ED Tracker 2.9 (Friday 13th)
- Thread starter pocketmoon
- Start date
I murdered my first 9150 with a soldering iron (long story!) my second works fine as a drop in for the 6050. Having calibrated a few of these beasties there is a fair bit of variation between sensors, especially in terms of accel noise. I have a nice PCB clamp to keep things still and level but they are so sensitive. I'm sure they can pick up when next door has the music on. Some calibrate in a singe go and other take a bit more care.
I'm currently working on getting a good magnetometer heading. It's very noisy on the 9150 (+/- a couple of degrees) and requires a hefty low pass filter to get a stable value. The plan is to perform background drift compensation based on the forward looking magnetometer reading (so no complex mag calibration or sensor fusion required). As you look from left to right or vica versa it will mark where the DMP yaw value crossed the mid-line and then retrospectively compare that with the magnetometer heading (which will lag a few ms due to the filtering).
End result should be comparable to the current auto-drift compensation but without the little 'pull' that you sometimes see when spring-back kick in.
cheers
Rob
if you want i can put all of it in a .zip file and upload it.
i used the 32bit ui files for this example. what version of processing are you using(mine is 2.2.1 64bit)? toxiclibs-complete-0020 is the correct version of the library.
edit: here's the zip file: https://www.dropbox.com/s/aubovyd5xr4f47a/EDTrackerUtil.zip
just choose "exctract here" and open the EDTrackerUtil.pde inside with processing, should work fine
I have just downloaded the zip, thanks for that but still getting the same results.
Stupid question time, but should I be trying this with the EDTracker plugged in?
very weird. i'm afraid i cannot help you any further
i just can't figure out what the problem could be. maybe some non-utf8 characters in the path to the processing directory / edtracker ui directory? (very wild guess )the tool should start up fine, no matter if the edtracker is plugged in or not.
That does sound intriguing. How high do you think that lag is going to be? Or would it be possible to combine the 6050 and the 9150? I'm thinking of using the 6050 for immediate movement detection and the 9150 as a means to correct the drift. Probably too much problems to get working right.
The MPU-9150 is a MPU-6150 with the magnetometer onboard so you could just add one to your MPU -6150 and do the fusion in the Arduino.
I'm really glad my custom board has now been adopted and is included in the guide as I feel I've contributed.
Last edited:
Congratulations! Unfortunately the only thing I'm capable of contribution right now are a couple of half ideas. This project is really something, and everyone involved is doing amazing work. Can't thank you guys enough.
Last edited:
Mine does exactly the same as this, just brings up a white square. I am running Windows 7 64bit, I have scanned through this thread for a solution but nothing I can find changes it.
Anyone discovered what causes it and how to fix?
Found the problem.
Was trying to run the UI in processing from a gash folder and not from */my Documents/Processing. Have moved the util folder and library to there and now it works inside Processing 2.2.1
EDIT:
The Util running under Processing has found the EDTracker2 straight away as soon as I plugged it in so all is (sort of) well in Bob_H land just need to get back into the instructions and get this thing calibrated now.
Last edited:
Right. It's working at last. Got it calibratedusing EDTrackerIICalibrate with apparently no probs - got both dots ob the centre of the retiule after 4 or 5 presses of 9 key. Loaded up EDTrackerII changing the expo values from 8 to 16 as per Brumsters suggestion.
The yaw drift is doing strange stuff though. It jumps down instantly to either 0 or -1 to -3 ballpark then slowly ramps up to a positive number before sort of settling then jumping back down to 0 or negative and repeating the ramp up. Rince and repeat. Yhe extremes of the ramp reduce over time but the yaw drift is defo not stable (although it does work in the game).
Will have another go at going through the full calibration in the morning but does it look to you guys like the MPU9150 is duff? Or do I just need to give it much more time to settle down while doing tha calibration and yaw drift?
The yaw drift is doing strange stuff though. It jumps down instantly to either 0 or -1 to -3 ballpark then slowly ramps up to a positive number before sort of settling then jumping back down to 0 or negative and repeating the ramp up. Rince and repeat. Yhe extremes of the ramp reduce over time but the yaw drift is defo not stable (although it does work in the game
).Will have another go at going through the full calibration in the morning but does it look to you guys like the MPU9150 is duff? Or do I just need to give it much more time to settle down while doing tha calibration and yaw drift?
The yaw drift is doing strange stuff though. It jumps down instantly to either 0 or -1 to -3 ballpark then slowly ramps up to a positive number before sort of settling then jumping back down to 0 or negative and repeating the ramp up. Rince and repeat. Yhe extremes of the ramp reduce over time but the yaw drift is defo not stable (although it does work in the game
Hmm.. Initiating a recentre will cause it to drop down to zero and then ramp up/down to a value??
Found the problem.
Was trying to run the UI in processing from a gash folder and not from */my Documents/Processing. Have moved the util folder and library to there and now it works inside Processing 2.2.1
EDIT:
The Util running under Processing has found the EDTracker2 straight away as soon as I plugged it in so all is (sort of) well in Bob_H land just need to get back into the instructions and get this thing calibrated now.
Hey Bob_H glad you managed to sort things out.
If you should get the time, could you explain your steps in more detail please as this has had me stumped for hours now and I think my brain is very close to shutting down :S
Phew, I finally got there! Thanks for the help guys
Yep, sorry, just got the PC turned back on. Yep, in the Processing folder that's in the same place as the Arduino folder you put the sketches for loading into the EDTracker in.
Onward and upward
Hopefully it's sorted now. I put the calibration sketch back in and started up the UI. I saw there was a +8° drift on it but approx every minute there would be a -270° drift spike. Hitting the 9 key a few times has gotten rid of the spike and the unit is now steady at +0.02° yaw drift with no spiking. Have just put the EDTrackerII sketch back in and am now monitoring it while letting the unit go through its 10 minute settle down before going any further. Fingers crossed!
Update:
OK looks like this thing is cooked and ready to serve
Looks like something went pear shaped in the calibration that was causing the spikes. It's sitting here in front of me now with the EDTracker2.0 sketch in it and on the UI the yaw drift is down to a steady +0.02° which, I assume, is pretty damned peachy.Just need to get it working in game now without it borking up my X52 Pro stick axis'
Last edited:
How to secure the EDTracker
I wasn't sure whether to ask this here or to start a new thread but here goes ---
Having taken delivery of all the components to make my EDTracker I plan to get the soldering iron out and see how many holes I can burn in the kitchen table like normal.
Once it's made and calibrated (following the excellent guidance) I want to stick it on the top of some headphones.
So the question is - how are people securing their EDTracker?
I want to do it in a way that is reversible (lesson learnt with epoxy resin) but secure enough to take the forehead hitting desk moments that accompany my flying skills (slight exaggeration)
I've seen one picture which implied it was more about securing the USB cable rather than the board itself. I know that there is some discussion about a caddy for it which could help but I'm not aware if anyone has that yet? Any guidance would be appreciated.
I wasn't sure whether to ask this here or to start a new thread but here goes ---
Having taken delivery of all the components to make my EDTracker I plan to get the soldering iron out and see how many holes I can burn in the kitchen table like normal.
Once it's made and calibrated (following the excellent guidance) I want to stick it on the top of some headphones.
So the question is - how are people securing their EDTracker?
I want to do it in a way that is reversible (lesson learnt with epoxy resin) but secure enough to take the forehead hitting desk moments that accompany my flying skills (slight exaggeration)
I've seen one picture which implied it was more about securing the USB cable rather than the board itself. I know that there is some discussion about a caddy for it which could help but I'm not aware if anyone has that yet? Any guidance would be appreciated.