Thanks!
I don’t mind if others want to investigate this issue. It’s just that I, personally, am happy with the work-around that I have, so no one needs to help figure out the why behind it for my benefit.
I’m all for knowledge for it’s own sake.
Since it is a one-off a reasonable explanation is that when the Tandem receiver tried to “pair” it picked up the old transmitter which rejected the pairing code. In principle this should not be a big problem; the receiver should keep trying, the old transmitter will step out for 5 minutes and that should be sufficient time for the receiver to try again and pair up with the new transmitter.
It is entirely conceivable that the Tandem software doesn’t quite cut it. I will say that it took xDrip+ a while to get to what I regard as a vaguely satisfactory behaviour and xDrip+ can be updated every day; no FDA approval required 
TL;DR:
The Dexcom design is inherently defective. It uses a four digit pairing code (10,000 possibilities) and the transmitter ID (the thing being paired with) is identified by the leading characters “DXCM” then a four digit id code specific to the transmitter (3,844 possibilities).
The pairing device (receiver; Tandem, 'phone etc) just gets the first four digit code. So it has to attempt to pair with anything whose name starts “DXCM”. In software engineering terms it is in a promiscuous mode (that really is the term.)
In principle it is possible for the receiver to attempt to pair with anything which appears on bluetooth; a sort of “poly” mode. However there might be built in limitations in the bluetooth hardware, the bluetooth specification, the Android/iOS/Tslim operating systems which prevent this. It’s generally only something bad actors do, though in these days of CCTV surveillance maybe they are the good guys? I certainly don’t know.
The problem is that without this “poly” mode each potential partner has to be reviewed sequentially; in this case there is no multi-tasking possible for pairing.
So maybe it is reasonable for Tandem to utter the “30ft” comment; a lot of the time for a lot of the people 30ft means a rebar reinforced concrete wall in the way. Divide the customer support requirement by four and you make four times the profit.
A simple general solution is to use a tin can; while my own tinfoil arm experiment to make the G7 inaccessible to it’s many potential partners failed I’m pretty sure if I dropped the thing into a tin can (think Popeye with the crinkly lid; close the lid) it would never have tried to pair with Olive again.
2 Likes
Thanks! Good explanation, and useful advice.
Plus, I love the mental image. I’m old enough to have watched Popeye the Sailor on Saturday morning cartoons. 