These are chat archives for nightscout/intend-to-bolus

4th
Jun 2015
Scott Leibrand
@scottleibrand
Jun 04 2015 04:29
@bewest you around?
Ben West
@bewest
Jun 04 2015 15:40
@scottleibrand sort of
Toby Canning
@TC2013
Jun 04 2015 15:55
@bewest How much effort do you think it will take for Geoff to make our Java based OpenAPS work with the CC1111?
My motive in asking this is to see if it is even worth the effort--given that it is unknown how much, if any, the communication distance will improve.
Eric
@ecc1
Jun 04 2015 17:18
I'm able to talk to my pump from a CC1111 dongle at about 10 feet but so far I've only tested small commands and responses (like pump ID and battery level). CRC errors will go up quickly with packet size so it may not be reliable enough at that distance for retrieving bolus history etc.
Toby Canning
@TC2013
Jun 04 2015 17:19
@ecc1 Thanks! 10 feet with minor obstructions or 10 feet line of sight only?
Eric
@ecc1
Jun 04 2015 17:28
Pretty much line of sight, but haven't done extensive tests yet.
It's frustrating that they use such a primitive RF protocol (for backward compatibility, I'm sure) when it could be made much more resilient to noise etc.
Pete Schwamb
@ps2
Jun 04 2015 17:53
What kind of antenna on your cc1111 dongle?
Pete Schwamb
@ps2
Jun 04 2015 18:26
Does the OpenAPS have a set of tests cases written around it? I.E. given these input parameters (IOB, BG, COB), expect this output (set basal to x)?
Toby Canning
@TC2013
Jun 04 2015 18:28
It is in the repository I added you to, in the module called APSlogic
Pete Schwamb
@ps2
Jun 04 2015 18:28
ok, thanks.
Toby Canning
@TC2013
Jun 04 2015 18:28
Scott wrote pseudo code for everyone else, I can find the link for you
OpenAPS doesn't include the actual code (not a public version) due to a variety of factors, including US rules/regulations
Pete Schwamb
@ps2
Jun 04 2015 18:34
cool, that’s useful for understanding how decisions are made. I was just curious, and trying to understand how feasible it would be to put that code in embedded sw.
Toby Canning
@TC2013
Jun 04 2015 18:35
My guess is that it is doable
Pete Schwamb
@ps2
Jun 04 2015 18:35
having a codified set of tests/expectations could help verify correctness across implementations.
Toby Canning
@TC2013
Jun 04 2015 18:35
We thought the same thing!
Pete Schwamb
@ps2
Jun 04 2015 18:35
And would greatly help with porting the implementation to a new platform.
Toby Canning
@TC2013
Jun 04 2015 18:36
It would be nice to have a way to "replay" a set of scenarios to validate when changes are made
Dana Lewis
@danamlewis
Jun 04 2015 18:46
Agreed re: tests, if anyone wants to start making a list/documenting :)
(Also re pseudo code, if you need to find it again and don't have it bookmarked, there's a link buried deep at the bottom of the first email in the "getting started" thread in openAPS-dev)
Eric
@ecc1
Jun 04 2015 19:23
@ps2 It's a PCB trace antenna. Same on both the TI and Ciseco dongles; looks like it's the one in the reference design available on the TI website.
Pete Schwamb
@ps2
Jun 04 2015 19:27
I’ve had a hard time getting good range out of these: http://www.ti.com/diagrams/med_cc1111emk868-915_cc1111_usb_dongle_low_res.JPG
That’s a ceramic chip antenna
This style of pcb trace (big) http://www.ti.com/graphics/tool/CC1110DK-MINI.jpg, works quite well, as does the coil antenna I use on the RileyLink (http://www.mouser.com/images/linxtechnologies/images/ANT-xxx-JJB-RA.jpg).
And by good range, I mean across walls/floors of house.
Eric
@ecc1
Jun 04 2015 19:41
Oh, my mistake then. My dongles use the chip antenna you posted first. It doesn't work through walls very well for me either.
Toby Canning
@TC2013
Jun 04 2015 19:52
@ps2 Have you heard of a service that would assemble the rileylink and sell the completed product?
Pete Schwamb
@ps2
Jun 04 2015 19:58
Yeah, I looked at (http://www.smallbatchassembly.com/). For a small run (40pcs), Looks like it would be about $62/board ($36 in parts, $26 assembly).
Toby Canning
@TC2013
Jun 04 2015 19:59
That's pretty awesome.
Pete Schwamb
@ps2
Jun 04 2015 19:59
That would be ordering the parts myself, supplying them to SBA, and dealing with all selling/shipping. Would probably do it through tindie or something like that.
Toby Canning
@TC2013
Jun 04 2015 19:59
How long does it take to put one together yourself?
Pete Schwamb
@ps2
Jun 04 2015 20:01
Also not sure that SBA would mount the antenna; that might be a post-assembly job as well.
A couple hours, if everything goes well. It sometimes doesn’t. :)
Also, I’m worried that my parts orientations in Eagle might not match up to how they come on the tape, and that I’d have reversed parts. Not sure how to deal with that.
Toby Canning
@TC2013
Jun 04 2015 20:09
I'm going to see if I can find a service that will put together 3 for me if I send the parts. I really don't care if it somewhat pricey if they do a good job!
Do you think there will be any issue if I go with a bigger battery? Preferably, I'd like to get 36 hours of runtime
Toby Canning
@TC2013
Jun 04 2015 20:31
@ianjorgensen Send me a private message when you get here
Pete Schwamb
@ps2
Jun 04 2015 20:34
I would guess most li/lipo batteries would work. Here is the datasheet on the charger: http://www.mouser.com/ds/2/268/21984e-53484.pdf