Replacing the #freestylelibre – building the DIY CGM

Replacing the #freestylelibre – building the DIY CGM
Replacing the #freestylelibre – building the DIY CGM

There’s a kind of thrill to deciding to build your own bits and pieces of technology. It’s a bit of a challenge and knowing what should be in place once complete is exciting.

And there are further reasons for doing this given some of the issues that Abbott is now experiencing with delivering Libre supplies. It appears that while they opened up the list to additional users, they failed to match supply to the demand this has generated. As a result, to the average user it appears that every month they have supply issues of some sort that means any new orders get put on hold for a week or two. This generates a feedback loop, whereby all users then try and buy as many sensors as they are allowed, as quickly as possible, as soon as the supplies become available. I wonder how hard the guys at Abbott are kicking themselves, if they even are noticing? It’s not a great performance from a fairly major pharmaceutical company.
Anyway, Libre supply issues aside, let’s get back to the alternative. A cheap CGM solution that doesn’t cost the earth and delivers as good, if not better results! While I’m not the first to do this, and I know I won’t be the last, it’s fun to go back to something I last did in anger some twenty odd years ago. I might even re-learn coding if this goes well as I’ve some ideas that could be quite interesting for use of the Wixel and phone apps.

Wireless Bridge

Having ordered all the parts from eBay and various websites, they turned up over a week or so, including a Dremmel like tool, 15W soldering iron and multiple electronic components. Following the instructions on the xDrip website, building the wireless bridge was reasonably straightforward and took about 45-60 mins. Once connected up, it all appeared to be working correctly. The lights even flashed!
I’ve yet to flash the ROM on the Wixel, but this is a process that’s not so different from how we used to upgrade Android on phones 5 years ago, so I’m not overly concerned about undertaking that.


Now the transmitter resuscitation has proven to be more of a challenge. Once you start breaking into it, you realise just how much redundant plastic there is in the G4 transmitter, and that the way it is made means it is thoroughly a disposable item. Getting the Dremmel out, I started to grind away at the plastic. And I continued to grind away at it. It is, to put it mildly, a right pain in the behind. My original intention was to remove the batteries rather than grind them out, because there are some not very pleasant substances in batteries that i don’t want to turn into an aerosol. 
It became clear that the way these things are manufactured that this may not be an option. It seems the circuit is made with the batteries soldered into position, and then the entire component is encased in plastic. Some sections of it have way too much plastic. One might argue that the excess plastic means that the transmitter could be made significantly smaller, especially given the Libre sensor size and the slimmer G4 transmitters that were released. 
Back to my grinding, and having got down to the battery level, and destroyed the wheel I was using, I was forced into stopping:
As the images show, I have removed a lot of the plastic on the front end and just about hit the battery. I’ll be purchasing a new wheel and probably removing a whole lot more plastic in an effort to force the batteries out, as I don’t believe that grinding through them in a domestic setting is a particularly good idea. While the Zinc is considered to be very low toxicity to humans, it’s not so great for dogs and airborne Sodium Hydroxide just somehow doesn’t seem like a good idea. So I need to find a workshop with suction facilities to do this or I find another way. In theory, if the various items on the PCB can be recovered, a similarly sized arrangement could be made and a safer battery replacement module created. A bit of real hardware hackery and reverse engineering.


The xDrip software, by comparison with the above, is dead easy to install and set-up, and it allows for a very straightforward install and set up. As long as you have all your log-in details for your nightscout database, it uploads all the data for you. This is the item that gets the biggest and easiest tick in the box.


NightScout is a bit more of a faff to set up. You need to build a database, for starters, plus an application environment. Whilst it’s not an overly complex item to set up, it involves setting up a number of web services in order to make it run effectively. Having said that, it’s really a deployment challenge more than anything else (even if Azure can be a temperamental beast to engage and get to connect to GitHub) and the instructions on the NightScout Wiki are excellent. As a result, aside from a bit of time, you don’t need to sit down and take a big hit on the set up. you can do it as and when you have time.

Having said all that, the set up of Azure, GitHub and Mongo was all done from an Android phone, which probably made it more fiddly, but it’s an interesting aside that I can set up a cloud based application environment using a Smartphone as the only input device that I needed. That’s really very impressive!

So far then, I’m about three quarters of the way through. I don’t think there’s a vast amount more needed to get this complete. The sticking point is the transmitter. There’s a rather a lot of grinding and cutting to go, but I really don’t want to go through those batteries. 
Ah well… Step by step…


  1. Can you please tell me where the instrtuction of xDrip website is… THanks.
    “Having ordered all the parts from eBay and various websites, they turned up over a week or so, including a Dremmel like tool, 15W soldering iron and multiple electronic components. Following the instructions on the xDrip website,”

Leave a Reply

Your email address will not be published.