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

16th
Oct 2017
jaylagorio
@jaylagorio
Oct 16 2017 00:00 UTC
Oh, even better. SSH login is pretty secure even over a plain-text transport.
Scott Leibrand
@scottleibrand
Oct 16 2017 00:04 UTC
If anyone wants to volunteer to work on getting a local web server set up we could use that too. @tim2000s has done some prototyping....
Paul Dickens
@thebookins
Oct 16 2017 00:39 UTC
I'd be keen to be involved with that - needed for local calibrations and session start/stop events with xdrip-js too
tynbendad
@tynbendad
Oct 16 2017 00:50 UTC
@scottleibrand :+1:
Glenn Primack
@primags
Oct 16 2017 01:34 UTC
I'm trying to troubleshoot why my rig stopped talking to my pump. When I use mmtune I can see the two lights flash on my explorer board. I get the following output. Can someone tell me if this error is something to worry about? I saw a fix for this was to reboot, but I have rebooted several times.
image.png
Scott Leibrand
@scottleibrand
Oct 16 2017 02:07 UTC
@bdr1177 what's the current state of the rig?
cat monitor/temp_basal.json
cat monitor/pumphistory.json | jq .[0]
cat enact/suggested.json | jq .reason
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:09 UTC
Give me just a minute to get it up and running again. I had just switched over to my other rig.
Scott Leibrand
@scottleibrand
Oct 16 2017 02:09 UTC
k
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:15 UTC
@scottleibrand Okay. Just got it running again. I had to reinstall dev branch because of my failed attempt to downgrade to master. Do I need to let it loop a few times?
Scott Leibrand
@scottleibrand
Oct 16 2017 02:21 UTC
I'd run those commands as soon as you get the error
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:21 UTC
Ok. Waiting for it
Scott Leibrand
@scottleibrand
Oct 16 2017 02:21 UTC
we can always run them again after another run
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:30 UTC
root@natopenaps:~/myopenaps/monitor# cat temp_basal.json
{
  "duration": 30,
  "rate": 0.0,
  "temp": "absolute"
}root@natopenaps:~/myopenaps/monitor#
 {
    "_type": "TempBasalDuration",
    "duration (min)": 30,
    "_description": "TempBasalDuration 2017-10-15T21:01:52 head[2], body[0] op[0x16]",
    "timestamp": "2017-10-15T21:01:52",
    "_body": "",
    "_head": "1601",
    "_date": "b481154f11"
  },
  {
    "_type": "TempBasal",
    "temp": "absolute",
    "_description": "TempBasal 2017-10-15T21:01:52 head[2], body[1] op[0x33]",
    "timestamp": "2017-10-15T21:01:52",
    "_body": "00",
    "_head": "330c",
    "rate": 0.3,
    "_date": "b481154f11"
  },
  {
    "_type": "BGReceived",
    "link": "c14953",
    "_description": "BGReceived 2017-10-15T20:33:03 head[2], body[3] op[0x3f]",
    "timestamp": "2017-10-15T20:33:03",
    "_body": "c14953",
    "_head": "3f17",
    "amount": 186,
    "_date": "83a1546f11"
  },
  {
    "_type": "CalBGForPH",
    "_description": "CalBGForPH 2017-10-15T20:33:03 head[2], body[0] op[0x0a]",
    "timestamp": "2017-10-15T20:33:03",
    "_body": "",
    "_head": "0aba",
    "amount": 186,
    "_date": "83a1346f11"
  }
Scott Leibrand
@scottleibrand
Oct 16 2017 02:33 UTC
that's not the output of my command. ;-p
(but more useful)
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:33 UTC
root@natopenaps:~/myopenaps/enact# cat suggested.json | jq .reason
"Warning: currenttemp rate 0.3 != lastTemp rate 0 from pumphistory; setting neutral temp of 0.3. 28m left and 0.3 ~ req 0.3U/hr: no temp required"
Scott Leibrand
@scottleibrand
Oct 16 2017 02:33 UTC
what's the current time there?
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:33 UTC
10:33 pm
Scott Leibrand
@scottleibrand
Oct 16 2017 02:34 UTC
so the last record in your pumphistory is from 9:01pm
have there been other temps set in that time?
I assume so: it says there's one running
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:35 UTC
Yes, several temps.
Scott Leibrand
@scottleibrand
Oct 16 2017 02:35 UTC
so the safety check is working as intended and preventing oref0 from potentially overdosing insulin
now the question is why your pumphistory is not updating
try service cron stop; killall -g oref0-pump-loop; cd ~/myopenaps; openaps monitor-pump; service cron start and then check pumphistory again
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:38 UTC
root@natopenaps:~/myopenaps/monitor# cat pumphistory.json | jq .[0]
{
  "_type": "TempBasalDuration",
  "duration (min)": 30,
  "_description": "TempBasalDuration 2017-10-15T22:35:39 head[2], body[0] op[0x16]",
  "timestamp": "2017-10-15T22:35:39",
  "_body": "",
  "_head": "1601",
  "_date": "a7a3164f11"
}
Scott Leibrand
@scottleibrand
Oct 16 2017 02:40 UTC
ok, good, it's updating now
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:40 UTC
But just gave the warning again.
Scott Leibrand
@scottleibrand
Oct 16 2017 02:40 UTC
ok, run all three commands again?
and maybe check monitor/pumphistory-zoned.json and monitor/pumphistory-merged.json too
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:42 UTC
root@natopenaps:~/myopenaps/monitor# cat ~/myopenaps/monitor/temp_basal.json
{
  "duration": 29,
  "rate": 0.3,
  "temp": "absolute"
root@natopenaps:~/myopenaps/monitor# cat pumphistory.json | jq .[0]
{
  "_type": "TempBasalDuration",
  "duration (min)": 30,
  "_description": "TempBasalDuration 2017-10-15T22:35:39 head[2], body[0] op[0x16]",
  "timestamp": "2017-10-15T22:35:39",
  "_body": "",
  "_head": "1601",
  "_date": "a7a3164f11"
}
Scott Leibrand
@scottleibrand
Oct 16 2017 02:45 UTC
ls -la monitor/pumphistory*
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:45 UTC
root@natopenaps:~/myopenaps/enact# cat suggested.json | jq .reason
"Warning: currenttemp rate 0.3 != lastTemp rate 0 from pumphistory; setting neutral temp of 0.3. 22m left and 0.3 ~ req 0.3U/hr: no temp required"
root@natopenaps:~/myopenaps/monitor# ls -la ~/myopenaps/monitor/pumphistory*
-rw-r--r-- 1 root root 7506 Oct 15 22:44 /root/myopenaps/monitor/pumphistory.json
-rw-r--r-- 1 root root 93051 Oct 15 22:44 /root/myopenaps/monitor/pumphistory-merged.json
-rw-r--r-- 1 root root 7674 Oct 15 22:44 /root/myopenaps/monitor/pumphistory-zoned.json
Scott Leibrand
@scottleibrand
Oct 16 2017 02:47 UTC
hmm, I wonder if it might be easiest if we do a screen share so I can poke at it directly
Brian Rabinovitz
@bdr1177
Oct 16 2017 02:47 UTC
sure, that's fine
Scott Leibrand
@scottleibrand
Oct 16 2017 02:50 UTC
ok, PMing.
Andrew Warrington
@andrew-warrington
Oct 16 2017 05:50 UTC
@scottleibrand thank you. Here it is with the extra info. I hadn't even noticed that IOB and COB were 0 in the screenshot...
Capture2.JPG
I had an error in Papertrail configuration so I didn't capture this there. (That's fixed now, but too late to catch this moment.) So, here's a link to an excerpt of the log uploaded to Google Drive: https://drive.google.com/open?id=0ByL5vTXq7-zBNjRHdTJ6LWpENGs
I included 4 loops in there. The 2nd and 4th loops include the bolus decisions.
Didn't post here directly because I thought it would polute.
br
@brmccollum
Oct 16 2017 05:54 UTC

Hi @scottleibrand I have been semi successfully looping on oref0 for the past two weeks. My rig rarely ever makes it 24 hours without having issues. About two days ago it quick controlling my pump. However my pump still sends data (bolus', primes, etc) to my nightscout.

I found these errors in the log and I can't recall if have seen them before while I have been looping:
Refresh at node.js:906:3 ed pumphistoryError: carb_ratio undefined out of bounds and meal.json
&
Old settings refreshISF of null is not supported

My loop also shows me the settings it wants to send to the pump. IE {"insulinReq":0,"bg":218,"reservoir":null,"temp":"absolute","snoozeBG":105,"rate":0,"predBGs":{"IOB":[218,218,217,216,214,212,209,206,202,198,194,189,184,179,175,170,166,161,157,152,148,143,139,135,132,128,125,122,119,116,113,111,109,107,105,104,103,101,101,100,99]},"minPredBG":999,"IOB":2.611,"reason":"COB: 0, Dev: 25, BGI: -4.13, ISF: 53, Target: 118, minPredBG 98, IOBpredBG 99; Eventual BG 105 < 118, setting 60m zero temp. ","COB":0,"eventualBG":105,"recieved":true,"duration":60,"tick":"+0","timestamp":"2017-10-15T00:03:23.957484","deliverAt":"2017-10-15T07:03:19.112Z"}

Any suggestions. I have also been wanting to try oref1 but wanted to get my rig solid first.

Andrew Warrington
@andrew-warrington
Oct 16 2017 05:58 UTC
I agree avoiding those bolus might not have prevented the hypo. In my old APS I simply had a rule (based on personal experience): Do not bolus when 20-minute glucose trend is down, even in the presence of COB, even with negative IOB. The idea was, if glucose is going down under those conditions, then there is something at play the system is not aware of, so the safest course of action is not to bolus. So I thought I would put it out there.
Marcus Lindfors
@Mackwe
Oct 16 2017 05:59 UTC
Hi! Moved from 0.5.3 (I think) master to 0.6.0 dev in order to use the new insulin action curves. After that I got some relly odd behaviour.. I got a friend who's got the same thing. Not sure what versions she uses though. Problem; After bolusing, within about 10 mins, the rig starts zerotemping. Every time. By the time the BG slowly falls in the right direction it has been zerotemping for quite a bit of time, which in turn gives that any kind of carb intake will skyrocket the BG. Autotune tells me my numbers are correct. I even ran with 1.5/0.5 to allow it some space. Same result. Where do I start fiddling? Gut instinct tells me the new way of thinking is SMB (giving 75% of the recommended bolus and let the rig handle it), but we don't use SMB on our kids. Or do I change the carb ratio? Or is the ISF skewed due to the new curve..? Can someone please explain this?
Scott Leibrand
@scottleibrand
Oct 16 2017 06:05 UTC
@Mackwe are you entering carbs?
Marcus Lindfors
@Mackwe
Oct 16 2017 06:06 UTC
Yes. Through pump.
Scott Leibrand
@scottleibrand
Oct 16 2017 06:06 UTC
at the time of the bolus? or is this a prebolus, and then you enter carbs later?
Marcus Lindfors
@Mackwe
Oct 16 2017 06:07 UTC
@scottleibrand I enter the carbs into th pump and bolus as per suggestion. Normally about 5-10 mins before meal.
Scott Leibrand
@scottleibrand
Oct 16 2017 06:07 UTC
and what does the reason field say for zero temping?
Marcus Lindfors
@Mackwe
Oct 16 2017 06:07 UTC
@scottleibrand So the insulin and carb info enters the system at the same time.
Scott Leibrand
@scottleibrand
Oct 16 2017 06:09 UTC
@brmccollum something is borked with what OpenAPS thinks your pump settings are. I would try drilling down to reproduce the carb_ratio undefined out of bounds and ISF of null is not supported errors and see why you're getting those. (search for "drill down" in the docs)
Marcus Lindfors
@Mackwe
Oct 16 2017 06:09 UTC
BG 106, Temp Basal started 0.00 for 30 m, COB: 14, Dev: 33, BGI: -2.03, ISF: 22.6, Target 5.0, minPredBG -0.9, minGuardBG -5.3, IOBPredBG 2.2, COBPredBG 3.4; minGuardBG -5.3<3.6
The carb ratio in settings/carb_ratio.json matches the pump.
Scott Leibrand
@scottleibrand
Oct 16 2017 06:10 UTC
ok, so minGuardBG -5.3<3.6 means that oref0 thinks BG will drop too low before the carbs kick in
Marcus Lindfors
@Mackwe
Oct 16 2017 06:10 UTC
It doesn't. After each of these boluses the BG rises. By quite a bit.
Not seldom we go from about 90 to 180 before it turns. It zerotemps throughout the rise..
Scott Leibrand
@scottleibrand
Oct 16 2017 06:11 UTC
that's probably due to oref0 making conservative assumptions about the carbs hitting over 3h at first, until it sees actual carb absorption to tell it they're going to hit faster
it should stop zero temping once it's rising fast enough: can you get another reason field from later on there?
or PM me your NS URL and I can look
Marcus Lindfors
@Mackwe
Oct 16 2017 06:12 UTC
Can this behavior be changed? I have never touched the 5m_carb thing..
Scott Leibrand
@scottleibrand
Oct 16 2017 06:13 UTC
yeah, I need to figure out what it's doing first to see whether the algorithm is a poor fit or if there are any settings that need adjusted
Marcus Lindfors
@Mackwe
Oct 16 2017 06:15 UTC
@scottleibrand I am sure it's a settings issue. I just need to understand where..
Scott Leibrand
@scottleibrand
Oct 16 2017 06:16 UTC
it looks like minGuardBG is getting set based on something other than the COB predBGs. lemme look at the code there
Marcus Lindfors
@Mackwe
Oct 16 2017 06:17 UTC
Please! Thank you!
Scott Leibrand
@scottleibrand
Oct 16 2017 06:17 UTC
yeah, I see a logic fail for the case of UAM not enabled
Marcus Lindfors
@Mackwe
Oct 16 2017 06:18 UTC
Yay! Sounds fixable!
Scott Leibrand
@scottleibrand
Oct 16 2017 06:22 UTC
openaps/oref0#721
Marcus Lindfors
@Mackwe
Oct 16 2017 06:23 UTC
@scottleibrand Fantastic! Will pull & rebuild this pm. I'll report back as soon as I know more. Thanks!
Scott Leibrand
@scottleibrand
Oct 16 2017 06:24 UTC
:+1:
thanks for clearly identifying the problem and making it easy to find and fix the bug
Scott Leibrand
@scottleibrand
Oct 16 2017 06:29 UTC
@andrew-warrington COB was 18g and 14g respectively for those two SMBs, and COBpredBG 152 and 137 respectively
oh, and target was 85
Adjusting targets for high BG: min_bg from 103 to 85; target_bg from 103 to 85; max_bg from 103 to 85
you might want to disable advanced target adjustments
there are some other features in the PR pipeline that will do a better job of safely high-temping when BG is high, so we might end up defaulting that off before too long anyway
Scott Leibrand
@scottleibrand
Oct 16 2017 06:36 UTC
"adv_target_adjustments": false in preferences.json
br
@brmccollum
Oct 16 2017 07:09 UTC

@scottleibrand I tried to drill down. Didn't get much luck. (<---poor linux skills may be to blame) I did see this a lot in my logs: AlreadyInUseException("%s already in use" % self.device) mmeowlink.vendors.serial_interface.AlreadyInUseException: /dev/spidev5.1 already in use

After restarted Cron in cycled through this for a bit:

```Listening for 30s silence before mmtuning: .No pump comms detected from other rigs
mmtune: "916.612", 5, -57
Broadcast message from root@dlk (Sun 2017-10-15 23:59:02 PDT):

Rebooting to fix radio errors!
The system is going down for reboot at Mon 2017-10-16 00:01:02 PDT!

Starting pump-loop at Sun Oct 15 23:59:03 PDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
Radio ok. Listening: .No pump comms detected from other rigs
Old pumphistory: RefreshNo more radio errors; canceling reboot

Broadcast message from root@dlk (Mon 2017-10-16 00:00:01 PDT):

The system shutdown has been cancelled at Mon 2017-10-16 00:01:01 PDT!

ValueError: need more than 0 values to unpack

Error, retrying```
Then another bit later it looped for the first time in day and set a temp basal on my 715 pump. However I really have no idea why?

Velibor Maric
@vebaba
Oct 16 2017 07:33 UTC
@scottleibrand Now, when I was ready to drill down an error, autotune works again (after midnight's scheduled run). I think the error got scared of me ;)
Andrew Warrington
@andrew-warrington
Oct 16 2017 08:03 UTC
@scottleibrand thank you. I have disabled the target adjustments for now. Jumping into the PRs for a read, too!
Andrew Warrington
@andrew-warrington
Oct 16 2017 08:12 UTC
Oops, just found something else which may have contributed in some form. "curve" in preferences.json was set to bilinear, and I am on Fiasp. Have changed that now to ultra-rapid (am running 0.6.0-dev). Will observe carefully.
Velibor Maric
@vebaba
Oct 16 2017 11:07 UTC
Yesterday, I mistakenly entered CH twice (I started bolus as square and immediately canceled it). I hoped that canceled bolus will cancel CH but it didn't. I didn't know how to alter data so I eat cake (poor me). Now, I've read this topic https://github.com/nightscout/cgm-remote-monitor/issues/637 and learned how to delete data through Reports tab on NS. Of course, I realized that I should enter new Treatment w/o CH next time but all kind of mistakes can arise which needs editing.
Question: Is this safe way to delete record? Will Rig take it in account?
Dana Lewis
@danamlewis
Oct 16 2017 12:03 UTC
@vebaba you can't delete data from the pump
Next history read will re-upload it
So I would set a high temp target for safety
Velibor Maric
@vebaba
Oct 16 2017 12:05 UTC
@danamlewis I understand. Thankfully, IKEA almond cake is delicious for these kind of mistakes ;)
Dana Lewis
@danamlewis
Oct 16 2017 12:24 UTC
@vebaba :cake: :smile:
Be careful and know if you adjust pump treatment timing it creates a double entry in NS aka double carbs. So just keep an eye out for that and if you spot double carbs, that's why and you can delete the second carb entry. But not the first since it's on pump.
Velibor Maric
@vebaba
Oct 16 2017 12:30 UTC
@danamlewis I agree. In time when I learn more, I'll make PR to ask to make OAPS disregard CH from canceled boluses on the pump. Maybe people would prefer that option...
Dana Lewis
@danamlewis
Oct 16 2017 12:33 UTC
If you can figure out a way to do that - definitely there are folks who would like that!
jcwarrior
@jcwarrior
Oct 16 2017 14:04 UTC
Hi! I'd like to get a backup rig (explorer board + edison). Enhanced radio devices seem doesn't keep selling all packacke with jubilinux pre-installed. Whats the easiest way to get a complete rig now (explorer+edison+jubilinux)?
Jacob H
@jdhigh
Oct 16 2017 14:53 UTC
@jcwarrior If you need it quickly, just buy it without the jubilinux preinstall and flash it yourself. It really only takes like 20 minutes, and there's step by step instructions on the web page. It's NOT as bad as they say...
jaylagorio
@jaylagorio
Oct 16 2017 14:54 UTC
Agreed, flashing an Edison is probably on the more technical side of things, but honestly isn't that bad
jcwarrior
@jcwarrior
Oct 16 2017 14:56 UTC
it's not urgent, but I don't know if they have stopped selling it definitively. I'm technical guy, not afraid for flash, but it's cool when they have already done it :-)
thanks @jdhigh @jaylagorio
Scott Leibrand
@scottleibrand
Oct 16 2017 15:22 UTC
AFAIK they're still selling them, just have to flash more to get them back in stock. But if it's for a backup rig, buying from a third party to flash yourself is a good idea. Among other things it preserves Edison stock for future new loopers while they work on ramping up production of the Explorer HAT for the Pi.
sdneufer
@sdneufer
Oct 16 2017 16:01 UTC
@scottleibrand What is the Explorer HAT for Pi?
Dana Lewis
@danamlewis
Oct 16 2017 16:10 UTC
@sdneufer it's an explorer for a Pi based rig instead of edison. More details soon.
Dana Lewis
@danamlewis
Oct 16 2017 16:17 UTC
Whoop whoop! New offline mode worked splendiferously! @jaylagorio , you're going to love. Ssh in to rig in the overhead, no cables needed :smile:
sdneufer
@sdneufer
Oct 16 2017 16:23 UTC
@danamlewis
john36m
@john36m
Oct 16 2017 16:32 UTC
What's the command to find out what version of openaps I have on my rig?
cameronrenwick
@cameronrenwick
Oct 16 2017 16:32 UTC
ok @danamlewis ... I need more info!!
@john36m npm list -g oref0
john36m
@john36m
Oct 16 2017 16:34 UTC
I've got 0.5.2
Scott Leibrand
@scottleibrand
Oct 16 2017 16:34 UTC
@cameronrenwick the offline branch of oref0. PR is openaps/oref0#716
cameronrenwick
@cameronrenwick
Oct 16 2017 16:34 UTC
prefect...thx.. goin there next!
jaylagorio
@jaylagorio
Oct 16 2017 16:34 UTC
@danamlewis That sounds super awesome!!
sdneufer
@sdneufer
Oct 16 2017 16:35 UTC
@danamlewis So purchase 2 more explorer/edisons now or wait? Size?
Dana Lewis
@danamlewis
Oct 16 2017 16:35 UTC
@john36m definitely want to grab 0.5.5 for Autotune improvements
Scott Leibrand
@scottleibrand
Oct 16 2017 16:35 UTC
@sdneufer size is slightly larger, based on the larger footprint of the Pi 0
Dana Lewis
@danamlewis
Oct 16 2017 16:35 UTC
@sdneufer if you have working rigs, I'd wait.
Scott Leibrand
@scottleibrand
Oct 16 2017 16:36 UTC
the Explorer HAT also has a small screen, which will be very useful for some things
I suspect most people will have some use cases they want an Edison for (smaller, longer battery life) and some that they want to use the new Explorer HAT rigs for (screen). so if you already have an existing Edison rig and don't need a new one urgently, waiting for the HAT wouldn't be a bad idea
cameronrenwick
@cameronrenwick
Oct 16 2017 16:38 UTC
is this a 3/6/12 month wait Scott or...
sdneufer
@sdneufer
Oct 16 2017 16:40 UTC
Interesting. Pi zero w based? Got one a bit ago. My best plans to try things get interrupted by .. everything else :(
Scott Leibrand
@scottleibrand
Oct 16 2017 16:45 UTC
@cameronrenwick hopefully by end of year
Velibor Maric
@vebaba
Oct 16 2017 16:45 UTC
@danamlewis Regard updating. I'm on 0.5.4. Is cd ~/src/oref0 && git checkout master && git pull && sudo npm install -g oref0 all I need to enter into the Rig or there are some previous steps? (stop, backup etc.)
Scott Leibrand
@scottleibrand
Oct 16 2017 16:45 UTC
@sdneufer the HAT is compatible with all Pi's, so for mobile use you'd want a Pi0W yes
@vebaba that should update your software. if you are upgrading from 4.x you'll want to re-run oref0-setup too
Jacob H
@jdhigh
Oct 16 2017 16:46 UTC
^ That's nice to know - I just got a sweet deal on a Pi3.
Scott Leibrand
@scottleibrand
Oct 16 2017 16:46 UTC
at this moment the Pi0W is actually too slow to run the full oref0-pump-loop with SMB: we need to do some optimization. works great on the Pi3 though
john36m
@john36m
Oct 16 2017 17:03 UTC
I just upgraded from master 0.5.2 to 0.5.5 I noticed that my preferences.json file is quite a bit shorter. I could post all the differences if that's needed. Running autosens and autotune but not SMB. Is this normal or did I mess up something in the setup?
cameronrenwick
@cameronrenwick
Oct 16 2017 17:12 UTC
normal AFAIK @john36m .. after update you'll need to update or personalize pref
sdneufer
@sdneufer
Oct 16 2017 17:12 UTC
I wondered if the pi0w would be enough to run it. My guess was no, however the pi3 is big!
john36m
@john36m
Oct 16 2017 17:13 UTC
@cameronrenwick Thanks. I did my personal tweeking.
cameronrenwick
@cameronrenwick
Oct 16 2017 17:14 UTC
I think really that's a safety aspect (and important) so that your prefs get added by you .. that's why most options start as "false" sort of thing...(if I understand this all correctly)
br
@brmccollum
Oct 16 2017 17:19 UTC
@scottleibrand my rig worked for about 9 hours, not it is stuck on this error: Old pumphistory: IOError: Lock at '/root/myopenaps/.git/refs/heads/master.lock' could not be obtained which I have never seen before.
Thacher
@thacherT1D
Oct 16 2017 17:55 UTC
Is there a way for my target range to be 100-120, but only have alarms if I'm under 70 or over 160?
is that all done in Night Scout?
Dana Lewis
@danamlewis
Oct 16 2017 17:56 UTC
yes
cameronrenwick
@cameronrenwick
Oct 16 2017 18:49 UTC
dunno if others would find value in this or not but one of the SMB things I'd like to see is in displaying it in NS that rather than having "0.1U" I'd be very happy w the little blue circle and simply ".1" When SMB is fully ramping up we tend to see this big "ghost" of numbers and letters, making it difficult to see the values of SMBs. My endo was looking at this with me the other day and he asked how I knew how much each SMB was. Anyway, dunno if this is something that people would like or is even doable but FWIW, my 2 cents
this approach may help to see .1 or .2 or .xx
Samuchco
@Samuchco
Oct 16 2017 18:51 UTC
@cameronrenwick in case you need, you can also see clearly the SMBs on the Reports under the Tratments tab
Scott Leibrand
@scottleibrand
Oct 16 2017 19:01 UTC
@cameronrenwick in nightscout/cgm-remote-monitor#2919 I made the font smaller on the treatment labels, which helps somewhat...
I like your idea about removing the label for 0.1U SMBs: question will be how hard that is to implement
or maybe that's not your idea: did you mean that, or just removing the U ?
looks like both are easy
Scott Leibrand
@scottleibrand
Oct 16 2017 19:08 UTC
Screenshot 2017-10-16 12.08.14.png
Scott Leibrand
@scottleibrand
Oct 16 2017 19:15 UTC
Screenshot 2017-10-16 12.14.45.png
first screenshot is just the U removed. second is removing label entirely for 0.1U, and removing the U from 0.2 to 0.5 and having the full label (w/o a space in between) for >0.5U
jaylagorio
@jaylagorio
Oct 16 2017 19:16 UTC
Consistency - I like having the U or not, not having it sometimes.
Samuchco
@Samuchco
Oct 16 2017 19:17 UTC
I'm changing pump from a 522 to a 722. After configuring th settings on the new 722, do I just change the serial number on the rig, or more than that.
jaylagorio
@jaylagorio
Oct 16 2017 19:22 UTC
Should just be changing the serial number in your setup command an rerunning
cameronrenwick
@cameronrenwick
Oct 16 2017 19:31 UTC
@scottleibrand I honestly think we all know that we're talking "unit" so for me the U is redundant. Equally, if there didn't need to be the "0" ahead of the ".x" I'd be fine with that too. Assuming this suggestion doesn't mess up any non-SMB bolus (and values like X.X) I think just seeing ".3" etc would clean up the display
jaylagorio
@jaylagorio
Oct 16 2017 19:31 UTC
^^
cameronrenwick
@cameronrenwick
Oct 16 2017 19:32 UTC
on your 1st screen shot some of the zeros overlap the .2's and that's what I'm hoping to avoid
sdneufer
@sdneufer
Oct 16 2017 20:03 UTC
image.png
No 0 and no U is my vote!
tynbendad
@tynbendad
Oct 16 2017 20:23 UTC
I've never really looked at raw values from dexcom much until recently. I see this in the docs "If you are a G4 user, this will also bring RAW BG data into the rig during sensor restarts or ??? times. The rig will loop using RAW BGs so long as the BG value is under 150 mg/dl. ". Does that mean people found the g4 raw values reliable enough under 150? And should that extend to g5? What happens with raw above 150?
Jake Punshon
@JakePunshon
Oct 16 2017 21:44 UTC
I'll put in my 2 cents and vote that the leading 0 and U can be dropped. I also prefer the smaller font.
Samuchco
@Samuchco
Oct 16 2017 21:46 UTC
I have my son in 0.6.0 autotune. I have a question. In his settings, Max_IOB is 6. I have noticed that the SMBs begin to bolus only after the IOB drops bellow 6. Is that correct?
Dana Lewis
@danamlewis
Oct 16 2017 21:52 UTC
@tynbendad yes re looping on raw. Right now if above 150 it does not loop. Debating a change to that as we speak. Not decided. But FYI G5 afaik doesn't have raw
Samuchco
@Samuchco
Oct 16 2017 22:02 UTC
Screenshot (7).png
ZandMarc
@ZandMarc
Oct 16 2017 22:02 UTC
afbeelding.png
since yesterday my rig is giving me all sorts of headaches... Anybody know what this could mean?
Sorry, that picture wasn't very clear. What I meant is this in the logs: Oct 16 22:15:13 Marcels1stRig autosens-loop.log: ++++++23h++++++u(xxxxxx0hxxxxx30g(xxxxxxx1hxxxxxxxxxxxx2hxxxxxxxxxxxx3hxx)++++++++++4h++++++++++u(xx5hxxxxxxxxxxxx6hxxxxxxxxxxxx7hxxxxxxxxxxxx8hxxxxxxxxxxx9hxxxxxx35g(xxxxxx10hxxxxxxxxxxxxx)u(x)--------13h------+30g(xxxxx14hxxxxxxxxxxxx15hxxxxxxxxxxxx16hxxx)--+++---17h-----++--u(x60g(xxxxxxxxxx19hxxxxxxxxxxxx20hxxxxxxxxxxxx21hxxxx)u(xxx40g(xxx22hxx
jaylagorio
@jaylagorio
Oct 16 2017 22:06 UTC
@Samuchco Yes, that's right. OpenAPS won't add to IOB via SMB or high temps when you're above the MaxIOB.
fidoman
@efidoman
Oct 16 2017 22:14 UTC
@sdneufer :+1: on the No leading 0 and no U
Samuchco
@Samuchco
Oct 16 2017 22:19 UTC
Thanks @jaylagorio that means that if he is eating say, 60 gm carbs and his bolus is 10 units (at 6 to 1) then even if you temp Target to eating soon right before or after he boluses, the temp Target will only affect the temp basals but not the SMBs. SMBs will kick in only after he has consumed more than 6 units of insulin.
Does it make sense to raise the MAX_IOB to around 12-15? so that in case he starts going high he can get SMBs? ofcourse he would have to under-bolus so as not to crash
Samuchco
@Samuchco
Oct 16 2017 22:25 UTC
Screenshot (7).png
a different issue. Gave him a juice and logged the carbs (15gm) and it gave him SMBs immediately. Shouldn't it recognize between reg carbs and correction carbs?
I decided to Temp Target upwards: to 140/120
tynbendad
@tynbendad
Oct 16 2017 22:59 UTC
@danamlewis thx, is the raw change debate in an issue or somewhere? Would like to see any discussion... g5 with dexcom readers don't have raw but you can get it thru xdrip*.
jaylagorio
@jaylagorio
Oct 16 2017 23:02 UTC
I don't know if I follow exactly what you're saying, but what I'm saying is that as soon as he hits 6 units IOB (whether you pushed it or OpenAPS pushes it) OpenAPS will stop doing anything that will add to that. So if you bolus 10 units you're not going to get OpenAPS doing anything to add to it until it falls below 6 units. If you want that to happen you'd have to increase the MaxIOB.
It does not recognize the difference between rescue carbs and regular carbs though
Samuchco
@Samuchco
Oct 16 2017 23:10 UTC
@jaylagorio yes, that is exactly what I meant
It would be useful for the system to recognize rescue carbs vs regular carbs
jaylagorio
@jaylagorio
Oct 16 2017 23:12 UTC
Yeah, that's been talked about here, I think that's on the list of ideas to flesh out
Samuchco
@Samuchco
Oct 16 2017 23:13 UTC
so if you are inputting the carbs via NS treatments it should be easy to differentiate
Scott Leibrand
@scottleibrand
Oct 16 2017 23:17 UTC
the challenge is dealing with all the different ways people enter carbs. we need a good safe least common denominator that always works
right now, the recommendation is "enter a high temp target with your rescue carbs, or OpenAPS will treat them as regular meal carbs"
jaylagorio
@jaylagorio
Oct 16 2017 23:18 UTC
^^
Scott Leibrand
@scottleibrand
Oct 16 2017 23:18 UTC
@ZandMarc looks like autosens is detecting and excluding a lot of meal periods (30g(xxx etc) and some UAM (high IOB) periods (u(xxxx).
ZandMarc
@ZandMarc
Oct 16 2017 23:20 UTC
@scottleibrand yeah, I found the conversation about that on Facebook. Disabled UAM for now.
Is there a way to make the rules in Autosens less "aggressive", like u suggested on Fb?
Scott Leibrand
@scottleibrand
Oct 16 2017 23:22 UTC
let's revisit this conversation this evening: I'm too distracted with various interruptions right now to meaningfully engage
ZandMarc
@ZandMarc
Oct 16 2017 23:22 UTC
:+1:
garykidd
@garykidd
Oct 16 2017 23:38 UTC
Another option to better visualize SMBs would be for a 60 or 90 minute look back windowad opposed to the 3 hour minimum timeframe that currently exits in NS.
jaylagorio
@jaylagorio
Oct 16 2017 23:39 UTC
@danamlewis Looks like my prescription source just started to stock Fiasp. Order placed - one step closer to better no-bolus looping!
Scott Leibrand
@scottleibrand
Oct 16 2017 23:40 UTC
@garykidd and maybe limit the look-forward window to the selected timeframe as well....
garykidd
@garykidd
Oct 16 2017 23:40 UTC
Good idea
Scott Leibrand
@scottleibrand
Oct 16 2017 23:40 UTC
now you just need to find where in the code that is set. ;-)
garykidd
@garykidd
Oct 16 2017 23:41 UTC
Will see if I can find it.
Martin Haeberli
@mhaeberli
Oct 16 2017 23:41 UTC
any ideas about why bluetooth tethering, which has worked well for a long time with a particular Edison / iPhone combination, seems solidly broken now - Edison reboots and iPhone reboots notwithstanding ?
like - what else should I / we try?
Scott Leibrand
@scottleibrand
Oct 16 2017 23:57 UTC
Screenshot 2017-10-16 16.57.11.png
That's no 0 and no U, and no label at all for 0.1U