DIY Head Tracker For A Tenner

Peeps!

New EDTracker 9150 (magnetometer) test firmware is available. If this looks good I'll be aiming for a "production" release at the weekend, along with an update UI and full instructions.

Instructions for use for at http://reprapdad.wordpress.com/2015/01/13/edtracker-latest-9150-test-firmware/

ta ta

Rob
Shiny.

Quick question, it should be possible to remove the 6050 MPU from my existing edtracker, and just slot in a 9150 MPU in its place right?

Assuming I manage not to set anything on fire, I figured it would be quicker and cheaper than building an entire new unit.
 
I'm interested in that solution too, if is possible.
It is certainly possible because the 9150 and 6050 are pin compatible. You should be able to replace one with the other, then recalibrate and load the 9150 software.

But it will depend on your de-soldering skills and equipment, obviously. De-soldering braid and/or a sucky tube thing would probably do the job, it's only 8 pins.

If you get a MPU9150 board from HK or China from a certain auction site, it's less than GB £6 delivered (in a couple of weeks), which is a good value upgrade for the 6050 I think.

Since it's not my EDtracker and it won't cost me anything if it goes wrong, I say go for it!!
 
Last edited:
New EDTracker 9150 (magnetometer) test firmware is available. If this looks good I'll be aiming for a "production" release at the weekend, along with an update UI and full instructions.

This was the first 9150sketch to work properly for me. I went though your instructions but had to do an Edtracker 2 2.20.7 flash between stage 5 and 6 of your instructions.

This is the first of the 9150's that worked pretty well. I use it with opentrack and had some minor niggles. I think the sketches with the mag are using quite a few more resources. I started off fiddling with filters in opentrack but there was always a micro stutter.

Tweaking got it mostly there with a slight blur, Switched down resolution from DSR 1440p to vanilla 1080p and the microstutter was gone until I landed in the fancy plush Alliance Capital station. Quick investigation showed my framerate had dropped to low to mid 50's with the Station commodity window up. With the old 6050 I didn't have any stutter it was totally smooth.

Having said that, it is now working really well at the lower res. It doesn't need re centering unless I put it down on its side for a few minutes. Easy solution is to perch it upright against something instead of lying it on the desk.

So Far so good. Was a bit surprised about the microstutter. I figured it would be using CPU not GPU (unless that is something more to do with Opentrack)

With Opentrack, old version 2.2 was easier to filter than the new 3.0 RC8 and ran smoother. RC8 seemed to have some friction/stiction which felt weird and couldn't be tuned out without some unwanted side effect.
 
Last edited:
I am assuming its the latest hardware (Model EDTTRK002), will I need to update to 9150 would you know, and is it quick to get going from scratch?
The standard EDTracker is the non-magnetometer versions. The 9150 version was an upgrade option in the shop so it depends if you selected that or not. We've revamped the shop a bit to split the variants out into distinct products now.
 
I have a question - is there a way to lower the sensitivity when looking straight ahead?
Only the deadzone, as you mention, or using opentrack to add manual curves. Next round of changes will introduce a soft-zone option for both 6050 and 9150 EDtrackers - combining an initial exponential curve which becomes linear, so you get the smoothing benefit of expo with the linearness of linear :) Will also add a user configurable filter to smooth out jitter.

- - - - - Additional Content Posted / Auto Merge - - - - -

So Far so good. Was a bit surprised about the microstutter. I figured it would be using CPU not GPU (unless that is something more to do with Opentrack)

I'll look into that. I wouldn't expect the 9150 tracker to cause a drop in fps unless the UI is eating up some cpu, which it shouldn't when minimised. Microstutter sounds more like a timing issue with the EDTracker sending joystick updates clashing with your framerate. Do you run vsync on or off?
 
It is certainly possible because the 9150 and 6050 are pin compatible. You should be able to replace one with the other, then recalibrate and load the 9150 software.

But it will depend on your de-soldering skills and equipment, obviously. De-soldering braid and/or a sucky tube thing would probably do the job, it's only 8 pins.

If you get a MPU9150 board from HK or China from a certain auction site, it's less than GB £6 delivered (in a couple of weeks), which is a good value upgrade for the 6050 I think.

Since it's not my EDtracker and it won't cost me anything if it goes wrong, I say go for it!!

I did this with one of mine. It's no easy task mind you. I unsoldered the pins but they were still unmovable. So I got a hacksaw and carefully cut the 6050 board away from the pins and then soldered the 9150 directly on top.

If you have some desolder wick, I would advise giving that a shot as you may find it pulls the solder away better than the pump action remover that I have.
 
I have a question about pesky drift.
My device will be stable for a short time, maybe 15 minutes or less and then will drift to the left. A little at first but after 15 minutes or so it becomes unusable needing to be recentered every couple of minutes.
I have many times recalibrated it. This seems to go fine.
I have tried many variations on drift compensation - even manually compensating for the sudden increase in drift this works temporarily but the drift increases so fast it is impractical. The automatic drift correction is not useful because as soon as it calculates values they are no longer useful.
Does anyone else have this experience or any ideas about a solution? Or maybe some ideas on how to troubleshoot.
Temperatures are stable. I have also tried the device in different physical orientations but it does not matter the drift is still to the left. Mouse headlook is off in game (heheh..) I think I have tried everything obvious but I may be missing something.
Thanks in advance.
 
Last edited:
I did this with one of mine. It's no easy task mind you. I unsoldered the pins but they were still unmovable. So I got a hacksaw and carefully cut the 6050 board away from the pins and then soldered the 9150 directly on top.

If you have some desolder wick, I would advise giving that a shot as you may find it pulls the solder away better than the pump action remover that I have.
I have a sucky thing. Bad news about having to take a hacksaw to the MPU though, was hoping I could keep the old one intact just in case.

I think I'll order one of those Chinese 9150s and give this a go anyway. Wish me luck!
 
Ha! Both my boards were from China and they work perfectly and all in with EDTracker PCB was £21. I seem to recall 3 figures for track ir and I can't imagine paying that much when my 9150 EDTracker is working so brilliantly now. AND!!!! I got to 'build' it myself!!

I ran it for a while in the 6050 mode as it was better than the experimental code, until 9150 V3. I can't wait to see what the next code iterations produce.
 
I have a question about pesky drift....

Couple of things to check ;
What do you have AutoCentre set to in the UI?
Do you play with the UI minimised ?
If you place the tracker down and hit 'recentre' does the view in the cockpit become central (toggling headlook on and off will show if it isn't central) ?
 
I have a question about pesky drift.
My device will be stable for a short time, maybe 15 minutes or less and then will drift to the left. A little at first but after 15 minutes or so it becomes unusable needing to be recentered every couple of minutes.
I have many times recalibrated it. This seems to go fine.
I have tried many variations on drift compensation - even manually compensating for the sudden increase in drift this works temporarily but the drift increases so fast it is impractical. The automatic drift correction is not useful because as soon as it calculates values they are no longer useful.
Does anyone else have this experience or any ideas about a solution? Or maybe some ideas on how to troubleshoot.
Temperatures are stable. I have also tried the device in different physical orientations but it does not matter the drift is still to the left. Mouse headlook is off in game (heheh..) I think I have tried everything obvious but I may be missing something.
Thanks in advance.


Unfortunately, I've never managed to get mine to work properly either... It detected fine by Windows, connected and appears in game. I switch it on (with a hotkey in game), look at the screen and press the button to re-centre, and it works that first time. The moment I switch it off with the hotkey to do something else, then switch it back on again with the hotkey, it immediately flies across to the left screen and stays there, regardless of how I move my head. It has to be re-centered again before it will work, but only that once, before flying across to the left screen again!

Recalibration follows deManglers' experiences above. Works for a short time, but then is gone again. The unit is strapped to a headset and is orientated as per the setup program.
 
Last edited:
Couple of things to check ;
What do you have AutoCentre set to in the UI?
Do you play with the UI minimised ?
If you place the tracker down and hit 'recentre' does the view in the cockpit become central (toggling headlook on and off will show if it isn't central) ?

I've had the same problem as deMangler.

So far I've been assuming that it's down to temp change between the temp on my desk that I calibrate at (around 28c) and the temp it actually runs at on my head (when I'm in game after its warmed up the temp is typically 32c).
Would this make difference? It seemed not to, but although I ran the calibration at 32c, by the time I got to run the drift calculation it had dropped below 31c. How sensitive is the calibration to temp changes?

The thing you mention that I haven't done is that I don't minimise the ui, I keep it maximised in the background.

The other thing I'm now thinking is that turning off autocentre may help (I've tried each setting other than off). I think it may be that my head physically drifts slightly left, because the drift is far worse on the starport screen when content is on the left. I've set headlook dead zones in game to remove the wobble from slight movement as I look around the screen, but if the tracker autocentre is running it stands to reason that it will be auto centering in the background. So no autocentre is my next move unless anyone has better ideas.
 
Unfortunately, I've never managed to get mine to work properly either... It detected fine by Windows, connected and appears in game. I switch it on (with a hotkey in game), look at the screen and press the button to re-centre, and it works that first time. The moment I switch it off with the hotkey to do something else, then switch it back on again with the hotkey, it immediately flies across to the left screen and stays there, regardless of how I move my head. It has to be re-centered again before it will work, but only that once, before flying across to the left screen again!

Recalibration follows deManglers' experiences above. Works for a short time, but then is gone again. The unit is strapped to a headset and is orientated as per the setup program.

Hope I'm not suggesting something obvious here, but, in game, do you have the tracker set to the axes or Look Up, Look left, etc? If the latter, that can make it look to the extremes of the screen and not track along with your head position.

As I said, apologies if this is not the case but it's worth checking.

- - - - - Additional Content Posted / Auto Merge - - - - -

I've had the same problem as deMangler.

So far I've been assuming that it's down to temp change between the temp on my desk that I calibrate at (around 28c) and the temp it actually runs at on my head (when I'm in game after its warmed up the temp is typically 32c).
Would this make difference? It seemed not to, but although I ran the calibration at 32c, by the time I got to run the drift calculation it had dropped below 31c. How sensitive is the calibration to temp changes?

The thing you mention that I haven't done is that I don't minimise the ui, I keep it maximised in the background.

The other thing I'm now thinking is that turning off autocentre may help (I've tried each setting other than off). I think it may be that my head physically drifts slightly left, because the drift is far worse on the starport screen when content is on the left. I've set headlook dead zones in game to remove the wobble from slight movement as I look around the screen, but if the tracker autocentre is running it stands to reason that it will be auto centering in the background. So no autocentre is my next move unless anyone has better ideas.

I had exactly the same problem with my original 6050. I just couldn't get the drift to the left under control, no matter how carefully I calibrated it. After a while playing, I'd be recentering every 20-30 seconds.

I convinced my other half to get me a 9150 version for Christmas (been dispatched so should be soon!) so hopefully I won't have any more drift problems.
 
The thing you mention that I haven't done is that I don't minimise the ui, I keep it maximised in the background.

It's absolutely critical to minimise the UI.

Autocentring has no effect if the UI is on-screen. We do have this in the user guide but it's easy to overlooked :)

[edit] The next ui will change this so it's always active. Which should solve a lot of issues.
 
Last edited:
It's absolutely critical to minimise the UI.

Autocentring has no effect if the UI is on-screen. We do have this in the user guide but it's easy to overlooked :)

[edit] The next ui will change this so it's always active. Which should solve a lot of issues.

Any idea why the tracker i am using and some others has an issue where the screen will appear to flick to the right for a split second when using said device?
its the 6050 chip.

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.

page 224 bottom 3 posts talk about it.
 
Last edited:
I did this with one of mine. It's no easy task mind you. I unsoldered the pins but they were still unmovable. So I got a hacksaw and carefully cut the 6050 board away from the pins and then soldered the 9150 directly on top.

If you have some desolder wick, I would advise giving that a shot as you may find it pulls the solder away better than the pump action remover that I have.

Yes, the voice of real experience there. Some thin gauge (~1.5mm maybe) desoldering braid/wick is worth trying.

Whenever I think about soldering and de-soldering, it always seems about ten times easier in my mind than it actually turns out to be in practice. For psychological reasons I never register that I'm a ham-handed, clumsy idiot, or that solder is designed to act exactly contrary to my intentions.

But it is possible.:D
 
Back
Top Bottom