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

4th
Apr 2018
Martin Haeberli
@mhaeberli
Apr 04 2018 02:45
ps - can I place comments in preferences.json? if so, how formatted?
Jeremy Cunningham
@jpcunningh
Apr 04 2018 02:51
@mhaeberli , you can enter records for comments that are like those shown below:
{
        "": "This is a comment",
        "": "This is another comment",
        "max_iob": 0,
        "max_daily_safety_multiplier": 3,
        "current_basal_safety_multiplier": 4,
        "autosens_max": 1.2,
        "autosens_min": 0.7,
        "": "This is yet another comment",
        "rewind_resets_autosens": true,
        "adv_target_adjustments": true,
        "unsuspend_if_no_temp": false,
        "enableSMB_with_bolus": false,
        "enableSMB_with_COB": false,
        "enableSMB_with_temptarget": false,
        "enableUAM": false,
        "curve": "rapid-acting"
}
Martin Haeberli
@mhaeberli
Apr 04 2018 02:56
thx, nice!
Dave Acklam
@dcacklam
Apr 04 2018 07:18
@idea
@ideagit If only there was a way to backfeed basal insulin levels & UAM impact into xdrip+ from NS.... It reads treatments and bolus IOB, but not the impact of variable basals.....
Debby Muskar
@debbymuskar
Apr 04 2018 14:10
Hi all. It's been a long time since I had a problem that I could resolve over in the FB group, but I think I need a little more expertise. I have had problems with both of my rigs not starting the dnsmasq in the reboot, which is causing a lot of looping issues. I have reflashed one of them, which did not help. I ended up with the same issues where the rig could not get my BG readings from my phone and my Bluetooth tether wasn't working either. I ran "systemctl status dnsmasq.service" and got the following:
'root@debbymuskar:~# systemctl status dnsmasq.service
● dnsmasq.service - dnsmasq - A lightweight DHCP and caching DNS server
   Loaded: loaded (/lib/systemd/system/dnsmasq.service; enabled)
  Drop-In: /run/systemd/generator/dnsmasq.service.d
           └─50-dnsmasq-$named.conf, 50-insserv.conf-$named.conf
   Active: failed (Result: exit-code) since Wed 2018-04-04 08:56:03 CDT; 59s ago
  Process: 420 ExecStart=/etc/init.d/dnsmasq systemd-exec (code=exited, status=2)
  Process: 362 ExecStartPre=/usr/sbin/dnsmasq --test (code=exited, status=0/SUCCESS)

Apr 04 08:56:02 debbymuskar dnsmasq[362]: dnsmasq: syntax check OK.
Apr 04 08:56:03 debbymuskar systemd[1]: dnsmasq.service: control process exi...2
Apr 04 08:56:03 debbymuskar dnsmasq[420]: dnsmasq: failed to bind DHCP serve...e
Apr 04 08:56:03 debbymuskar systemd[1]: Failed to start dnsmasq - A lightwei....
Apr 04 08:56:03 debbymuskar systemd[1]: Unit dnsmasq.service entered failed ....
Hint: Some lines were ellipsized, use -l to show in full.
root@debbymuskar:~#`
Alan
@Alzibiff
Apr 04 2018 16:27
I have just received an Edison and Explorer board which I ordered last week in order to build a second rig. All went according to the book until I got to the end of the flash step for the Edison. Rather than finish with it asking for a user name and password, I got a Ctrl+C prompt which would not go away ... so I did a Ctrl+C and the Edison seemed to carry on. However, during the boot up procedure, I get to lines which say failed and I think that is this failure which is stopping me from connecting to my home wifi despite all the wifi settings looking the same as on my original rig. The fail lines are: [FAILED] Failed to start Load Kernel Modules. and [FAILED] Failed to start Hostname Service.
Any suggestions on how to resolve this?
I have tried flashing again but get the same errors
The new Edison was already carrying jubilinux but I was not sure that it was the version compatible with OpenAPS which is why I went through the flashing process.
Alan
@Alzibiff
Apr 04 2018 16:38

One last point - running the next stage - the Bootstrap script - halts because it cannot connect to my wifi: Attempting to bring up wlan0:

ifdown: interface wlan0 not configured
wpa_supplicant: /sbin/wpa_supplicant daemon failed to start
run-parts: /etc/network/if-pre-up.d/wpasupplicant exited with return code 1
Failed to bring up wlan0.

Martin Haeberli
@mhaeberli
Apr 04 2018 17:46
what is relevance of enableSMB_always in light of other enableSMB flags?
in other words, if I use enableSMB_always: a ) does it do anything more than having all other enableSMB flags set true; b ) is there any point in having any of the other enableSMB flags set true?
Scott Leibrand
@scottleibrand
Apr 04 2018 17:47
yes and no
an allowSMB with high target set to false is the only one that overrides enableSMB_always (when you have a high temp target set)
Martin Haeberli
@mhaeberli
Apr 04 2018 17:51
so if one wanted to go ‘full SMB’ one can just enableSMB_always … does that end up with no basals (assuming communications maintained) and just a trail of SMBs?
no basals aka permanent zero temp ...
Scott Leibrand
@scottleibrand
Apr 04 2018 17:52
after meals it often looks like that, but usually it's just a slightly less dense cityscape
Martin Haeberli
@mhaeberli
Apr 04 2018 17:52
k thx
Scott Leibrand
@scottleibrand
Apr 04 2018 17:53
it still sets high temps when you don't have enough bolus IOB, and then lowers them (often to zero) after the SMB goes through
so it also depends whether you're doing no-bolus, or a partial manual meal bolus
Martin Haeberli
@mhaeberli
Apr 04 2018 17:54
current operating mode is - forgotten bolus / UAM; sometimes carb announcement / 0 bolus; sometimes carb announcement / partial bolus ; sometimes carb announcement / full bolus ...
after some excitement with a CGM sensor replacement 3 hours before bedtime (which almost inevitably leads to highs at bedtime / calibration) - high_temptarget_raises_sensitivity seems to have worked well ONCE she was back down from ~300 to maybe 220 or so… I was having trouble figuring out how to encourage it (remotely) to high temp enough to bring it down from 300...
I played with targets - eventually set an 80 temp target for a while; it got a bit more aggressive and even SMB’d, then started ramping down quickly. reset target to 140-140, things settled in from there.
Scott Leibrand
@scottleibrand
Apr 04 2018 17:58
are you doing the inverse (low target lowers sensitivity) as well?
Martin Haeberli
@mhaeberli
Apr 04 2018 17:58
haven’t yet but easily can
Scott Leibrand
@scottleibrand
Apr 04 2018 17:58
with that, an 80 target will set a sensitivityRatio equal to your autosens_max (default of 1.2x)
Martin Haeberli
@mhaeberli
Apr 04 2018 17:59
you know the drill with N=1 / YDMV ; I/we try not to change too many parameters at once.
Scott Leibrand
@scottleibrand
Apr 04 2018 17:59
:+1:
James Babcock
@jimrandomh
Apr 04 2018 18:02
If the CGM is offline or questionable and you want to correct a high, you should look at the IOB in NS, do a fingerstick test, and make a correction bolus with the IOB subtracted off, similar to how you would've done it pre-OpenAPS
Martin Haeberli
@mhaeberli
Apr 04 2018 18:03
good suggestion!
James Babcock
@jimrandomh
Apr 04 2018 18:03
If sensor replacement consistently leads to highs, it may mean the basal settings are too low, and OpenAPS is compensating by high-temping all the time. Check autotune's report and adjust the pump settings to match.
Martin Haeberli
@mhaeberli
Apr 04 2018 18:03
@jimrandomh :+1:
James Babcock
@jimrandomh
Apr 04 2018 18:08
(Never rely on the insulin-on-board calculation performed by your pump, since it doesn't account for the temp basals that OpenAPS uses. If you use the bolus wizard, this will make it use the incorrect IOB, so, do the calculation yourself)
Martin Haeberli
@mhaeberli
Apr 04 2018 18:08
good tips, all…! thx
Paul Heltzel
@pheltzel
Apr 04 2018 18:28
@jimrandomh For a typical bolus, do you skip entering the BG in the bolus wizard on the pump and instead enter BG in NS? Since the IOB on NS would be accurate but the pump's would not?
Paul Heltzel
@pheltzel
Apr 04 2018 18:35
I've been getting some persistent highs after bolusing, running 170-180 and lasting for hours, even when eating very little carbs. My basals looked pretty locked in and autotune appeared to agree. I changed my site but I'm still having trouble. The sensor is in line with my meter.
Dana Lewis
@danamlewis
Apr 04 2018 18:36
what do your logs say at that point in time? are you running in to settings caps?
Paul Heltzel
@pheltzel
Apr 04 2018 18:37
Thinking it could be the insulin. Although a couple times when I'm bolusing, like yesterday morning, the temp basal was at 0 or .3 or I'm not sure why they'd be that low where my basal setting would be more like 1.5 at that time.
Dana Lewis
@danamlewis
Apr 04 2018 18:37
i know how you can tell ;) - check the logs & see what it was saying at that time.
Paul Heltzel
@pheltzel
Apr 04 2018 18:37
Dana Lewis
@danamlewis
Apr 04 2018 18:38
copy and paste pertinent info? we don't have access to your papertrail (but also know, papertrail doesn't always display the full stuff, so grab a NS OpenAPS pill screenshot to start with for fastest analysis)
Paul Heltzel
@pheltzel
Apr 04 2018 18:38
Sorry, sure.
Dana Lewis
@danamlewis
Apr 04 2018 18:39
no worries
part of the learning curve :)
fwiw, here's my take on which types of logs and where are good for what and when. https://diyps.org/2017/05/10/why-guess-when-you-dont-have-to-openaps-logs-why-theyre-handy/ <--- just my opinion though, everyone has personal preferences, but that's a good starting point IMO.
and now I miss my Pebble Steel. nostalgic sigh
Mike
@libxmike
Apr 04 2018 18:43
@danamlewis what's wrong with your Pebble Steel? I love my Time Steel!
Dana Lewis
@danamlewis
Apr 04 2018 18:43
I switched to the Pebble 2 HR to get heart rate. But I adored my Steel.
Mike
@libxmike
Apr 04 2018 18:44
how is the HR ? was considering to get one but don't like the plastic
Paul Heltzel
@pheltzel
Apr 04 2018 18:44
image.png
Scott Leibrand
@scottleibrand
Apr 04 2018 19:04
@libxmike when I switched from the Steel to the HR, I instantly stopped having watchband issues because it was so much lighter. and the HR is cool for running etc.
@pheltzel what does autotune say about your CR?
Paul Heltzel
@pheltzel
Apr 04 2018 19:06
@scottleibrand It was v. close 4.000 | 3.957
Scott Leibrand
@scottleibrand
Apr 04 2018 19:24
This message was deleted
Paul Heltzel
@pheltzel
Apr 04 2018 19:26
image.png
Scott Leibrand
@scottleibrand
Apr 04 2018 19:26
it also looks like you were riding fairly flat after those meals, but started high and weren't coming down. what does it say about basals?
k, looks like it's suggesting your mid-day basals are too low
how many days has autotune been running?
Paul Heltzel
@pheltzel
Apr 04 2018 19:27
Just a few--this is early on for sure
Scott Leibrand
@scottleibrand
Apr 04 2018 19:29
ok. probably best to change one thing at a time, but I suspect you'll end up needing higher afternoon and evening basals.
autotune will keep adjusting them gradually, until/unless it gets to 1.2x (autosens_max) times your pump values
Paul Heltzel
@pheltzel
Apr 04 2018 19:35
Cool, ok thanks Scott.
I think I'm realizing the morning my basal was at 0 when I went to bolus was because the log shows my temp basal was 3U/hr shortly before that so there was a bit of IOB that I hadn't expected.
I'll up the basal and see how that goes, appreciate it.
Paul Heltzel
@pheltzel
Apr 04 2018 19:42
Just a question--not looking to make a change at the moment, while I make this basal change: If I set the target on the pump to 90 rather than say 110 will that affect anything with the basals (i.e. a more aggressive target on the pump making the AP less aggressive on the basal end?)
And is it ok to enter the BG in the bolus wizard in the pump (rather than entering BG in NS)?
Scott Leibrand
@scottleibrand
Apr 04 2018 19:45
if you enter BG in the pump bolus wizard, and you don't have a high enough high-end correction target, it will try to do a correction bolus, over and above the high temps that OpenAPS is already doing.
setting a low temp target on the rig is a good way to get it to be more aggressive without committing to a full correction bolus (in case you start dropping immediately afterward). by default low targets don't make OpenAPS very much more aggressive, but there is a preference you can set that causes it to set the sensitivityRatio based on the temp target, which makes it respond much more strongly.
but you definitely won't get anything like "a more aggressive target on the pump making the AP less aggressive on the basal end" - a low target will always make the rig more aggressive, and a high target will make it less so.
Paul Heltzel
@pheltzel
Apr 04 2018 20:12
Great, thanks. I've got the high end set to 250, per the docs, to avoid double dipping.
James Babcock
@jimrandomh
Apr 04 2018 20:27
@pheltzel At the end of the bolus wizard before you confirm, it gives a summary saying how much it'll give, how much is for carbs, how much is for correction, and how much it adjusted for its notion of IOB (if at all). What does it say on that screen if you enter a meal bolus the way you normally would?
I don't normally enter BGs into my pump at all, I only enter them into the CGM receiver
Paul Heltzel
@pheltzel
Apr 04 2018 20:48
I see what you mean, it shows 0 for correction since the high end of the target is set to 250 on the pump. So it's a non issue.
Alan
@Alzibiff
Apr 04 2018 22:02
I managed to get my second rig build back on track (see above) after a couple of hours googling and typing in Linux commands whick I didn't fully understand. Fix was the old, shut down and reboot! Wish I had done it sooner but the main thing is I can continue with the rest of the build tomorrow as bed is calling now :-)
fidoman
@efidoman
Apr 04 2018 23:57
Is there a way to confirm the curve that is used by OpenAPS by looking in the pump-loop.log file? I changed to fiasp/novolog (50/50) mixture today with good results so far. I changed the preference for curve to "ultra-rapid" but wondering if there is a way to see that in the log?
Scott Leibrand
@scottleibrand
Apr 04 2018 23:59
I don't think we currently log that. you could make oref0-calculate-iob log something, but that gets run many many times by determine-basal, so probably wouldn't want it going to pump-loop.log