Well, technically spin and about a meter to the right lol. But awesome to know! Thanks!
DIY Head Tracker For A Tenner
- Thread starter Dead Fred
- Start date
Well, technically spin and about a meter to the right lol. But awesome to know! Thanks!
Urrgh, Windows USB strikes again! I'm sorry, but I'm out of ideas, you've done the sensible thing and if it detects on another PC then I think you've ruled the device out as the problem....
as you correctly say, it should show up as a composite USB HID device... for the joystick part. The COM port is just used for flashing/calibration, basically serial I/O to send commands and receive results back... plus a different COM port for the Arduino bootloader (for flashing). This bit is obviously working.Urrgh, Windows USB strikes again! I'm sorry, but I'm out of ideas, you've done the sensible thing and if it detects on another PC then I think you've ruled the device out as the problem....
Well a fresh install of Windows 7 has cleared everything up. Thinking back, I'd been experiencing some USB oddities for a while now. One or two devices unable to load drivers etc.
I started head tracking with Freetrack and a homemade 3 point LED hat. It was fantastic but it became apparent that Freetrack was no longer being developed so when FTNOIR introduced a point tracker I jumped ship and enjoyed the extra smoothness of the improved filtering. It occasionally jumped around a bit if I looked too far in one direction and one of the LEDs lost tracking but on the whole was a great success.
Now, EDTracker has taken things to a whole new level. The range and smoothness of movement is beyond anything I could have expected or hoped for. I will probably still use FTNOIR and the magic hat for Rise Of Flight as there are benefits to 6DOF in that game but for pretty much everything else it'll be EDT and Opentrack from now on.
You really should legally protect the design somehow. Maybe just having it out in the open like this will provide some protection from copycats. And thanks for sparking my long-forgotten electronics skills. I have another couple of Arduinos on order for projects (unrelated to EDT) that have bubbled into my head.
Just to check, your not trying to plug into USB3 are you? i did that and windows couldn't detect what it was. was fine once in usb2
I tried all sorts while it wasn't working. It seems fine everywhere now Windows 7 has been refreshed, although USB2 is the most convenient socket.
Ok, time to tease you all a little bit, because we're at a good place right now where confidence is starting to build on "EDTracker Pro", to the point where we have a working prototype now, albeit the wireless still needs work. Some of you that have followed the thread will know we had this in the background, it's been an ongoing piece of work since middle of last year... I started designing it late summer and finished it off in October. It then went through a redesign to reduce costs as much as possible and we sent the design off to fabrication just before christmas. Then, early January, 30 of the following turned up on my doorstep
We've invested a bit of cash from the sale of DIY EDTrackers into buying enough parts to make 30 prototypes which, at such low volumes, is not cheap but the idea is, we get 30 prototypes done, get them out there in the field being tested and price up construction of the device somewhere more cost-effective like China. If it works, great, if it doesn't - well, nothing ventured nothing gained... but none of the EDTracker gang want to spend the rest of their lives soldering
So I set off with the solder paste, hot air gun and painfully started to hang-assemble the first few prototypes... a nerve-racking process!
Several weeks later, with slow process made on the oscilloscope figuring out why certain aspects weren't working (I stuffed up the xtal pads :doh, we got the basic one working with just the AVR on it. Then we added the MPU-6050 and proved that working. Then yesterday Rob and I got the HMC magnetometer (much better than the 9150) working. So prototype #005 rolled off the production line and gets us to where we are today with the DIY EDTracker....
Yes, it's Lipo powered for the wireless versions. Charges off USB and in basic tests I had it lasting well past 12 hours before I gave up measuring any more. All on a single board designed to fit into our good friend, the Hammond enclosure, that many people use for devices today. The eagle-eyed amongst you will also notice take-offs for 3 infra-red LEDs, just as an option for those that want one, two or three point tracking for 5/6DOF. Untested and unproven yet, will greatly depend on power demands (possibly wired-only mode) but since a few people asked for it, made sense to put it on there just in case.
Just about fits in the box...!
And bob's your uncle. No buttons any more (not really needed) and still a sensible size...
Now before we get swamped with requests.... schtop schtop.... this edtracker isn't ready yet.... we have lots of coding to do to squeeze down the firmware to fit all of the features into this 32Kb AVR. Lots of field testing (yes, we haven't forgotten those who asked, we will contact people when we're ready I suspect we will look to people who've been pivotal in helping us along this journey) and then - the bit I'm dreading - finding out how much it's going to cost to produce in volume. So we're still some weeks/months away, and it might not even succeed, but so far it is looking promising, with a mag-corrected wired EDTracker Pro working as intended.
Excited and knackered all at once, me
We've invested a bit of cash from the sale of DIY EDTrackers into buying enough parts to make 30 prototypes which, at such low volumes, is not cheap
but the idea is, we get 30 prototypes done, get them out there in the field being tested and price up construction of the device somewhere more cost-effective like China. If it works, great, if it doesn't - well, nothing ventured nothing gained... but none of the EDTracker gang want to spend the rest of their lives soldering So I set off with the solder paste, hot air gun and painfully started to hang-assemble the first few prototypes... a nerve-racking process!
Several weeks later, with slow process made on the oscilloscope figuring out why certain aspects weren't working (I stuffed up the xtal pads :doh
, we got the basic one working with just the AVR on it. Then we added the MPU-6050 and proved that working. Then yesterday Rob and I got the HMC magnetometer (much better than the 9150) working. So prototype #005 rolled off the production line and gets us to where we are today with the DIY EDTracker....
Yes, it's Lipo powered for the wireless versions. Charges off USB and in basic tests I had it lasting well past 12 hours before I gave up measuring any more. All on a single board designed to fit into our good friend, the Hammond enclosure, that many people use for devices today. The eagle-eyed amongst you will also notice take-offs for 3 infra-red LEDs, just as an option for those that want one, two or three point tracking for 5/6DOF. Untested and unproven yet, will greatly depend on power demands (possibly wired-only mode) but since a few people asked for it, made sense to put it on there just in case.
Just about fits in the box...!
And bob's your uncle. No buttons any more (not really needed) and still a sensible size...
Now before we get swamped with requests.... schtop schtop.... this edtracker isn't ready yet....
we have lots of coding to do to squeeze down the firmware to fit all of the features into this 32Kb AVR. Lots of field testing (yes, we haven't forgotten those who asked, we will contact people when we're ready I suspect we will look to people who've been pivotal in helping us along this journey) and then - the bit I'm dreading - finding out how much it's going to cost to produce in volume. So we're still some weeks/months away, and it might not even succeed, but so far it is looking promising, with a mag-corrected wired EDTracker Pro working as intended.Excited and knackered all at once, me
Brilliant.
I get 9+hrs out of my wireless headphones so anything approaching that is just going to be perfect.
The extra 3 IR LED connections are to give full 6DOF I assume. Think the ones I use might deplete that battery pretty quickly. Still, I seriously cannot wait for this to become reality.
Awsome work and congrats to all the team.
BTW +1 on the choice of the HMC Mag. Use them in my hexacopters, utterly reliable and low noise.
I get 9+hrs out of my wireless headphones so anything approaching that is just going to be perfect.
The extra 3 IR LED connections are to give full 6DOF I assume. Think the ones I use might deplete that battery pretty quickly. Still, I seriously cannot wait for this to become reality.
Awsome work and congrats to all the team.
BTW +1 on the choice of the HMC Mag. Use them in my hexacopters, utterly reliable and low noise.
Last edited:
Yep, in conjunction with something like Opentrack or whatever 3rd party tool people prefer. I doubt we'll look to do the software ourselves, it just doesn't make sense when there are guys like Stanislaw out there who know the ins and outs of it far more than us.
The LEDs are driven by PWM so that should hopefully reign in their power usage a little, and we can tune/tweak brightness versus battery life that way.
Cheers buddy ;-) yes, Rob's been a fan of the HMC5883 right since the early days. It's just simple and reliable. Way forward when you've got the control like this to put on there what you want...
It has been moved from "Fabled" status to "Hotly Anticipated" status
Brilliant news. I know when I recently got into RC cars again for a while. A electronics engineer designed and had his programmable electronic speed controller and had it fabricated in china. With his own specked components toshiba fets etc. He got the whole thing done incredibly cheap. He did have to order quite a few units though. And the bad side is that knockoffs did eventually appear on Ebay.
Personally I'd love a wireless unit. The microUSB socket tends to be the weak point on these cheap boards. I had to find the lightest cable I could find and then fix the cable to my headphones or hair band with some extra slack at the plug because over time it's almost inevitable that it's going to get yanked on occasion. (With the hairband it's so light I often forget it's on my head lol).
Personally I'd love a wireless unit. The microUSB socket tends to be the weak point on these cheap boards. I had to find the lightest cable I could find and then fix the cable to my headphones or hair band with some extra slack at the plug because over time it's almost inevitable that it's going to get yanked on occasion. (With the hairband it's so light I often forget it's on my head lol).
looks great, I assume you'll kickstarter the production? my current EDTracker (9150) is working very well, but assuming the price for this doesn't end up too far North I'll be in
Yea, superb toy! Bought a Arduino Micro and a 6050 from Ebay, together less than 10€, soldered some wires, calibrated it - ok, then I had some fuzz configuring it in E: D, but this was my fault. Don't bind it to digital headlook, use the analogue entry...
I don't have the magnetometer version, and I have no big issue with drift, although calibration was done QnD.
BUT I have trouble using it. When flying, I have problems aiming at the right point, when mining, I have problems targetting fragments and some more. Right now I tend to switch headlook on when dogfighting and off for all other situations.
Anyway I believe this is my personal problem, maybe I have to get used to this device some more hours. Again: great toy, fantastic work!
I don't have the magnetometer version, and I have no big issue with drift, although calibration was done QnD.
BUT I have trouble using it. When flying, I have problems aiming at the right point, when mining, I have problems targetting fragments and some more. Right now I tend to switch headlook on when dogfighting and off for all other situations.
Anyway I believe this is my personal problem, maybe I have to get used to this device some more hours. Again: great toy, fantastic work!
Yes, once you pay for the tooling it costs pennies per item to run off a few thousand more clones. Protecting the software and updates somehow might be possible in this case, though it only makes it slightly harder to clone the whole package. Branding, accessories and packaging, distributing manufacture across different companies, and controlling the sole genuine retail channel might help. But it won't stop cloning if there's a profit in it. Tough business model for the long run.