Having unsuccessfully played around with trying to bring Dexcom G4 transmitters back from the dead, failing three times, I finally bit the bullet and paid for a transmitter. And some sensors.
And finally completed the xBridge build.
I am now a fully paid up member of CGM in the cloud, alarms, and everything…
So how did I get here? Well the build and soldering was fairly straightforward. I wrote about that previously, here. But I hadn’t attempted to build the Wixel app and get it set up as I didn’t have a functioning transmitter.
Well wait no more. With a functioning transmitter, I was able to build the files and get it loaded up. But it wasn’t dead straightforward.
No I tend to operate on Windows and Linux, and to be honest, I think the windows route is much harder, especially with Windows 10. Missing DLLs and odd set-ups mean that while making the file is easy, getting the PC to talk to the Wixel proved to be rather more of a challenge than I was expecting. After a few attempts, I got fed up and switched to Linux.
Now Linux is a platform that is seemingly designed for this. If you are used to using a command line, then it is much easier to set up. as ever, making the appropriate file was no problem. But once again I ran into a couple of little issues with getting it onto the Wixel. After moving directories around and lots of silliness, I eventually cottoned on that I needed to su to root to run the appropriate utility and got on with it. Once in root, it was all very quick and simple, and I’m very pleased with the result. Everything works very well, and with HAPP configured, it’s interesting to see what the Open Loop APS throws up in terms of what I eat and how my levels manage.
I’m currently running the software on a Sony Xperia M4 Aqua, although I’ve a cheap Chinese Android phone on the way with a huge battery, with the aim of seeing how long I can get the thing working between charges. I’m hoping that I can get away with three or four days, but we’ll see!
Here we have two screenshots. The first is xDrip, doing its stuff:
That was quite a good trace. It seems to work well and track comfortably.
The second is HAPP, the hackability artificial pancreas project, by Tim Omer and Friends. I’m interested to see how this one plays out and what I end up doing with my basal as a result:
Now this wouldn’t be a proper Cloud based thing without it linking to nightscout, which of course I’ve done. I’d already got that set up, so it was a very simple next step to link it up. It also means that I can keep an eye on things using my iPhone rather than everything else, which can stay suitably hidden away, not causing issues and getting in the way. This is mainly being done for the technical curiosity that I have rather than anything else, but if I have nightscout set up I can at least use an app on my Apple watch to glance at my bg level.
I’ve set up the Sony Smartwatch 3 with the Android phone, and let’s face it, at £89, it was a bit of a bargain, and now have a real-time monitor on my wrist, to simply glance at. HAPP data is also displayed, as you can see in the picture, and if it recommends additional bolusing or a temporary basal rate, I can simply accept it from the watch. All in all, a very satisfactory system.
Sometimes, you’ve got to love the technology…
One thing I would note though, is take care in the soldering and covering process. Having got it built, I failed to properly cover the battery contacts effectively and managed to short the battery out. This meant I was without the system for 12 hours as I was out. It also did something odd to the Wixel, and the only way to fix it was to reset the Wixel and reload the app. Not really an issue, but something to be aware of.
Now, all I need is a pump to talk to the phone, then we can start to close the loop. HAPP so far has support for the Dana R, but the more available pumps (Medtronic) for example are less easy to get our hands on. Likewise, the Roche Spirit Combo has a hidden bluetooth PIN, and so far Tim hasn’t been able to get the pump to reveal its secrets. If someone at Roche wanted to help us with doing that, it would be very welcome. I believe that other “Official” AP projects have been given this, so it would fantastic if Roche would help us to achieve the same ends!
We’re also welcoming any contributions of unwanted pumps, so if you have one of the below pumps from Medtronic that you no longer use, please get in touch!
- 523/723 (with firmware 2.4A or lower)
- 554 (European Veo, with firmware 2.6A or lower)