It's been a busy year, it's about time we updated you all on progress with wireless!

2016 Success

As we approach the festive season and the end of 2016, we thought we'd look back and reflect on the past 12 months (and more!), what we've achieved, the successes (and failures) and where we're heading for the next 12 months. 2016 saw big uptake in both the home-made DIY EDTracker and our more commercial EDTracker Pro model. We continue to receive great stories from you guys on the ingenious ways you put EDTracker to use. Outside of head tracking, we've had stories of people using EDTracker as an alternative flight controller, as part of DIY VR headset solutions, as accessibility products for people with disability, as a gun position tracker for a multi-player VR experience and even by game audio engineers to help them during development of positional audio for VR games. The DIY website (www.edtracker.org.uk) took 8.9M hits and in November alone 11,328 people fired up the DIY EDTracker UI - that's an average 377 launches every day! The DIY GUI software has been downloaded an amazing 59,870 times so that suggests the number of users out there is will into 5 figures, even accepting people will download more than once. Likewise the Pro has seen increasing growth over the year, with 5.6M hits over the 12 month period and pretty much the same (11,116) number of launches of the UI within November.

We had to register for VAT earlier in the year but we decided to absorb the cost and kept our prices low - EDTracker is now even greater value for money compared to the competition. We have shipped to pretty much every corner of the globe. We sponsored the Elite Dangerous-themed community event, Lavecon, back in July. It was great to put a little back into the community that helped EDTracker come to fruition. We've also seen player bases extend greatly across other great games such as DCS, Prepar3d/FSX, Arma 3 and War Thunder (amongst many others). We even had one ingenious user work out a way of getting EDTracker working for head look in No Man's Sky! And, of course, we have invested money back into the product to help develop enhancements including a new wireless model for 2017.

On behalf of everyone at EDTracker, we would like to thank everyone who has helped us with this venture - there are far too many to list - including all of the users who spread the good word (through no pushing from ourselves) across the social media and YouTube channels of the internet.

Wireless Update

We know - many of you have been waiting for the release of the wireless version of EDTracker Pro. We receive enquiries pretty much daily asking when it will be available. We are now in a position to tell you some good news, and we're also going to be completely open with you on why it's taken so long!

Original batch of prototype PCB

Back in late 2014 (yes, honestly!) we'd already made a start designing the wireless model, and by Spring 2015 the first PCBs arrived in our laps. We set about hand-assembling the prototypes, which took an age in itself as all the components are surface mount and a mix of 0604 and 0402 package sizes... but, nonetheless, the first prototypes started to breathe and we managed to get some degree of performance out of them. However, they used a PCB trace antenna and, despite our attempts at tuning it, we could not get a signal performance we were happy with.

This is not unusual with trace antennas, so we soldiered on resolutely!

For a brief period, other things took priority so we set about updating the website and getting our internal business sorted, and registering the company and the EDTracker brand. We then re-visited the design and re-worked it to a surface-mount ceramic chip antenna, which would simplify the build and save us space on the board. The device at this stage was still based on the Atmel 8-bit microcontroller, the ATMEGA32U4 (as serves us well in the wired model, and is also used by the majority of people who build a DIY device). We were making use of 2.4GHz ISM wireless - initial tests with bluetooth were laggy and disappointing. This would mean we needed to design and build some sort of USB dongle to fit into the user's PC, to receive the signals and present them to the operating system.

After much digging around, we managed to source a supplier in China of suitable dongles, so we headed down that road rather than making our own - it was simply more cost effective. Things were looking promising.

SMA test donkey
Finished Early Prototype

The prototypes went out to various people to test and, on the whole, the results were promising but there was quite a variance in the range that people were getting, and the behaviour they were seeing. We were racking our brains trying to understand how several identical prototypes behaved differently from one person to another. Eventually we narrowed it down to the dongles, which seemed to have a less-than-ideal antenna design but also a rather "variable" approach to quality control! We suspect some of them may have been using fake Nordic chips - there are certainly reports out on the internet suggesting this could be a problem. Our antenna also needed a little more tuning; the process of RF component design is one with a steep learning curve!

On top of this, we then discovered that Honeywell, who make the excellent magnetometer that we use in EDTracker Pro, were announcing discontinuing of the HMC 5883L magnetometer that we used. We still had a while before they would dry up, but we needed to plan ahead for a redesign at some point. Since submitting the device for wireless testing and CE/FCC approval was an expensive business, we really wanted the design to be final. So it made sense to redesign now (this was early 2016) - do it right, do it once.

We made another batch. The new design would move to the Invensense MPU-9250 combined gyro and magnetometer that was becoming very popular in DIY build; it seems an obvious evolution from the MPU-6050 we were already using (itself just a gyro). The long and arduous task of building them by hand started once more, and porting the code to the new chipset, but we had the results fairly quickly and the product looked good to go. We found a supplier of a revised dongle design that showed much better performance and reliability. Then we got thrown a curve ball.

In short, 1 in 10 of the MPU-9250 ICs showed sub-standard performance way outside of their advertised specification. The magnetometer in them was just too noisy. To cut a long story short, conversations with the vendor were never going to fix it so we had to take it on the chin and do another re-design. Now summer 2016, this was pretty annoying but we were so close it wasn't time to give up. So we took the opportunity to re-group, to evaluate the entire design... if we're going to have to re-design, we might as well take the opportunity to benefit from everything we'd learned to date.

On the analyser!
Final prototype

We evaluated several magnetometers and settled on a replacement. But we didn't leave it at that. We re-positioned the antenna to maximise performance following the tests we'd done with the previous prototypes. We re-positioned the gyro and mag ICs to be perfectly axis aligned on the device. And the ultimate change - we have moved to a vastly more powerful microcontroller, a 32-bit ARM Cortex M0 based chip with 8x the memory of the previous device, 48MHz CPU frequency (versus 16MHz of the old) and better power management features too.

We found a great and local prototyping house who worked with us quickly to see the prototypes built for us in just 2 weeks - while the cost was significantly more than hand-assembling, we just couldn't afford the delays any more. And so, in November 2016, what we hope will be the final prototypes arrived on our doorstep. Initial testing was flawless (at this point, the design was in by several revisions so it was pretty much "down pat" at this point!). The signal strength was vastly superior - we now get an indoor range way in excess of what any person is really going to need. And the responsiveness from the faster processor is pretty mind-blowing when you come from a wired device - wireless now shows performance as smooth as the wired (if not, we suspect, slightly better).

So the question everyone is asking, of course, is "when can I buy it?". Right now (December 2016) it's about to be submitted for RF testing to get its CE and FCC approvals. Then we can legally sell it across the world. Of course, being the Christmas season, this means the approval is not likely to come this side of the New Year, so we expect the results in January. At that point we can give the factory the "go" to start manufacture, and the typical turnaround from the factory is around 6 weeks. So the realistic answer is, *probably* around 1st March... although as with all these things, so many external factors out of our control could delay it that we can't promise it. But we are certainly at the point where we can see light at the end of the tunnel.

Believe me when I say we're just as frustrated at the delays as you guys probably are! But these things happen, and the one thing we pride ourselves on is our integrity and attention to getting the product right... and we were not going to cut corners just to punt something out to satiate the masses. This will be the first truly Wireless (ie. a device emitting RF signals, as opposed to "wire-less" or "without wires") head tracking product on the market and we believe it will be worth the wait. So thanks for your patience and bearing with us.

Here's wishing you all a wonderful festive season, a great new year... and a successful 2017!

Final solution