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

7th
Feb 2019
Jens Heuschkel
@juehv
Feb 07 16:21
I have some news for Enlite users: Medtronic discontinued Enlite v2 and it will be replaced by Guardian Sensor v3. They will produce v2 until Januar 2020 but want to force their customers to use v3 (with the new Pumps and a new transmitter) afterwards. According to the hotline, v3 is not compatible to the Minilink transmitter, hence integrated CGM with compatible Medtronic Pumps will not work any more.
renegadeandy
@renegadeandy
Feb 07 18:29
Hi all. My loop is continually giving me too high a basal when I think it should be giving me a far lower basal to allow my glucose to rise. For example, right now I am 4.8, I've been decreasing tiny amounts over the past 45 minutes from about 5.2,loop is keeping me at 1.1 units per hour, I know I'm going to go low, can feel it, and I've experimented with the loop, this will bring me to about 4.3 when it will start a temp basal of 0.0 but it's too late, I'll tend to go low without an intervention. How can I reduce background basals when my blood is happily in range. Its just a little aggressive in these flat zones I find. Especially overnight
Dana Lewis
@danamlewis
Feb 07 18:31
a) can you grab your logs, or a screenshot of your NS including the OpenAPS pill? b) paste your preferences.json and your autotune
renegadeandy
@renegadeandy
Feb 07 18:34
Absolutely!
Dana Lewis
@danamlewis
Feb 07 18:35
(go ahead and edit out your NS host there)
renegadeandy
@renegadeandy
Feb 07 18:35
I don't think I can do
I can't paste images from the gitter fone app.... And I can't edit messages either. Nightmare
I'll need to do it later
Dana Lewis
@danamlewis
Feb 07 18:36
ok. or plan to change your port, etc.
if it's easier, put the images in dropbox or something, and you can paste the link instead
renegadeandy
@renegadeandy
Feb 07 18:41
https://photos.app.goo.gl/dQjEQTi23whL6M3x9 @danamlewis this should be the 2 images
You cn see from the main nightscout view, anyone can see that this line is only going to end with a low. I have got a lot of these examples. I would want my loop to detect this far earlier and likely half my basal around 5.5 mmol, and at this stage (now 4.6, loop is giving 0.75 units per hour) it should give 0.0
renegadeandy
@renegadeandy
Feb 07 18:49
Or a lower amount than 0.75, because I'm going to go low....
Dana Lewis
@danamlewis
Feb 07 18:51
ah, you're using AAPS. have you run autotune manually? it sounds like you want to adjust your settings to provide different inputs to the calculation. autotune can help you do that. you may also want to try raising your target slightly if you find yourself drifting low, until you sort out your settings.
renegadeandy
@renegadeandy
Feb 07 18:52
Aaps?
Dana Lewis
@danamlewis
Feb 07 18:52
oh, are you just using AndroidAPS (AAPS) for visualizing your rig info? or are you using AAPS directly? either way: it makes predictions on your settings. it's keeping you at neutral temp because your eventual is at your target (4.8). If you want that to do differently; you can adjust your target to be higher; or change your settings, which changes the calculations and thus the predictions and influences the behaviors
renegadeandy
@renegadeandy
Feb 07 18:55
Ahhh, yes it's simply for visualising nightscout, it does a better job than xdrip I've found. I'm using openaps :)
Aaps is making no decisions
Am 4.5 now, openaps rig has dropped to 0.45 per hour, but still, a little too late imo. How do I adjust that eventual target. I think 4.8 as a target is a little low.
4.4 now (just validating my prediction)
renegadeandy
@renegadeandy
Feb 07 19:00
The reason for context in using the Aaps app, is that it shows the openaps pill info. The xDrip+ app doesn't, don't know why
Dana Lewis
@danamlewis
Feb 07 19:00
you can change your baseline target in your pump
that's fine to use it for visualization
check your autotune too - you may be bumping up into the edge of the safety settings & need to adjust further to allow it to tune even further
(cat-autotune)
renegadeandy
@renegadeandy
Feb 07 19:03
Ah yes, bolus wizard has target 4.8-8, so push the 4.8 higher? I didn't know that openaps uses that as a baseline. Either way, if I am below the baseline shouldn't it suspend to allow my bg to rise back to 4.8?

Autotune is ```root@andyaps:~# cat-autotune Parameter | Pump | Autotune | Days Missing

ISF [mg/dL/U] | 43.200 | 47.580 |
Carb Ratio[g/U]| 8.000 | 8.566 |
Basals [U/hr] | - | |
00:00 | 1.100 | 1.151 | 11
00:30 | | |
01:00 | | 1.113 | 13
01:30 | | |
02:00 | | 1.066 | 5
02:30 | | |
03:00 | | 1.042 | 6
03:30 | | |
04:00 | | 1.035 | 5
04:30 | | |
05:00 | 1.200 | 1.077 | 6
05:30 | | |
06:00 | | 1.074 | 8
06:30 | 1.7 | |
07:00 | | 1.179 | 8
07:30 | | |
08:00 | | 1.197 | 8
08:30 | | |
09:00 | | 1.190 | 6
09:30 | | |
10:00 | | 1.216 | 8
10:30 | | |
11:00 | | 1.190 | 10
11:30 | | |
12:00 | 1.600 | 1.220 | 13
12:30 | | |
13:00 | | 1.149 | 11
13:30 | | |
14:00 | 0.900 | 1.080 | 9
14:30 | | |
15:00 | | 1.061 | 9
15:30 | | |
16:00 | 0.950 | 1.058 | 9
16:30 | | |
17:00 | | 1.060 | 9
17:30 | | |
18:00 | | 1.083 | 11
18:30 | | |
19:00 | 1.300 | 1.087 | 12
19:30 | | |
20:00 | | 1.060 | 12
20:30 | | |
21:00 | 1.000 | 0.994 | 12
21:30 | | |
22:00 | 0.900 | 1.014 | 10
22:30 | | |
23:00 | | 1.049 | 10
23:30 | | |```
Dana Lewis
@danamlewis
Feb 07 19:04
if you don't want your target to be 4.8, yes, adjust it. (OpenAPS only uses the bottom number) remember, OpenAPS responds on current and predicted BG. you may want to look through some of these examples, too: https://openaps.readthedocs.io/en/latest/docs/While%20You%20Wait%20For%20Gear/Understand-determine-basal.html (the page overall is a good read, and these visuals may help you interpret your logs: https://openaps.readthedocs.io/en/latest/docs/While%20You%20Wait%20For%20Gear/Understand-determine-basal.html#openaps-algorithm-examples
renegadeandy
@renegadeandy
Feb 07 19:05
Yuck at that formatting, sorry. Phone is bad for this
Dana Lewis
@danamlewis
Feb 07 19:05
yes, your baseline basals during the day are a lot higher; if you adjust those down, autotune will probably tune them down even further, and that may make a big difference in getting your baseline to be more stable without those drifts down.
(if it were me, I'd lower your 1.6 and 1.3 basals to be slightly lower, and see how autotune responds & how things go from there)
renegadeandy
@renegadeandy
Feb 07 19:08
I've upped my baseline from 4.8 to 5.1, actually that Autotune output is wrong
Those are not my existing basals
I've altered them about a week ago since those!
Wow....
Why does it think those are the basals!
Any thoughts?
When reviewed auto tunes recommendations about a week ago, and I adjusted my basals accordingly.... It hasn't picked them up. I set them as Pattern A on my pump, I wonder if autotune doesn't look at the current pattern and instead is hard coded to the base basal
?
I figured I'd use a pattern for the auto tunes suggestions, allowing easy rollback if it caused me problems, but I wonder if that's why it's not working
Raymond Richmond
@PedanticAvenger
Feb 07 19:29
@renegadeandy Those are the basals it is recommending in the autotune output. In operation that is what it tries to achieve if it is possible within the configured safety limits.
Generally most people try to set their pump basals to match autotune suggestions (or at least reasonably close) and over time achieve a steady-state.
renegadeandy
@renegadeandy
Feb 07 19:54
No @PedanticAvenger @danamlewis -- Autotune has incorrectly picked up my current basals. That's not what my basals are. Because I believe autotune doesn't pick up the correct basals if I currently am using a basal 'pattern' which is a feature of medtronic pumps. Can somebody please confirm?
Dana Lewis
@danamlewis
Feb 07 19:56
I think you'll be best positioned to confirm - try switching your pump back to standard, go ahead and do a manual run, and see if it picks up in the autotune report. then switch back to pattern a, run again, and see if it changes.
renegadeandy
@renegadeandy
Feb 07 19:59
well, ill need to switch back to standard, alter all my standard basal rates, then do a manual run
Bit of a pain...
If it has to be that way then sure, ill give it a go
Dana Lewis
@danamlewis
Feb 07 20:00
two choices: 1) do nothing, keep the status quo of the situation. 2) do something. your call :)
renegadeandy
@renegadeandy
Feb 07 20:50
I guess I was just surprised nobody knows if patterns break autotune...
happy to be the one that proves it though
Dana Lewis
@danamlewis
Feb 07 20:52
nobody may have tested that on the latest version. also, are you running master or dev?
renegadeandy
@renegadeandy
Feb 07 21:02
dev
Was it tested on master?
renegadeandy
@renegadeandy
Feb 07 21:14
Just want to know what was designed to work against what is working
Carl
@c-robertson
Feb 07 21:43
Having a problem with my Edison rig, just started today after a powercycle... basically looks like the OS is hosed up somehow. Doesn't respond to SSH login either via hostname or IP, when I try a serial connection over a COM port, it just doesn't respond. Only clue is that in the Win 10 device manager, it shows up as 'MERRIFIELD' which apparently was some sort of Intel codename...
thinking a reflash is in order, wondering if other have experienced a similar issue...
renegadeandy
@renegadeandy
Feb 07 22:02
@danamlewis when I run autotune manually, the --start-date, what should that be?
ive executed without passing a startdate, so :)
renegadeandy
@renegadeandy
Feb 07 22:08
I can confirm, that if you use a Pattern basal set, autotune does not pull that in. It always pulls from the base basal sets from the pump. I think thats important to note to the community @danamlewis
Ive added a note to the docs, and created a pull request @scottleibrand @danamlewis openaps/docs#1416
Dana Lewis
@danamlewis
Feb 07 22:41
thx for confirming & for the PR; merged
renegadeandy
@renegadeandy
Feb 07 22:49
When do new deployments get pushed to the live doc site?
Dana Lewis
@danamlewis
Feb 07 23:00
immediately within a few minutes
renegadeandy
@renegadeandy
Feb 07 23:25
I see it, with a couple of alterations :) np
renegadeandy
@renegadeandy
Feb 07 23:40
Would anyone here please help me test my xDripAPS. I have updated it , and am looking to support it from a fork, as the original is dead as far as I can tell : https://github.com/renegadeandy/xDripAPS