components bundle. Although I have no soldering experience after lots of swearing I thought I had it all together.

Now I can flash the Calib file from the GUI, but it is not detected after that. When I plug the EDtracker into the USB, on the Arduino an orange LED flashes once and the red LED stays lit, whilst on the MPU board the LED fades up then fades out over about a second and then stays out.

I've checked my soldering, and although far from pretty I can't see any bad connections, anyone have any idea what I've messed up here?

No-one got any ideas here then?
 
Sorry, missed that one. Definitely something wrong with the hardware there - sounds like a problem with the power getting to the MPU. Probably not a dodgy connection, sounds more like the voltage regulator or a capacitor on it is "marginal" :-S ?!
 
Sorry, missed that one. Definitely something wrong with the hardware there - sounds like a problem with the power getting to the MPU. Probably not a dodgy connection, sounds more like the voltage regulator or a capacitor on it is "marginal" :-S ?!

Cheers. So it looks like a faulty component on the board?

Bit annoying as I doubt hobby components will give me a refund on a board soldered up to a PCB! (And de-soldering is beyond my capabilities and equipment!)
 
If you like, I'd happy to take a look at it and see what I can do if you want to post it to me. Just pay for return postage and whatever needs swapping out, if anything. At worst sounds like just a new MPU6050 module, which is only a few quid...?

PM me if you're stuck...
 
Back
Top Bottom