CGM on the 640g: Connectivity

I thought it might be worth a quick post on the CGM hardware that comes with the Medtronic 640g and to focus, for this post, on the wireless connectivity, with a view to getting NightScout on this system at some stage... (for us, this is probably some way off, but I hope some of this might be useful to others who may be in more of a hurry to connect to the Cloud!)

You get the (Enhanced / New) Enlite Glucose Sensor and a Guardian 2 Link transmitter.
There's an insertion device for the Enlite, which, fortunately, hides the needle out of sight...

 

The Guardian 2 Link transmitter comes with its own mini (AAA battery powered) charger, taking up to an hour to charge. Ours took around 40 minutes (or four Mr Men books in Janki's time-frame) to recharge after 7 days of use) and a test clip, which is also used for cleaning (it's waterproof once this is connected or once it's clipped on to the Enlite - we've had no problem with swimming and baths).

I've been following the fantastic NightScout work and the port of that work to the Medtronic Veo with the original Guardian link transmitters (the UK Facebook group is here). What is known is that the kit developed for the Veo CGM won't work with the 640g. They use a different radio frequency for a start: the original Guardian transmitter uses a ~900MHz band (868 or 916MHz depending on your region) whereas the 640g's Guardian 2 uses the 2.4GHz band and the IEEE 802.15.4 standard.

I was quite surprised to learn that the 2.4GHz band was being used (well I guess the Contour Next Link 2.4 BG meter name as a give away!). It's a very crowded spectrum, probably the most popular of the "Industrial, Scientific, Medical (ISM)" bands, where low powered devices can operate without license. Cordless phones, WiFi routers and devices (this laptop, my mobile phone's WiFi connection, smart tellies, some printers, cameras etc). video senders, the baby monitor, the door bell are all likely to be using the same frequency space, with the potential for conflict.

Please note this doesn't mean it will cause an error or malfunction, as the device should have adequate error checking on-board to spot if its Sensor Glucose (SG) reading data-packet has just been hijacked by the door bell... but it might mean interrupted / lost transmissions from the Guardian 2 and the pump and also to/from the Bayer Contour Link 2 BG meter.

Using the IEEE 802.15.4 standard should mean that it'll stay clear of the most common 802.11 (Wi-Fi) channels: these are probably the most high-powered 2.4GHz transmitters in your house and so the most likely to interfere. Indeed, the manual confirms it operates at 2420, 2435, 2450, 2465 and 2480 MHz, with a 5MHz bandwidth on each channel. In the 640g configuration, the pump is the coordinator (ie it decides which channel is best to use) with the transmitter acting as the node (ie it listens for the beacon from the pump to correctly identify the channel to broadcast on). The transmitter also re-transmits up to 10 hours of missed data packets (your raw, uncalibrated, SG readings) to the pump should comms be lost.

The 640g manual stipulates that you shouldn't use the pump within 30cm of cordless and mobile phones and 1m of Wi-Fi devices (Bluetooth devices are 10cm). Um, OK, I'll try, but that's virtually impossible in most households (or in most workplaces)... When she's older, Janki can either leave her mobile at home or leave it turned off like her Gran ;-) It's not uncommon (or unreasonable) for Medical Devices to stipulate very cautious RF levels and my experience in a hospital environment is that it usually takes vastly more RF power to cause any ill effects, but I certainly don't intend to ignore them and will not be deliberately placing a mobile next to the pump or resting the pump on top of a laptop. These limits are all about restricting the likely signal strength from each device at your pump / transmitter: a mobile phone is a powerful RF transmitter - close to the 2.4GHz band; WiFi lower strength on the 2.4GHz band, Bluetooth even smaller, again on the 2.4GHz band, and hence the reduced distance requirement for each device group as RF power decreases).

I should say, we've not experienced any issues so far in terms of data loss, obvious interference or extended delay in data delivery. I would be interested to see how it performs in particularly dense wireless environments, such as a crowded coffee shop (goodness knows how it would get on a tech conference, but I don't think Janki is likely to select that over Peppa Pig World for a while...) or with the usual interference suspects, such a cheap video senders and remote control cars, some of whom struggle to keep their transmissions to a single channel.

The most likely scenario appears to be that if a channel is taken and the CGM-pump connection is interrupted, the pump will scan for a new channel after five-ten minutes (ie once it's missed its regular five minute data packet), find a clear channel, re-associate with the transmitter and get the delayed SG data, so maybe with a maximum 10-15 minute delay in packet delivery, unless you're in such an RF unfriendly environment that all channels are being hogged by someone's remote controlled drone off Ebay or something similar... In practice, we see between zero - two 5 minute packet delays in each 24 hour period (with the exception of using the CGM swimming of course, when the signal is going to be attenuated by the water).

What isn't mentioned - and what appears to be confirmed absent by scanning - is Bluetooth (BT). There was a hint that the Guardian 2 transmitter might be BT enabled (which would be based on IEEE 802.15.1 standard, but now maintained by the BT Special Interest Group, and also operates at 2.4GHz), allowing it to be ready for the Guardian Mobile monitoring app that was previewed last year. That may have to wait for the Enlite 3 (see Slide 8 of this PDF presentation), as this appears to confirm that it's the Enlite 3 that'll do the on-board algorithm work to be able to send SG data in mmol/L to an app. This bit is speculation with a sprinkling of insight so please feel free to correct / update!

I've not had time yet to scan and dump the Guardian 2's data packets to a terminal window to take a look. One key thing that isn't known (as far as I know) is whether the data packets are encrypted or sent in the clear. There's a lot more digging to do, but that'll have to come after sorting out our basic pump settings. I hope this is of some help for now.

Update: See also this comparison of connectivity on two different 640g systems.