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

27th
Jul 2017
LilDucky
@LilDucky
Jul 27 2017 01:39 UTC
@fiberfan if @rodzilla55 does not take it then I would love it for R&D
Lin
@DengSir_twitter
Jul 27 2017 01:41 UTC
@scottleibrand I deleted the MLAB database, the openaps still did not run again, would I ask where should I check?
屏幕快照 2017-07-27 09.44.08.png
Joey
@Cooljoe04_twitter
Jul 27 2017 02:17 UTC
@cameronrenwick The more I am seeing and reading about Fiasp, especially seeing it in use with your openAPS setup, the more interested I am in getting my hands on it and trying it out. Wish they would hurry up and approve it in the US to make my life easier/make it easy for me to get! :smile:
Michael Spradling
@M1Sports20
Jul 27 2017 02:45 UTC
@scottleibrand: I see the three lines described nicely in "Understanding “Advanced Meal Assist” (AMA) and the three purple lines" I can't seem to find an equal section under smb section with the four purple lines.
i could just be missing it.
Scott Leibrand
@scottleibrand
Jul 27 2017 04:20 UTC
@M1Sports20 the fourth is the UAM prediction line, then. When UAM is enabled (via bolus IOB), and positive deviations from BGI are occurring, oref1 calculates the rate of peak deviation over the recent past, and projects the current rate of deviation to drop to zero at the same rate as from the recent peak until now. Sometimes that results in a prediction line below the IOB one (which projects current deviations to gradually decrease to zero over one hour), but with increasing or sustained deviations it can be much more up and to the right.
Please go ahead and PR something like that into the docs wherever you think is appropriate.
Or I can clarify further if you have any questions.
Scott Leibrand
@scottleibrand
Jul 27 2017 04:26 UTC
Side note: "positive deviations from BGI" == "insulin counteraction effect" in the Loop iOS app now, FWIW.
Tom Petrillo
@tompetrillo
Jul 27 2017 04:51 UTC
If my blood sugar oscillated 10-20 points through out the night with basals going high then going low (on openaps), what does that mean? Does that mean i need to raise my baseline basal rate? I'm not running any of the advanced features except for "meal".
Tom Petrillo
@tompetrillo
Jul 27 2017 05:00 UTC
also, what is COB? Carbs on board?
Tim Street
@tim2000s
Jul 27 2017 05:52 UTC
@tompetrillo It's probably worth you taking another read through the docs.
Phil
@drnoname82_twitter
Jul 27 2017 10:58 UTC
I have a problem with the openAPS pill in Nightscout. I just installed oref0 current master for the first time and started looping. The rig is looping as per the logs, it pulls CGM data from Nightscout, but the openAPS pill keeps displaying "unknown" and no forecasts are available. I set all the Nightscout variables according to the docs. Any ideas?
fidoman
@efidoman
Jul 27 2017 11:13 UTC
@rodzilla55 - there is a company working on a bionic insulin pump that has two separate chambers (one for insulin and one for glucagon). It uses one infusion set with two tubes. It is in early human clinical testing so no matter what the articles say about it, I expect it will take years to get through the FDA. http://www.bu.edu/today/2017/bionic-pancreas-clinical-trial/ I don't think it is possible to take an existing pump and modify it to support delivery of two hormones. The entire physical and software setup in today's pump only support one reservoir and singular medicine delivery. The obvious options for open source are to use two loopable pumps (one for insulin and one for glucagon) or wait for the bionic pancreas and hope we can send insulin and glucagon delivery commands to it OTA.
Joey
@Cooljoe04_twitter
Jul 27 2017 11:25 UTC

@ericfeibelman I find it interesting that the bionic insulin pump you were talking about uses a singular infusion set and not dual sites/infusion sets. That would mean in order to dose any glucagon to treat a potential low, they would essentially be first dosing at least an additional 6 units or so (in my case) of insulin just to get the glucagon pushed through the tubing and to the injection site. Obviously this value would vary based on the length of tubing they are offering and would be something they would have to factor into their calculations internally.

It is encouraging that someone is working on that kind of setup at least. Hopefully when they get around to releasing it, it is something that the openAPS community can hack the radio comms for to be able to use it for our scenarios!

fidoman
@efidoman
Jul 27 2017 11:30 UTC
@Cooljoe04_twitter - I think the bionic pancreas has 2 separate tubes going all the way to the single site of the injection. The fork in the road for insulin / glucagon is right at the place where the medicine goes into the skin. This way they can supposedly alternate glucagon vs insulin with only the canula amount as "fluff". I am creating a new channel for dual hormone therapy discussions because I think it will be useful to collect all these great thoughts and ideas.
fidoman
@efidoman
Jul 27 2017 11:35 UTC
https://gitter.im/dual-hormone-delivery/Lobby# - join if your interested in discussions about dual hormone delivery of both insulin and glucagon to more fully mimic the major pancreatic functions of both Beta and Alpha cells.
Paul Dickens
@thebookins
Jul 27 2017 11:35 UTC
Hi All, just troubleshooting some weird loop behaviour and realised that there was an old autotune.json file hanging around that was overriding the pump settings in profile.json. I had turned off autotune in cron but alas neglected to remove this file. Might be worth checking file age to avoid in future. Submitted openaps/oref0#571 to discuss.
Paul Dickens
@thebookins
Jul 27 2017 12:04 UTC
@Kdisimone just wondering if your looptips fix for the nasty Could not get subg_rfspy state or version. Have you got the right port/device and radio_type? error ever made it into oref0 dev? Our loop has been hanging at this error pretty frequently recently, so I just implemented your fix. It it's in dev then it should be fixed when I next update the rig?
rodzilla55
@rodzilla55
Jul 27 2017 13:54 UTC
@LilDucky @fiberfan It would likely be a few months before I can get to it... so if @LilDucky is ready to use it toward this same purpose (dual hormone pump) than please do run with it.
Samuchco
@Samuchco
Jul 27 2017 14:01 UTC
Had my son's (17 years old) first night with OpenAPS !!!!! Amazing!!!! a couple of lows (bG65). Now I need to re-read and start tweeking. A big sense of gratitude to ALL of you guys who participate. Let me know I can be of any help (not technologically related)
Dana Lewis
@danamlewis
Jul 27 2017 14:03 UTC
Congrats @Samuchco! Happy looping!
Paying it forward in either editing and helping improve the docs, and/or helping answering questions here is a great way to pay it forward :smile:
Samuchco
@Samuchco
Jul 27 2017 14:04 UTC
@danamlewis thanks Dana, I think you mean happy sleeping!
Dana Lewis
@danamlewis
Jul 27 2017 14:05 UTC
:D
rodzilla55
@rodzilla55
Jul 27 2017 14:25 UTC
I'm curious as to what some of you have your Pump Battery Warn and Pump Battery Urgent settings are? I have 1.25 and 1.15; But I feel like I am in Warning mode too soon. Thinking of going to 1.21-ish and 1.13-ish.
cameronrenwick
@cameronrenwick
Jul 27 2017 14:51 UTC
what kind of batteries are you using? Lith or alkaline? my urgent pump battery v is set to 1.3. While your pump may continue to function as a pump, going lower that 1.3 (on lith batteries in my case) means 1) that mmtune will often put my signal strength at -99 and 2) that as a lith battery has a flat response and then drops to zero, when my urgent battery alarm hits I replace it almost asap. I'd never sleep with that alarm active. Your pump may vary but...fwiw, my 2 cents
rodzilla55
@rodzilla55
Jul 27 2017 14:56 UTC
I'm using Alkaline... I do change it if getting close to urgent before sleeping. I have just been looping for 19 days now... I am on my 3rd battery and probably have a day left... so I am getting 6 to 7 days out of each battery so far. My battery drops to 1.3 pretty quick (2 days), but seems to take a long time to get to 1.2.
Andy Sharrow
@dramageek
Jul 27 2017 15:25 UTC
@thebookins That was added in, just under a different name. It's scheduled to check once per minute in cron.
https://github.com/openaps/oref0/blob/master/bin/oref0-radio-reboot.sh
fiberfan
@fiberfan
Jul 27 2017 16:05 UTC
@LilDucky PM sent.
jharper11
@jharper11
Jul 27 2017 19:40 UTC

Couple of questions regarding auto tune...
1) How would I link our multiple rigs in order to ensure they are all working with the most current recommendations? 2) Has anyone seen autotune recommendations that are waaaaaay off? We are struggling with post meal highs and autotune is recommending changes that seem to be going the wrong way. Am I autotuning wrong?

running OpenAPS with edison, xdrip and 722 - any help is appreciated!

Dana Lewis
@danamlewis
Jul 27 2017 19:41 UTC
You can copy files from one rig to another (details in second to last paragraph in this section: http://openaps.readthedocs.io/en/latest/docs/Customize-Iterate/autotune.html#phase-b-running-autotune-in-openaps-closed-loop-system)
First re "waaaay" off - I'd be curious if you changed your autosens min and max multipliers? I'd also suggest looking into the logs and looking specifically at how the data is being tuned to understand how it's coming up with the recommendations.
remember, you can always turn autotune off in the rig if needed, do some testing with OpenAPS and meals without it, and turn it back on if you want to isolate autotune-related problems. there may also be other stuff going on, so that'll make it easy to separate the impacts or confirm if it's autotune that's influencing those time periods.
jharper11
@jharper11
Jul 27 2017 19:46 UTC
Awesome Dana, thanks so much! I'll start with the files and work my way through.
Now how did my xdrip turn into an emoji???? Weird gitter trick?
Dana Lewis
@danamlewis
Jul 27 2017 19:47 UTC
yep, weird gitter formatting
we all just ignore it and know what it means :D
LilDucky
@LilDucky
Jul 27 2017 20:50 UTC
@rodzilla55 @fiberfan the pump would be for general R&D. Dual pump is not possible yet due to lack of glucagon here.
Paul Dickens
@thebookins
Jul 27 2017 21:48 UTC
@dramageek I had a look at https://github.com/openaps/oref0/blob/master/bin/oref0-radio-reboot.sh. It looks like that fix is addressing the spidev5.1 already in use error, and won't help with Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?. We're already running oref0-radio-reboot.sh but the rig still hangs when we get the subg_rfspy error.
Am I missing something?
David
@davidkeddydb
Jul 27 2017 21:53 UTC
Is anyone else using a DIA of 2 or am I the only one? With the work be doing with the new DIA curves being done, is the idea you can pick which algorithm works for you or will those curves replace the current in OpenAPS @scottleibrand
Dana Lewis
@danamlewis
Jul 27 2017 21:55 UTC
Curves will be selectable
David
@davidkeddydb
Jul 27 2017 21:55 UTC
Awesome 👏
Dana Lewis
@danamlewis
Jul 27 2017 21:58 UTC
Check out the issue if you want to follow curve development
David
@davidkeddydb
Jul 27 2017 22:01 UTC
@danamlewis trying to keep up since coming back from vacation. I can't get over all the is happening in just 2 weeks. Crazy. Do you guys frequent Bacco? My whole family loves the place
PieterGit
@PieterGit
Jul 27 2017 22:11 UTC
any volunteers for testing out of the box pebble pancreable support? See openaps/oref0#570
Cameron Chunn
@applehat
Jul 27 2017 23:20 UTC
@PieterGit I'm gonna try again tonight.