DIY Head Tracker For A Tenner

Is there anyone who has put an Edtracker in a nice maplins or other enclosure willing to do it for me? I am very very bad at fiddly things and I lack the tools. I will cover all postage costs/materials/Beer Money in advance.
I can't just mount it to my headphones as the spikes are marking the headband.
Can't do tic tac box as mine seems too thick to fit in properly.

You couldn't just do the Maplin box version in my sig? - http://forums.frontier.co.uk/showpost.php?p=567586&postcount=1419

If you choose bigger maplin box (60mm) instead of 50mm you won't even need to cut the PCB. But you will need to simply hold/pad it into place better inside the box.
 
Yes yours looks very good thank you. I'm just really bad at practical stuff. Figuring out how to stop the thing moving inside that box is how some people would feel being asked to design a rocket ship from scratch with a pen and paper.
 
Can't monitor it after flashing...

Brumster or others in the know,

Hiya mate, hope you can help. Finally got my board and built it all up tonight. All was going swimmingly.

Loaded up the Calib sketch with no probs and calibrated it fine, had the two dots in the middle no probs. Loaded up and flashed the main sketch which the box says was succesful, however now it says error finding it and in the GUI unknown device - not monitoring...

If I load upi the Calib sketch again now I get the same, not monitoring. In my system I can see the Leonardo fine on Com 3. I've tried changing to Com 3 in the GUI with no luck...

Any ideas or have I somehow bricked it?

Paul
 
Yes, those guys are awesome.

(Disclaimer: The following is an attempt at humor, Read at own risk!)

Although having watched a few videos of Mr. Brummie, I'm still trying to figure out what "soilent running" might be.... :rolleyes:
 

Tar Stone

Banned
Just an fyi,
Maplins sell round Velcro pads that are superb. They're very secure and stick onto the rounded surface of a headphone can and still have enough grip to solidly hold the flat surface of your tracker, if you're a side mounty like me.

I've been able to peel one off and it didn't leave behind a mess either.

Big thanks to the edtracker guys! Been using a DIY trackir with freetrack for about a year, the edtracker reduces CPU overhead and tracking lag to practically zero.
 
Phew done it ...

well that was a bit traumatic, but i got it up a running last night, i think i might take my time and do the drift compensation again today though.

I plugged the tracker into my pc and the drivers installed with no problem and i thought everything was going to go swimmingly but i guess inevitably it all went down hill from there. :rolleyes: Initially i couldn't get the GUI to find the tracker at all, no matter what i did so in the end i installed the arduino IDE and flashed the blink sketch onto it. you would not believe how happy someone can be seeing a flashing amber light lol.

in the end the only way i could get the calibration sketch onto the tracker was to use the command line utility which worked flawlessly. the GUI then found the tracker no problem and i could calibrate it and then flash the main sketch. :D

the only two things i need to do now is find a way to turn off the usb ports in my pc ( i think I'm just gonna have to crawl under the desk and turn the PC off every time) and i also want to incorporate the magnetometer (compass) into the tracker, I know a bit about programming in C so I'm gonna have a play with the sketch and hope i don't brick it :p ( any help with this would be most gratefully received)

edit: I bought the MPU 9150 that has the magnetometer incorporated
 
Last edited:
Temperature.

Just a thought on Drift. The device has a built-in temperature monitor, yes? Can the drift-compensation software be programmed to automatically increase/decrease a set amount, based on a temperature fluctuations? It seems to me that wearing the Tracker (on top of my head in this case), plus using it in-game, generates higher temperatures than when it's just sitting still on your desk during calibration/drift compensation set-up. I'm still finding my device is drifting off after only a few minutes use :( For those first few minutes though, it's agonisingly close to working as it should :eek:
 
Last edited:
Just a thought on Drift. The device has a built-in temperature monitor, yes? Can the drift-compensation software be programmed to automatically increase/decrease a set amount, based on a temperature fluctuations? It seems to me that wearing the Tracker (on top of my head in this case), plus using it in-game, generates higher temperatures than when it's just sitting still on your desk during calibration/drift compensation set-up. I'm still finding my device is drifting off after only a few minutes use :( For those first few minutes though, it's agonisingly close to being perfect :eek:

I think the talk was of if building up a map of yaw by temp and using that? Goodness knows if that possible or indeed even necessary?

As regards you device drifting off, I assume you've calibrated it carefully? After you do so, is the drift visible in the calibration tool? ie: Leave it for 10 mins and it is/isn't aligned? Or are you instead experiencing it in game?
 
The problem with a temperature compensation mapping is that the drift is not consistent. Meaning that the drift you get at a given temperature might have a different value when run next time at the same temperature.

This would only be good to get a rough estimate of the drift.

I had the best results for drift compensation when the device was at working temperature (i.e. on my head). I did this by wearing my headset (where it's attached to) for half an hour before doing the drift compensation.

The working temperature thing is also mentioned in the excellent guide, but it can be missed easily. Putting it in a box with a big exclamation mark could help (wink, wink, nudge, nudge). ;)

Regarding the 9150, it seems that while getting a 9-axis output of the DMP itself is possible, no one knows how (or won't tell anyway).
So, back to 6-axis output and raw data from the magnetometer.
If one searches hard enough, there are a lot of examples for making sense of the raw output, from a basic low pass filter to kalman filtering.

Sadly I haven't had the time yet to look further into this. Or maybe I'm just secretly hoping someone else does... :D
 
Sadly I haven't had the time yet to look further into this. Or maybe I'm just secretly hoping someone else does... :D

Nope, I'm with you on that - investigated high and wide for Invensense firmware knowledge on how to get the mag enabled on the DMP output, but Invensense are very coy about telling anyone anything of any use :(
 
I think the talk was of if building up a map of yaw by temp and using that? Goodness knows if that possible or indeed even necessary?

As regards you device drifting off, I assume you've calibrated it carefully? After you do so, is the drift visible in the calibration tool? ie: Leave it for 10 mins and it is/isn't aligned? Or are you instead experiencing it in game?
I try to calibrate carefully. After leaving it 10+ mins it is roughly aligned. I say roughly because the orange dot is the one that is most difficult to get central, it takes several presses of the 'Calculate Bias Values' button to get it central. After moving the Tracker the blue dot returns nicely to the central position, but the orange one is always slightly errant. It always wavers slightly after calibration and seems to get further away over time. What does the orange dot represent? The variation in temp is roughly between 13 and 19 degrees. It'd be nice if I could manually adjust the Drift comp.
 
Last edited:
Please don't mix up calibration and drift compensation, as they are quite different.
For both to work best, DON'T move your device during the process.

As mentioned above, drift compensation is best done at working temperature.

Wear your headset with the device attached for half an hour before doing drift compensation, this should get you better results.
 
Please don't mix up calibration and drift compensation, as they are quite different.
For both to work best, DON'T move your device during the process.

As mentioned above, drift compensation is best done at working temperature.

Wear your headset with the device attached for half an hour before doing drift compensation, this should get you better results.
I'm not mixing them up.

Here is the sequence:

I wipe the eeprom.
I quit and reload the UI.
I install the calibration software and have to press the 'calculate bias values' several times to get the orange dot in the centre, the blue dot is centralised after one attempt. I have to get the X & Z Accel into the -300's in order for the orange dot to centralise. Maybe this signifies a problem in itself.
I quit and reload the UI, install the Tracker Utility and perform the drift compensation.
After this is complete I save the settings. The orange dot does not return to the central position the way the blue dot does.

I think that drift compensation should be part of the calibration, but whatever.

Update: After some experiments, it's clear that the device needs to be kept at a very stable temp to avoid drifting. The drift compensation needs to be variable to match real-time temp readings.
 
Last edited:
Back
Top Bottom