DIY Head Tracker For A Tenner

Blairvoyach · Sep 21, 2014

Skyflyer,

That all looks correct, axis assignments are now in Mapping/Options where you generate your curves (end tab)

HTH

SentineI · Sep 21, 2014

brumster said:
I think you're misunderstanding how you flash it - it's not a URL, you have to fire up a flash utility and it's listed in there in a drop-down box

Ah, THANKS. I thought I must have been me being dumb, but couldn't work out where I was mistaken: I didn't realise the GUI program automatically finds the latest flash images without me doing anything.

NeilF · Sep 21, 2014

brumster said:
1. Plug in your device
2. Download this
3. Run it
4. Choose "EDTracker2 (v2.20.5)" in the drop-down
5. Press "FLASH"

More detailed instructions are in the guide.
If you've got some COM port problems, I know the command-line flash tool often does a better job of things.

Thanks... Initial thoughts are I'm getting the view ending up off center even more (eg: during/after combat)? I'll recalibrate again and check.

If I want to go back to the previous version to see how it compares, do I just choose sketch 2.20.4 (old) and flash that?

EDIT:-
I've recalibrated it so I'll see how I get on with the new sketch.

Here's the result after about 12-15mins.

I guess 0 is about as good as it can get?

Antmax · Sep 21, 2014

NeilF said:
Thanks... Initial thoughts are I'm getting the view ending up off center even more (eg: during/after combat)? I'll recalibrate again and check.

If I want to go back to the previous version to see how it compares, do I just choose sketch 2.20.4 (old) and flash that?

EDIT:-
I've recalibrated it so I'll see how I get on with the new sketch.

Here's the result after about 12-15mins. I guess 0 is about as good as it can get?

Nice one

.

Aquila Oculus · Sep 21, 2014

I've emailed and put myself on the waiting list for one of these, next best thing until the OR and won't break the bank, thanks

NeilF · Sep 21, 2014

So carefully calibrated it earlier today. And after 12-15 minutes of leaving it, I was getting results like this:-

As of this evening, it is drifting at a rate of a tenth per second. So in just a couple of minute:-

:S:S

peterbbb · Sep 21, 2014

Downloaded etc and device flashed and recalibrated for yaw drift. All seemed fine and perhaps less jittery. Play more later and I'll report anything untoward. This is using win 8.1. Thanks.

Cmdr Akheir · Sep 21, 2014

Got my kit from Hobbycomponents. I tell ya they deliver them fast. Waiting on PCB from Rob now.

I've somehow developed symptoms of coughing and fever which work will be notified lol. Think its too much space fever.

Neillm · Sep 21, 2014

NeilF said:
So carefully calibrated it earlier today. And after 12-15 minutes of leaving it, I was getting results like this:-

As of this evening, it is drifting at a rate of a tenth per second. So in just a couple of minute:-

:S:S

Yeah, I got exactly the same. Yaw drift of 0.02 this afternoon and now it's at 0.70. You can see on yours that the temperature difference is over 3.5 degrees between the two images. Mine is about 1.5 degrees.

As someone suggested elsewhere, I think we need to wear the headset to get it up to the proper operating (and on head) temperature, and then try and calibrate it at that temperature...using a light or something to warm it up.

Neillm · Sep 21, 2014

Neillm said:
Yeah, I got exactly the same. Yaw drift of 0.02 this afternoon and now it's at 0.70. You can see on yours that the temperature difference is over 3.5 degrees between the two images. Mine is about 1.5 degrees.

As someone suggested elsewhere, I think we need to wear the headset to get it up to the proper operating (and on head) temperature, and then try and calibrate it at that temperature...using a light or something to warm it up.

OK, just putting my hands over the tracker to raise the temperature by a couple of degrees and my drift dropped down to 0.04.

Cmdr Akheir · Sep 21, 2014

My PC room, is mostly heated by the GPU running. Would this cause issue in the tracker?

Neillm · Sep 21, 2014

I could be wrong, but based on what I've seen, if the temperature is different between the calibration and actual usage, there is likely to be a difference is the drift.

brumster · Sep 21, 2014

Yes, the calibration needs to be done with it up-to-temperature really (give or take a degree or two). Since the auto-centre is less 'viscious' now, if you HAVE got bad drift it might have masqueraded behind the auto-centering previously but come through now.

I'll do a bit more playing in ED just to confirm it's no worse than before but, for me, I played 30 minutes yesterday and 30 minutes the night before without any drift or change of behaviour (apart from the reduced auto-centre "kick", naturally).

Having said that, mine is running in an enclosure so the temperature is possibly more consistent. Maybe we should just get everyone to lag their MPUs

it's no issue to do such a thing, the temperature being higher is fine, just so long as it's consistent. Open a window, drop the room temperature 5 degrees, and you WILL notice a change.

peterbbb · Sep 21, 2014

Mine is in an enclosure, tic-tac box - quality stuff here - and I shaped it so that the 'door' closes and the electronics cannot move and the enclosure has two small holes. The device cannot be affected by draughts in any significant way and has been very stable, once it is up to temperature. This takes about 10 minutes. Which is about how long it takes me to get from turning the PC on to playing the game.

Armored · Sep 21, 2014

Are you giving out a printed version of your PCB design? I would love to try out etching for the first time and just doing this project from start to finish would be very awesome.

SentineI · Sep 21, 2014

@brumster
Why can't the act of pressing the reset button (recentering it) automatically recalculate the drift? This would automagically handle changes in drift over temperature & time (so no need for fancy Y=MX+C to estimate the drift like I suggested before). This also seems like it *ought* to be simple to implement?

I'm just getting annoyed that I calibrate it one day, and it seems fine (well recenter every 15 minutes or so), and then the next day it's drifting like crazy (recenter every few minutes).

Antmax · Sep 21, 2014

I used some quick drying 2 part epoxy to lock/secure the microUSB socket in place. I wonder if a dab of that on the temp reading part of the board would insulate it and make it change temp much less.

Trouble is, which part of the board is the temp sensor.

n0ahg · Sep 22, 2014

Antmax said:
I bought mine ready built from the Edtracker guys and used the official Arduino drivers you mentioned. They seemed to work in that the edtracker was originally recognized as an Leonardo but on uploading sketches it changed id to promicro and failed..

The steps I mentioned were the only ones that worked with 8.1. I heard that prior to 8.1 the latest official drivers worked fine.

It was a very common problem with the pro micro a couple of months ago. I didn't build my tracker though, so I don't know how it was originally flashed. Not in Win8.x though.

Works fine with the steps I mentioned and have had no problems since.

Sounds like someone is building the sketches with the device set incorrectly so the boot loader is being overwritten with the wrong one. This would then cause the official drivers to stop working.

SnowGhost · Sep 22, 2014

Regarding drift. I've had some thoughts. These thoughts come from someone who has ordered the bits, but doesn't have an EDTracker and also isn't a big tinkerer. So, maybe it's not possible. It would also add cost. But from looking at ebay, not that much cost. It would also add bulk, and going by the instructables page a much bigger circuit board (This is absolutely out of my area of knowledge)

What if you grabbed a ebay cheapo Wii remote and removed the IR Camera.

Hooked up the IR Camera to the arduino. http://www.instructables.com/id/Wii-Remote-IR-Camera-Hack/?ALLSTEPS

Then used a Wii sensor bar. The linked one is battery powered. Here is a USB powered bar

The sensor bar is nothing of the sort. It's actually got 4 IR LEDs in it, and the wii remote IR Camera does all the heavy lifting.

None of the above components are particularly expensive.

Then the EDTracker software just needs to keep an eye out for the IR signals, so it knows where straight ahead is. It should drop the drift to zero. It doesn't need a continual watch on the camera. Just every 5 minutes, when you're looking straight ahead anyway.

Neillm · Sep 22, 2014

After a bit of random calibration, I hit/fluked on a reasonably stable drift setting and actually had a couple of hours of good game time with only having to reset the view every 5 minutes or so. Using the Opentrack ability to bind a key to the reset which can the be mapped to a HOTAS button is great for this.

The recentering is far more subtle and manageable with the new sketch, so big thanks to pocketmoon for this.

After the two hours play, with the tracker mounted on the headset and on my head, the temperature was quite stable at around 22 degrees, so I need to do the calibration at this temperature. The random bit from earlier came from me wearing the headset and saving the compensation until it seemed stable. This is...err...not optimal and certainly not best practice but I wanted to play the game.

Interesting suggestion by Antmax about insulating the temperature sensor.