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

12th
Oct 2017
cameronrenwick
@cameronrenwick
Oct 12 2017 00:46
for ppl who are offline looping, do you have an offline and online rig that are separate? Just thinking with setup of G4 to dexcom receiver to rig this may be "cleanest". Any thoughts or advice appreciated
jaylagorio
@jaylagorio
Oct 12 2017 00:46
Nope, same rig. Goes with me wherever.
cameronrenwick
@cameronrenwick
Oct 12 2017 00:47
so when you're in a wifi/cell environment you just detach the dexcom receiver?
jaylagorio
@jaylagorio
Oct 12 2017 00:49
I'm not sure what you mean, on wifi/cell the receiver stays connected.
cameronrenwick
@cameronrenwick
Oct 12 2017 00:52
sorry .. should'a been clearer.. I have an xdrip+ that wirelessly connects to my rig and android phone. I need to have a totally offline rig for a trip I'm taking (and thus the rig/dexcom/pump arrangement). I plan on using my xdrip+ rig/phone for the majority of time in a wifi setting but when offline the dexcom wired into the rig. Makes sense that you could use the dexcom wired to your rig and be in a wifi environment... I just like to keep things separate for ease of carrying
jaylagorio
@jaylagorio
Oct 12 2017 00:58
Ah, I see what you mean. You might consider doing an oref0-setup.sh where you have the receiver connected to USB constantly and leaving it that way for the trip. It'll minimize trouble trying to switch back and forth to xdrip+ and having the receiver connected, I don't think there's a seamless way to have that work without running setup every time you switch.
cameronrenwick
@cameronrenwick
Oct 12 2017 01:00
exactly my thoughts too... I think I'll take both rigs and that way if something happens I've got the opposite one to fall back on. I just managed to purchase a second hand G4 receiver for $250 that's actually brand new so that's my "camping" rig now
jaylagorio
@jaylagorio
Oct 12 2017 01:01
That's a smart move. I have a spare rig should something go horribly wrong, but not a spare pump for when this one inevitably dies. It's my one point of failure.
cameronrenwick
@cameronrenwick
Oct 12 2017 01:04
lol.. we're so parallel. I had a huge score (better than Christmas!) last week. A friend gave me a brand new 554 Canadian veo with 2.7a firmware. It had been a demo pump for potential pumpers and she said she'd come across it. Brand spanking new... couldn't believe my luck!
my old pump a 522 is still limping along but...
jaylagorio
@jaylagorio
Oct 12 2017 01:04
Good for you, that's amazing!
cameronrenwick
@cameronrenwick
Oct 12 2017 01:05
yes.. I'm a very lucky guy... just hope to be able to pay it forward somehow for others...
jaylagorio
@jaylagorio
Oct 12 2017 01:05
That's the goal
Craig Brenner
@cbrenner
Oct 12 2017 01:28
With oref0 v0.5.5 , seems that I can no longer run Autotune manually via the command: "oref0-autotune --dir=/root/myopenaps --ns-host=https://('yourappnamehere'.herokuapp.com --start-date=2017-10-10 --runs=1 --end-date=2017-10-10" Any ideas as to what changed here?
garykidd
@garykidd
Oct 12 2017 01:58
@scottleibrand ...so weird thing here. manually bolused to try to head off steady rise in BG...all the while small SMBs were firing....at 914 I bolused 5 units...confirmed on pump...NS shows I'm looping...but 5 unit blous not showing in nightscout and my IOB pill shows like 7 units...thinking I should have more--pump shows like 9 units alone...and this doesn't include any of the basal IOB of course....
wondering if the manual bolus wasn't picked up because it hadn't yet confirmed delivery of a 0.2 SMB that had been effected 1 minute prior...wondering if anyone else has noticed/ experienced this?
Scott Leibrand
@scottleibrand
Oct 12 2017 02:03
@cbrenner you don't need --runs=1 any longer
if any of the docs still specify to use that, please PR them to remove it
garykidd
@garykidd
Oct 12 2017 02:16
figuring that documenting this may be important....let me know if I can provide any info/logs/etc....if this is of interest. Double checked it--it's not there.
should I manually enter bolus...don't want to f' it up (further).
Dana Lewis
@danamlewis
Oct 12 2017 02:24
@garykidd what does your OpenAPS pill and logs say?
"it's not there" where? check bolus history - is it on the pump's bolus history list?
garykidd
@garykidd
Oct 12 2017 02:27
they don't show it.
Dana Lewis
@danamlewis
Oct 12 2017 02:27
what is "they"?
garykidd
@garykidd
Oct 12 2017 02:27
neither OpenAPS pill nor logs...like it didnt happen...but pump shows that it did.
and it shows that I am looping all the while.
Dana Lewis
@danamlewis
Oct 12 2017 02:28
ok, what do your logs say? is it successfully looping otherwise? any error messages? is net IOB correct in the logs? (make sure the bolus isn't being attributed to basal)
garykidd
@garykidd
Oct 12 2017 02:31
I cant find errors in log---I did see way later (so thinking probably not relevant to this) that there were some duplicate carbs/boluses removed just while tailing....it's working fine/looping otherwise.
net IOB doesnt appear to be correct AFAICT.
Dana Lewis
@danamlewis
Oct 12 2017 02:31
that sounds related. what time was your bolus, when did you see these "duplicates" removed?
you may want to set a high temp target, or turn off SMB's (or both) temporarily while things get sorted
garykidd
@garykidd
Oct 12 2017 02:32
I entered manual bolus--or at least pump shows it being completed at exactly 914PM ET.
I cant find it now when it was removed.
Scott Leibrand
@scottleibrand
Oct 12 2017 02:34
if you'd like to PM me your NS URL I can take a look
garykidd
@garykidd
Oct 12 2017 02:35
PM is out
My BGs appear to be elevated relating to illness right now.
Scott Leibrand
@scottleibrand
Oct 12 2017 02:37
so I'm looking at the reservoir values in your pump pill, and I am not seeing evidence the bolus was actually delivered. you said 9:14pm EDT right?
garykidd
@garykidd
Oct 12 2017 02:38
exactly...that's my exact point
my pump has it in it's history though
Scott Leibrand
@scottleibrand
Oct 12 2017 02:38
actually I do see it, just a little later
it took a couple minutes to complete I suppose
at 9:16pm reservoir is 27.4, and at 9:20pm it's 22.2U
and before that, 27.65U at 9:13pm
so, that confirms the bolus was actually delivered...
garykidd
@garykidd
Oct 12 2017 02:40
OK, great...didnt think the IOB value in the pill reflected the bolus though...what am I missing?
Scott Leibrand
@scottleibrand
Oct 12 2017 02:40
IOB pill was at 4.83U at 9:13pm EDT, and jumped to 9.6U at 9:16pm EDT
so that definitely reflects it
and indicates OpenAPS knows about the bolus and is looping based on that knowledge
your only problem seems to be that the bolus didn't upload. I assume it's one of the same issues we're starting to try to investigate at openaps/oref0#709, so if you could collect your json files and upload them there, that'd be helpful
garykidd
@garykidd
Oct 12 2017 02:44
so NS just isnt showing it...what confused me was that the IOB within the openaps pill reflected a far lower number than my IOB on my pump....and with the pump not relfecting a lot of the other basal increases I had been experiencing...thinking it must be the effect of the rapid-acting curve --to an entent...
right.
will do.
thanks....so I guess I shouldnt manually add the bolus even though it doesnt show..
Scott Leibrand
@scottleibrand
Oct 12 2017 02:46
right, just leave it and continue as normal
garykidd
@garykidd
Oct 12 2017 02:47
also..comforting to know that despite looking strange..it's working like it should
May be worth noting...that the manual bolus coincided closely in timing with an SMB...may be unrelated--maybe not.
Might be worth noting in documentation how the application of the advanced non-linear curves may also further impact differences in what the pump shows as IOB v APS ( I realize this is largely intuitive for most--but there may be others like me that have somewhat unknowingly like to see those values somewhat in relative parity.
Samuchco
@Samuchco
Oct 12 2017 02:52
Not clear to me. Using 0.6.0 . All SMB, UAM preferences checked "true", Autotune. Yet the SMB starts kind of late. see attached picture. Any ideas?
garykidd
@garykidd
Oct 12 2017 02:52
what causes duplicate entries to exist?
Samuchco
@Samuchco
Oct 12 2017 02:52
Screenshot (4).png
Dana Lewis
@danamlewis
Oct 12 2017 02:54
@Samuchco look at the logs from where your arrows are to see what it says
logs/pill
also how's your connectivity? looks like you may not be looping consistently?
and do you have unentered carbs? or that 80g was entered late compared to when you ate?
Samuchco
@Samuchco
Oct 12 2017 02:59
He (my son) started going up by itself. He entered the carbs and bolus at meal time. I have noticed that the entries are unreliable. See my post to @scottleibrand a couple of days ago. I'm looking for logs at that time. Be right back
sysyc
@sysyc
Oct 12 2017 03:07
snipaste_20171012_104715.jpg
snipaste_20171012_104758.jpg
@scottleibrand why I see the loop is ok in nightscout but error in edison?
Samuchco
@Samuchco
Oct 12 2017 03:15
@danamlewis in the logs, I see a lot of TImed Out errors starting at 16:11 all the way to to 16:22 wher it uploaded/ns-status.json then lots of same Timed out errors
Samuchco
@Samuchco
Oct 12 2017 03:26
Screenshot (6).png
current screenshot
@danamlewis you are probably right about the connectivity. Although I see lots of temp basals throughout the period
garykidd
@garykidd
Oct 12 2017 03:53
OK...I've added some commentary to that #709, @scottleibrand along with relevant .json files. Let me know if I can do anything else on this--thanks again.
LilDucky
@LilDucky
Oct 12 2017 09:41
@scottleibrand @danamlewis installed on a Pi B with carelink (blergh!). Apart from killing off the 4 lines that force password changes and add the edison user to the groups, all went pretty well. I did sudo -i first, then rpi-update (with reboot). all works and queries the pump fine. there is a problem - openaps mmtune returns pump://JSON/read_temp_basal/monitor/temp_basal.json and the alias in openaps.ini is mmtune = report invoke monitor/temp_basal.json
This was on a clean card flashed with debian jessie
Jon Cluck
@cluckj
Oct 12 2017 11:17
@cameronrenwick I set my rig up to default camping mode, and interact with NS if possible. I disabled the wifi on the edison and am connecting only via bluetooth tether to my phone (which is doing all the network transitions because it's better at it ;)). xdrip is sending to both xdripaps on the edison and NS
I guess "local mode" is a better description of it than camping mode
Jon Cluck
@cluckj
Oct 12 2017 11:26
I should get much better battery life on the edison by only using bluetooth -- I'm testing that out today (I only started with the edison rig a few days ago)
Jon Cluck
@cluckj
Oct 12 2017 11:35
my next project is to see if I can get it to sleep/low power mode between loops :) (unless openaps already does this?)
kallnap
@kallnap
Oct 12 2017 12:01
@cluckj do you mean using only BT no BT tethering to your phone, and only using bluetooth serial connection to transmit the data between edison and cellphone?
Jon Cluck
@cluckj
Oct 12 2017 12:04
I'm using BT tethering to my phone for all the data
Brian Rabinovitz
@bdr1177
Oct 12 2017 13:08
FYI - just found out that the IFTTT google calendar service is currently experiencing technical difficulties, so if you have any temporary targets scheduled with google calendar, they may not run correctly. You can sign up for status updates here: http://status.ifttt.com/?utm_medium=Help&utm_source=Applets&utm_campaign=Troubleshooting_Applets
fidoman
@efidoman
Oct 12 2017 13:17
Anyone know where I can find the difference between dexcom g5 filtered vs unfiltered glucose values? Right now, I'm using unfiltered, but wonder if there is a good reference online somewhere explaining how each are derived.
kallnap
@kallnap
Oct 12 2017 13:39
@cluckj thank you for clarifying I was expecting that, but curious if using bluetooth serial rather than bluetooth tethering would be more energy efficient
fidoman
@efidoman
Oct 12 2017 14:47
The docs have a good explanation of filtered vs. unfiltered dexcom data. http://www.nightscout.info/wiki/labs/interpreting-raw-dexcom-data
Velibor Maric
@vebaba
Oct 12 2017 15:01
Hi all. Me again. I made it work!Now, I set it up like this: Libre --> Blukon --> Xdrip+ --> Rig. So, Nightscout is cut off and it doesn't receive data but via Rig
How do I transfer data from xdrip+ to NS? I read it a bit and I found in posts mentioning 'NS uploader'. What's that? Or, I should set it up ...XDrip+ --> NS --> Rig
Can someone pinpoint me to instructions as I'm searching half an hour trying to find it (I saw it, I know) :(
Velibor Maric
@vebaba
Oct 12 2017 15:09
I believe that I found instructions on xdrip to my original question...
Velibor Maric
@vebaba
Oct 12 2017 15:21
I found a problem, no api/v1...
Dana Lewis
@danamlewis
Oct 12 2017 15:26
@lilducky for your Pi/carelink, did you run the setup script fresh and tell it you're using carelink?
Velibor Maric
@vebaba
Oct 12 2017 15:39
What is Alarm HIGH in NS? High blood sugar?
Scott Leibrand
@scottleibrand
Oct 12 2017 15:50
@LilDucky that's not a problem, that's a solution, to the problem that mmtune only works with mmeowlink and a proper radio stick, not with Carelink. In any event, Carelink is completely unsupported nowadays, so I would recommend a cc1110 based radio.
Velibor Maric
@vebaba
Oct 12 2017 15:51
@vebaba It is. Why it sounds? I'm 8.2mmol and I set it @ 180mg/dl?
Scott Leibrand
@scottleibrand
Oct 12 2017 15:51
@cluckj let us know if you see a difference. I don't think the Edison turns off the wifi radio if you don't use it, do not sure how much benefit you'll get.
Velibor Maric
@vebaba
Oct 12 2017 15:59
I found the answer. It's because it is between BG_TARGET_TOP and BG_HIGH
@scottleibrand @danamlewis Everything works (for now). Thank you for your work and support. I'll post only OAPS questions in future, I promise :smile:
Jon Cluck
@cluckj
Oct 12 2017 16:00
I'm not sure if I will see a difference -- I got a little over 12h on a 2000mAh battery yesterday, with the wifi on but with no connected networks. I'm at 6h with the wifi disabled in /etc/network/interfaces (and showing 63% battery left).
Jon Cluck
@cluckj
Oct 12 2017 16:08
anything decided about the next-gen hardware platform, or a direction? I kinda want to trick out this rig (qi, e-paper display, etc) but I'd rather wait until things are more settled on that front to buy anything except batteries :)
Samuchco
@Samuchco
Oct 12 2017 16:11
your only problem seems to be that the bolus didn't upload. I assume it's one of the same issues we're starting to try to investigate at openaps/oref0#709, so if you could collect your json files and upload them there, that'd be helpful
@scottleibrand This is the issue I posted a couple of days ago. First the NS shows pill with carbs but no Units, later on shows a new pill with Carbs and units. How can I help (data)
Scott Leibrand
@scottleibrand
Oct 12 2017 16:16
There is an example in that issue of the types of json data (in a zip file) that would be useful.
@cluckj the Explorer HAT for Pi has a small OLED display and two buttons. No Qi.
Jacob H
@jdhigh
Oct 12 2017 16:29
The past few days I've noticed that in NS I'm getting "Last data received 20 (or 30) minutes ago" messages. However, the openaps pill is getting regular updates. What could be the cause of this? I've checked the mongo stuff and I'm below 500 MB used...
Also, and probably unrelated, but I've noticed that IFTTT is having a hard time from my android updating to NS. I push a button and it has about a 30% fail to run rate.
Jon Cluck
@cluckj
Oct 12 2017 16:39
@scottleibrand the qi is a relatively easy add-on to anything with a power input :)
Samuchco
@Samuchco
Oct 12 2017 16:58
On NS reports/Treatments, when we do a Temp Target it shows only the minutes of the Temp Target but doesn't specify what Temp Target it is.
Craig Brenner
@cbrenner
Oct 12 2017 17:21
Anyone having any recent issues with their IFTTT applets failing when trying to connect to the Google Calendar service to set recurring temp targets. I have seen my Applet fail with the reason, "There was a problem with the Google Calendar service". Worked fine up to a few days ago now random failures. Edited the Google Calendar connection in IFTTT and reconnected the service to IFTTT to see if that would fix anything. Still failures.
cameronrenwick
@cameronrenwick
Oct 12 2017 17:22
@cbrenner @bdr1177 posted earlier that google calendar was experiencing some probs today...it's earlier this am if you scroll up
today.jpg
@scottleibrand et al, I'm wondering if anyone has any suggestions wrt making 0.6.0 a bit more aggressive. I've moved my eSMB to 90 min and up'ed my IOB, max basal etc but just find that if I'm using IFTTT buttons for anything more than about 20 gm of carb, smb just doesn't keep up. I hear others eating 50-60-70+ gm so I'm just not too sure how they get that much insulin. I've attached my today's lunch where I used an IFTTT button for 70gm, tho I only ate about 50 thinking maybe that would help - it didn't. Any thoughts or suggestions .Just fyi..BG pre meal was 6.5.. peaked at 12.0
Dana Lewis
@danamlewis
Oct 12 2017 17:25
@scottleibrand ^ per our convo last night, I asked him to re-post for fodder for additional ideas.
Craig Brenner
@cbrenner
Oct 12 2017 17:27
@cameronrenwick thanks
Scott Leibrand
@scottleibrand
Oct 12 2017 17:31
@cameronrenwick which insulin (and curve) are you using? are you using autotune? what does it say about CR?
cameronrenwick
@cameronrenwick
Oct 12 2017 17:41
@scottleibrand I'm using Fiasp, and have the ultra-rapid curve selected. Since switching to 0.6.0 I haven't run a one off autotune (but it does so each night correct?) and it's put CR at 14 which I feel is pretty accurate for "me"
is there a one-off way to run autotune when using 0.6.0?
Scott Leibrand
@scottleibrand
Oct 12 2017 17:42
if you want to PM me your NS URL I can take a look at the purple prediction lines as it initially starts dosing and see what's limiting it
you can run it from the command line on your rig, but it's slow
cameronrenwick
@cameronrenwick
Oct 12 2017 17:43
PM'd
Scott Leibrand
@scottleibrand
Oct 12 2017 17:48
so looks like your basal is only 0.45u/h, but you need more than 0.7u per SMB to get insulin in fast enough for meals. you might try raising eSMB beyond 90m (15m at a time) to help with that, and back it off if you get any post-meal lows
I also working on some enhancements to make UAM to be more conservative for folks with <0.5u/h basals in situations where you have enough IOB to cause a low (if carb absorption suddenly stops) that couldn't be prevented by zero temping due to the lower basals. once that's ready for dev, it might allow you to safely be a bit more aggressive up front
cameronrenwick
@cameronrenwick
Oct 12 2017 17:51
do you mean max basal or just what NS has displayed? My max basal on my pump is set to 4.5 U/hr
Scott Leibrand
@scottleibrand
Oct 12 2017 17:53
I'm talking about your normally scheduled basal that maxSMBBasalMinutes uses: 0.45u/h * 1.5h = 0.675U, which rounds to 0.7U per SMB
I got 0.45u/h from your NS scheduled basal profile, so it should be fairly close to what you're typically running
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:03
@cameronrenwick I would love to see 0.6.0 be more aggressive.
OpenAPS-Lunch.PNG
that is what I am looking like on a "good" day
just moved "maxSMBBasalMinutes": to 120 upping 15 minutes every few days as per the docs
also set eating soon mode to 90 minutes
Scott Leibrand
@scottleibrand
Oct 12 2017 18:08
@diabetic_dude_twitter what's autotune say about your CR?
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:08
did run autotune last Friday and changed settings but.... it suggested a different CR
adjusted from 7
to 7.8
as it suggested
Scott Leibrand
@scottleibrand
Oct 12 2017 18:10
and which insulin / curve?
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:10
"curve": "ultra-rapid",
Scott Leibrand
@scottleibrand
Oct 12 2017 18:11
so on a meal like that, where you bolus for the full meal with fiasp, and then still rise despite a whole bunch of extra SMB, there's something else going on other than oref0 not being aggressive enough
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:12
yeah it has been common
Scott Leibrand
@scottleibrand
Oct 12 2017 18:12
I would be on the lookout for autotune to start tuning your CR downward if that pattern persists, as you plainly needed way more insulin for that meal than would've been predicted by your CR
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:13
every morning I have the same meal and normally I have to give 60 carbs for a smoothie
now I have the same smoothie but give 80+ carbs
Scott Leibrand
@scottleibrand
Oct 12 2017 18:13
you did 9U of bolus, and your IOB peaked at 13U after all the SMBs
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:14
what day are you looking at?
Scott Leibrand
@scottleibrand
Oct 12 2017 18:15
yesterday lunch, the one you screenshot
I'm adding the 7U and 1.7U boluses to get the 9U (8.7U)
Diabetic_Dude
@diabetic_dude_twitter
Oct 12 2017 18:21
yes for the meal yesterday 76 carbs was what I added
it was PHO that I get all the time.
Scott Leibrand
@scottleibrand
Oct 12 2017 18:22
This one is too far back to examine easily in any more detail, but if you (or anyone else) get another one of these, LMK and I can look at your NS a bit more.
cameronrenwick
@cameronrenwick
Oct 12 2017 18:25
sorry, remind me how we can access what 0.6.0 has determined for autotune values?
Scott Leibrand
@scottleibrand
Oct 12 2017 18:25
cat ~/myopenaps/autotune/autotune_recommend*
cameronrenwick
@cameronrenwick
Oct 12 2017 18:27
perfect.. thx :+1:
LilDucky
@LilDucky
Oct 12 2017 19:45
@danamlewis yes, I did select carelink to test it. @scottleibrand I suggest removing code that is not supported, I reran the oref0-setup for a CC1110 and it did not re-write the alias for mmtune correctly. Will check this again when I can.
Scott Leibrand
@scottleibrand
Oct 12 2017 19:47
well, not supported doesn't mean not useful. I don't like taking action to actively break people's setups, even if I expect them to do any necessary work to keep them working with our new stuff...
LilDucky
@LilDucky
Oct 12 2017 19:49
Also, from what remember, after rerunning the setup the pump.ini did not have a radio type entry for subg-rfspy
Scott Leibrand
@scottleibrand
Oct 12 2017 19:49
did you delete your myopenaps directory and start fresh, or install over it?
LilDucky
@LilDucky
Oct 12 2017 19:50
Install over
Scott Leibrand
@scottleibrand
Oct 12 2017 19:53
k. when changing something major like that, it's probably safer to mv the myopenaps directory and install from scratch. those transitions haven't been well tested, and probably aren't worth doing so.
LilDucky
@LilDucky
Oct 12 2017 20:30
Ok. Thanks. Will hose it and try it.
garykidd
@garykidd
Oct 12 2017 23:56
@scottleibrand what do you think about a NS pill for SMB IOB? Or maybe when you mouseocer IOB it could reveal that. ?