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

29th
Jul 2017
RedCyclist1
@RedCyclist1
Jul 29 2017 01:28
@scottleibrand Hey there...
Shawn
@dyjaks
Jul 29 2017 02:24
So I'm hacking away at getting BLE w/ dexcom to work directly on the edison but progress has been slow due to large real life business. Do we have any documentation for dev debugging? My python experience is trivial and I have no idea if you can just use pdb wrapped around openaps. It's looking like it may not be openxshareble that's broken but instead dexcom_reader.
@dyjaks
(Unless your talking G4 + Share)
Shawn
@dyjaks
Jul 29 2017 02:43
I was intending to make openxshareble work with the G5 but that might work just as well
has that ever been attempted to work on the edison?
ah, just found the gitter channel
Cameron Chunn
@applehat
Jul 29 2017 02:45
=)
Dana Lewis
@danamlewis
Jul 29 2017 02:46
Well also please fix openxshareble for G4 too please :smile:
:+1: props for continuing to work on this, btw
Cameron Chunn
@applehat
Jul 29 2017 02:47
Ignore her. Turn to the G5 darkside. We have blue teeth or something?
Shawn
@dyjaks
Jul 29 2017 02:47
Still think that might be dexcom_reader. I went over the CS history and nothing should have broken it. It looks like dexcom_reader is failing to instantiate it's dexcom object
Why I don't know
since it works for other parts
iananders
@iananders
Jul 29 2017 03:17
small autotune question - is there anyway to see autotune results from the past? i have it running as part of my loop and i only see the current one (autotune_recommendations.log). or do i need to be reviewing every day?
to be more specific - i guess i am looking for the main autotune screen that compares current profile vs the suggested one - i see the different profiles - maybe that's the best way to look back at changes in the past
Dana Lewis
@danamlewis
Jul 29 2017 03:22
Dig into the Autotune folder and you'll see each day's logs
You don't have to view every day, they're saved with a date stamp
iananders
@iananders
Jul 29 2017 03:28
@danamlewis maybe i should scroll down to the bottom of the file before i ask - sorry for being dense!
Dana Lewis
@danamlewis
Jul 29 2017 03:47
:+1: no worries
Scott Leibrand
@scottleibrand
Jul 29 2017 04:49
@RedCyclist1 if you have 0 at IOB (and no recent boluses that might make DIA relevant) and BG is dropping unexpectedly without anything like activity to blame, it probably means your basals are too high.
Deweyoxberg
@Deweyoxberg
Jul 29 2017 05:35

Evening all, flashing jubilinux to edison. I understand from the instructions if the process completes in a few seconds it wasn't successful and to retry flashall.bat ; have done that a few times with no change. Also verified sufficient RAM (Windows) along with specifying paging file; no change. The flash.log file does indicate attempts to flash from today with no further details:

Flashing Edison Board Fri 07/28/2017 23:14:27.25
Flashing Edison Board Fri 07/28/2017 23:16:27.89

Any thoughts?

I do get this in my command prompt while console is spamming away with whatever it's doing:

"Please plug and reboot the board
Dfu device not found Timeout
Did you plug and reboot your board?
If yes, please try a recovery by calling this script with the --recovery option"

Figured I would try flashing despite all warnings to save time :P ;)
katie disimone
@Kdisimone
Jul 29 2017 05:37
are you on mac or windows?
sorry…windows.
#DiabeticinJapan
@LuminaryXion_twitter
Jul 29 2017 05:38
@Katie Yes :)
katie disimone
@Kdisimone
Jul 29 2017 05:38
yes…hold on…I had posted the solution to this earlier. let me see if I can find it
#DiabeticinJapan
@LuminaryXion_twitter
Jul 29 2017 05:39
Thank you!
katie disimone
@Kdisimone
Jul 29 2017 05:39
now to remember what I typed
Deweyoxberg
@Deweyoxberg
Jul 29 2017 05:39
Hah! Isn't that always the case ;)
katie disimone
@Kdisimone
Jul 29 2017 05:39
(I think it was on July 10th if that helps)
@LuminaryXion_twitter Basically, what you need to do is run that flash command…and in the edison’s window when it’s rebooting…interrupt it with control-D. that will give you a boot prompt. At the boot prompt type run do_flash and that will actually get it to flash.
#DiabeticinJapan
@LuminaryXion_twitter
Jul 29 2017 05:42

mayeb this?

@Kdisimone
download jubilinux here http://www.jubilinux.org/dist/jubilinux-v0.1.1.zip

Open on Terminal window…enter cd Downloads/jubilinux

open another terminal window, connect your rig with both cables, and use sudo screen /dev/tty.usbserial-* 115200, enter computer password, login to rig

go back to the first terminal window and type ./flashall.sh

go to second terminal window (the rig’s window) and type reboot

walk away for 15 minutes and then come back. Flashing done
(if you have an existing jubilinux folder in your Downloads, you may want to delete it…the older version makes you enter control-d midway through flash and is a bit of a pita for that)

katie disimone
@Kdisimone
Jul 29 2017 05:42
nope…that’s my mac version
Deweyoxberg
@Deweyoxberg
Jul 29 2017 05:44
Was it around the same time as discussing leaving a console open and working around that lockout?
tynbendad
@tynbendad
Jul 29 2017 05:44
anyone understand git?
katie disimone
@Kdisimone
Jul 29 2017 05:44
bits and pieces
tynbendad
@tynbendad
Jul 29 2017 05:44
i tried "git checkout dev" and "git pull" but it doesn't seem to be using dev
nothing updated
katie disimone
@Kdisimone
Jul 29 2017 05:45
do git branch
see which branch you are on
tynbendad
@tynbendad
Jul 29 2017 05:47
root@benaps3:~/src/oref0# git branch
* dev
  master
katie disimone
@Kdisimone
Jul 29 2017 05:47
git pull should say up-to-date if it is up-to-date
Deweyoxberg
@Deweyoxberg
Jul 29 2017 05:47
@Kdisimone : swapped cables, suddenly WOOOSH
:D
tynbendad
@tynbendad
Jul 29 2017 05:48
root@benaps3:~/src/oref0# git pull
Already up-to-date.
but it doens't have the fix i'm looking for that is in dev branch
katie disimone
@Kdisimone
Jul 29 2017 05:49

@LuminaryXion_twitter here’s what I wrote down to the fellow I was helping out

I think the dfu-util file was left out of the jubilinux folder when you flashed. So it couldn't "find" a partition on the edison to flash. Had to do a series of steps where I issued the flash command in control prompt window, rebooted edison in putty window, pressed control-d while edison booted and then typed run do_flash in the interrupted boot. THEN it took

tynbendad
@tynbendad
Jul 29 2017 05:49
for segfault in openaps/oref0#553
#DiabeticinJapan
@LuminaryXion_twitter
Jul 29 2017 05:49
@Kdisimone THank you!
@Kdisimone , @YYCMichael says he understand. But luckily- this time it was just a cable issue.
katie disimone
@Kdisimone
Jul 29 2017 05:51
well, maybe for the next person too. lol! glad to hear it was cables!
@tynbendad did you do npm run global-install and then run again?
tynbendad
@tynbendad
Jul 29 2017 05:52
yes
katie disimone
@Kdisimone
Jul 29 2017 05:53
oh, it’s in the setup script that the file was changed
So, you can’t do it with runagain…will need to run the setup script plain
tynbendad
@tynbendad
Jul 29 2017 05:58
ok, i see the lines in the setup script so git is aok, will run setup, thanks!
:+1:
katie disimone
@Kdisimone
Jul 29 2017 05:59
yup, should install then and fix the issue
Dirk Gastaldo
@dirkgastaldo
Jul 29 2017 06:04
image.png
Good evening. In what folder are these files?

on your rig…the analogous spot would be

src/oref0/lib/determine-basal/determine-basal.js

oh sorry…you meant where are the json files?
Dirk Gastaldo
@dirkgastaldo
Jul 29 2017 06:09
yes
#DiabeticinJapan
@LuminaryXion_twitter
Jul 29 2017 06:11
how do you delete a character in insert mode in putty?
nevermind!
katie disimone
@Kdisimone
Jul 29 2017 06:12
@dirkgastaldo cd ~/myopenaps/monitor && ls
Dirk Gastaldo
@dirkgastaldo
Jul 29 2017 06:12
Thanks, Katie
katie disimone
@Kdisimone
Jul 29 2017 06:13
also useful is cd ~/myopenaps/settings && ls
both those directories have good info
Dirk Gastaldo
@dirkgastaldo
Jul 29 2017 06:13
excellent, ty
Dirk Gastaldo
@dirkgastaldo
Jul 29 2017 06:16
danke
Deweyoxberg
@Deweyoxberg
Jul 29 2017 06:51
@Kdisimone : all sorted, thank for your help. Question.. how do we tell how much battery power is left on the Edison? is that broadcast to nightscout?
stefanakerblom
@stefanakerblom
Jul 29 2017 08:53
@YYCMichael As far as I know no. You can see it in the pumploop log on the rig
LilDucky
@LilDucky
Jul 29 2017 10:11
hi all, new edison and explorer just arrived and setting up openaps (YAY!!!) now getting error running oref0-setup - the cloning oref0 step fails with "couldn't clone oref0"
tried this logged in as root and edison
trying again as edison and using sudo.
LilDucky
@LilDucky
Jul 29 2017 10:35
found the issue - needed to change the git: to http: in
mkdir -p ~/src; cd ~/src && git clone git://github.com/openaps/oref0.git || (cd oref0 && git checkout master && git pull)
LilDucky
@LilDucky
Jul 29 2017 11:29
Found the root cause - my firewall is as paranoid as I am and blocks ssh outbound by default, thus stopping git:
Ljuslykta
@Ljuslykta
Jul 29 2017 14:04
@YYCMichael I can see how much battery is left on the Edison by hovering over the battery symbol in Nightscout, so I it's possible
Shawn
@dyjaks
Jul 29 2017 14:38
Not really an aps question but can a dying transmitter look like a noisy sensor? My last three have all looked horrible and my g5 transmitter is 130 days old. I use xdrip and it says the battery is fine.
mkericson
@mkericson
Jul 29 2017 14:49
What are you all doing to fix/secure cracks near the cartridge interface on a 522 paradigm?
katie disimone
@Kdisimone
Jul 29 2017 15:39
@YYCMichael yes, the rig's battery is shown on NS's battery pill (looks like my Gitter profile pic). When charging, the pill will always read something like 67% because it is reading the charging voltage, but as soon as you unplug from charger then it will give you actual % of battery life)
cameronrenwick
@cameronrenwick
Jul 29 2017 15:52
@mkericson I've used this stuff https://notaglue.com/ with fantastic success. I was pretty skeptical that it would work but it did really well. I had a large chunk out of the where the battery cap screws into the pump (like about a 1/4 of the whole circular area that's threaded). I took a single wind of plumbing Teflon tape and wrapped it around the battery cap, applied the liquid "not a glue" hardened it with the UV led and it's never given any trouble since - that's over 4 months. I'd recommend using it
Eric
@ecc1
Jul 29 2017 17:12
@dyjaks I doubt there's anything wrong with dexcom_reader, except perhaps in its assumptions about the state of the BLE connection. I have an independently-written version working quite reliably between Linux (Edison or RPi) and G4 Share receiver over BLE. The authentication and communication over BLE are implemented here: https://github.com/ecc1/dexcom/blob/master/share.go
That in turn uses a BLE library similar to the Adafruit/Python one that @bewest used, but is dbus-only (the current preferred way to use Bluetooth in Linux): https://github.com/ecc1/ble
Ljuslykta
@Ljuslykta
Jul 29 2017 17:40
If I terminate a Bluetooth connection between the Edison and my Android phone, is there a simple way to reestablish that connection?
mkericson
@mkericson
Jul 29 2017 18:24
Thanks
elodaille01
@elodaille01
Jul 29 2017 18:25
Hi @scottleibrand , here is the error message I have : (I had to rerun oref0 ) ..what I am suppose to do to fix it please :

Starting supermicrobolus pump-loop at Sat Jul 29 20:15:02 CEST 2017 with 25 second wait_for_silence:
Waiting up to 4 minutes for new BG: ls: cannot access monitor/pump_loop_completed: No such file or directory
Radio ok. Listening: .No pump comms detected from other rigs
Preflight OK. Old pumphistory-24h, waiting for 25 seconds of silence: Radio ok. Listening: .No pump comms detected from other rigs
Old pumphistory-24h refreshed
Old pumphistory, waiting for 25 seconds of silence: Radio ok. Listening: .No pump comms detected from other rigs
Refresh syscall: 'open' }
ed pumphistory
Old settings refreshCould not parse autotune_data
ed. find: `monitor/pump_loop_completed': No such file or directory
pump_loop_completed less than 5m ago. RefreshIndexError: bytearray index out of range

Unsuccessful supermicrobolus pump-loop at Sat Jul 29 20:19:39 CEST 2017

thank you very much ;-)
elodaille01
@elodaille01
Jul 29 2017 18:37
thank you, it's ok now
Steve Lund
@902Lund
Jul 29 2017 18:39
Hi all. I'm still new to Openaps and not very tech savy. I seem to have my loop running now and am trying to figure out what all the data in the logs means. My big issue now seems to be with Night Scout. I followed the directions in Openaps documentation for setting up NS but, it just doesn't seem to be working. I deleted the NS app in Heroku and tried to reinstall with the same results. NS is showing some data from my pump such as bolus wizard entries that I make but, there is no graph showing the CGM values...also the IOB showing is showing that the source is CarePortal which I believe is incorrect. I'm using an Edison and Explorer board with a 522 pump and Enlite CGM. Wondering if anyone has any suggestions to help me troubleshoot? I looked through the rig-NS troubleshooting section in the docs without any success.
RedCyclist1
@RedCyclist1
Jul 29 2017 18:59
@Kdisimone Hi Katie, thought I would ask you on SMB. I have been running it for awhile but with the recent updates, I am running high after I eat. I tend to eat high carb diet for cycling... When I eat, I hit eating soon, then I put in 1 unit from pump, then update Care Portal with planned carbs to eat. I still get very high numbers..at first. Wonder if you might have any ideas or suggestions.
Dana Lewis
@danamlewis
Jul 29 2017 19:09
If you're eating high carb, you will likely find you need to do more bolus. Think about smbs and how much they can give every few minutes. It Might take 45 min or more to build up to 5 units, and given the timing of insulin, that's not ideal.
So if 1unit is giving you highs, then next step would be experiment with up front larger boluses when you KNOW there are carbs that will be kicking in in 15 min
RedCyclist1
@RedCyclist1
Jul 29 2017 19:20
@danamlewis Hey there, didn't see you online.. Thanks for that. So as a rule would you bolus for something like 30% of total carbs eaten?
Dana Lewis
@danamlewis
Jul 29 2017 19:23
YDMV. 30% could be a lot of insulin, or not very much. I think about the number of carbs and how many units of insulin I need total and how many I can do without causing a low when that insulin peaks.
RedCyclist1
@RedCyclist1
Jul 29 2017 19:24
@danamlewis Also when talking with Scott yesterday, I am hitting lows in am... today at 5 I went low, but in atunes I think I am under what ws recommended for basal. My ISF is the same as atunes as well as carb sensitivity.. Just can't seem to stop the early morning lows. Thinking about setting a higher target from 5 - 8am.
Dana Lewis
@danamlewis
Jul 29 2017 19:24
I would get closer to what you'd do for the first part of a meal without SMB, and gauge from there, vs work your way up from zero. Again, thinking about the timing of what SMB can/can't do.
RedCyclist1
@RedCyclist1
Jul 29 2017 19:24
image.png
@danamlewis ahhh that is helpful..I will give that a try
Dana Lewis
@danamlewis
Jul 29 2017 19:25
And if you set different targets, I might back up and set it higher at 4 so it'll do better stopping a drop by 5
RedCyclist1
@RedCyclist1
Jul 29 2017 19:27
ahh yup, makes sense
RedCyclist1
@RedCyclist1
Jul 29 2017 19:37
@danamlewis Last question, when I see I am running at 140 w arrow up, I hit eating soon with a target of 80, but I end up under 50 as the output. Thoughts here?
@danamlewis This is just during the day and not related to eating.
Dana Lewis
@danamlewis
Jul 29 2017 19:38
By output do you mean eventual BG or one of the purple lines?
RedCyclist1
@RedCyclist1
Jul 29 2017 19:39
Sometimes all the purple lines
Could there be some interaction between an increased basal that hits in the middle of the execution of Eating Soon
Scott Leibrand
@scottleibrand
Jul 29 2017 19:48
When your BG is rising, the purple prediction lines will be based on that trend continuing. If you set and eating soon and trigger SMBs to correct that, and then your BG starts to drop, the positive deviations will disappear and the purple lines will predict a low, so it will zero temp to counteract. If ISF etc. are set properly, it should still have time to ward off an actual low by zero temping as you come down. If you actually end up low, then you might need to consider only using eating soon mode if you're actually planning to eat soon, and just let oref0 handle unexpected rises without trying to make it any more aggressive.
RedCyclist1
@RedCyclist1
Jul 29 2017 19:53
@scottleibrand Thanks Scott and Dana, I have been back on a roller coaster BGs and this is all helpful information. I have been laser focused and it just gets worse... I will back off a bit, and let the system try to correct those. What I have chnaged this far is updated basals slightly and adjusted early morning targets. I will back of on small corrections with eating soon and see how it unfolds. Thks
@danamlewis This was really helpful, will give this a try too: I would get closer to what you'd do for the first part of a meal without SMB, and gauge from there, vs work your way up from zero. Again, thinking about the timing of what SMB can/can't do.
Scott Leibrand
@scottleibrand
Jul 29 2017 19:59
to expand on that: if you bolus normally for a meal, SMB won't do much. If you find yourself sometimes going low after your normal meal bolus, you can back it off a bit and let SMB fill in the gaps on meals where it's needed. I think most people using SMB end up doing something like 70% of their normal meal bolus when they have a decent carb estimate.
RedCyclist1
@RedCyclist1
Jul 29 2017 20:02
@scottleibrand Just to be sure I am not missing something, the above is dosing from the pump, correct? It should not be different if I do it from Care Portal or the pump, I am putting in the same carbs depending on which input I use?
elodaille01
@elodaille01
Jul 29 2017 20:12
Are we suppose to have done an 'eating soon' announce in order to see oref1 fight the rise due to the meal?
Dana Lewis
@danamlewis
Jul 29 2017 20:15
@elodaille01 it depends on what you've enabled in preferences for the ways for you to enable SMB and/or UAM.
elodaille01
@elodaille01
Jul 29 2017 20:16
all is set to true @danamlewis
(regarding SMB and UAM)
I tried to enter less carbs than what was going to be eaten..and I didn't succeed..
Dana Lewis
@danamlewis
Jul 29 2017 20:20
If you have set all to true, then you have 3 ways to enable SMBs: via a low temp target, after a large enough bolus, or after carb entry. All that depends on what the system says is needed; so if you don't need as much excess insulin, it may not need to SMB. You should look at the logs to see why it did what it did at any point in time.
elodaille01
@elodaille01
Jul 29 2017 20:22
yes I spend my evening watching my logs ;-)..
there is no microbolus..but a big rise..I certainly did not bolus enough
oh I see now...Autotune did not run since my latest oref0 upgrade (almost 2 hours ago)...so..no autotune -> no SMB
Scott Leibrand
@scottleibrand
Jul 29 2017 20:29
@RedCyclist1 if you enter any carbs from the pump, it will only use pump-entered carbs. if there are no pump carbs, it will check NS. so pick a method and stick with that (at least for the day)
elodaille01
@elodaille01
Jul 29 2017 20:32
@scottleibrand (sorry @RedCyclist1 ), if we enter only some of the carbs really eated, SMB does not do the job ?
RedCyclist1
@RedCyclist1
Jul 29 2017 20:46
@elodaille01 Scott keep me honest, but that is correct, there is no need for SMB to activate otherwise you would go low
Scott Leibrand
@scottleibrand
Jul 29 2017 20:47
if you enter fewer carbs than you actually eat, then as OpenAPS has decays COB to zero, it will gradually switch over to using UAM (if enabled) to continue dosing extra insulin if it appears there is still carb absorption occurring.
RedCyclist1
@RedCyclist1
Jul 29 2017 20:48
@elodaille01 Wait I read your comment wrong... yup what Scott just said.
Scott Leibrand
@scottleibrand
Jul 29 2017 20:49
I just merged a new feature to dev that (if enabled via a new preference) allows you to keep SMB/UAM active for 6h after a carb entry (rather than only for DIA hours) to allow it to do that for larger meals that take longer to absorb.
elodaille01
@elodaille01
Jul 29 2017 20:51
@scottleibrand , just merged..right now (less than 2 or 3 hours ago ?) ;-)
Scott Leibrand
@scottleibrand
Jul 29 2017 20:52
did you know you can subscribe to notifications for all activity on the oref0 repo? :)
elodaille01
@elodaille01
Jul 29 2017 20:52
oops..how could I do it please ? (and sorry )
Scott Leibrand
@scottleibrand
Jul 29 2017 20:53
#561 by scottleibrand was merged an hour ago
no need to apologize there. not many people know that. :) go to https://github.com/openaps/oref0 and click the watch button at the top
RedCyclist1
@RedCyclist1
Jul 29 2017 20:54
@elodaille01 I didn't know that
:)
elodaille01
@elodaille01
Jul 29 2017 20:54
ok, thank you ;) :)
I am the happy fiftieth to watch ...
Iain
@iainct
Jul 29 2017 20:58

Hey, i am experiencing recurring communication issues. From the logs

Temp refreshIndexError: bytearray index out of range
IndexError: bytearray index out of range
grep: enact/smb-enacted.json: No such file or directory
IndexError: bytearray index out of range

and

root@edi_1:~/slicaps# openaps monitor-pump
pump://JSON/read_clock/monitor/clock.json
monitor/clock.json  raised  No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
Traceback (most recent call last):
  File "/usr/local/bin/openaps-report", line 82, in <module>
    app( )
  File "/usr/local/lib/python2.7/dist-packages/openaps/cli/__init__.py", line 51, in __call__
    self.run(self.args)
  File "/usr/local/bin/openaps-report", line 75, in run
    output = app(args, self)
  File "/usr/local/lib/python2.7/dist-packages/openaps/cli/subcommand.py", line 52, in __call__
    return self.method.main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/reports/invoke.py", line 40, in main
    output = task.method(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/uses/use.py", line 44, in __call__
    self.before_main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/medtronic.py", line 64, in before_main
    self.check_session(app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/medtronic.py", line 141, in check_session
    self.session = self.get_session_info( )
  File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/medtronic.py", line 104, in get_session_info
    fields = self.create_session( )
  File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/medtronic.py", line 129, in create_session
    self.pump.power_control(minutes=minutes)
  File "/usr/local/lib/python2.7/dist-packages/mmeowlink/handlers/stick.py", line 222, in power_control
    raise CommsException("No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?")
mmeowlink.exceptions.CommsException: No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
battery is fine and is at close range
I have not found a reliable way to recover from this when it happens...
Scott Leibrand
@scottleibrand
Jul 29 2017 21:02
did you try both powercycling the rig and mmtuning?
Iain
@iainct
Jul 29 2017 21:05
yes, sometimes this works, the trouble is when it starts it can not run for many hours....
if i update the oref it will start looping again
i have also experienced a similar issue on my rig that uses a TI stick with an edison (on sparkfun breakout board)
Scott Leibrand
@scottleibrand
Jul 29 2017 21:07
do you have another pump (even a non-loopable one) to test with?
if so, you might see if you can talk to that pump when the rig is in this state
Iain
@iainct
Jul 29 2017 21:07
I will give that a test next time it comes up
is there anything particular in the logs that is worth recording?
elodaille01
@elodaille01
Jul 29 2017 21:09
sorry @scottleibrand (to disturb your conversation with @iainct ) , I don't see any new preference..perhaps is it in the dev script ?
Scott Leibrand
@scottleibrand
Jul 29 2017 21:12
yeah, you need to actually read the new code if you want to use new features before they're documented. :)
@iainct if you can successfully talk to another pump, but not your normal one, you may have a pump with a flaky radio
Iain
@iainct
Jul 29 2017 21:23
@scottleibrand hmmm, have there been cases of that?
Scott Leibrand
@scottleibrand
Jul 29 2017 21:25
I've seen ones that hardly work at all. not sure if anyone's works sometimes but not others
I would also be interested in seeing your mmtune output when the problem is happening
you might have that in your pump-loop.log from the last time
Iain
@iainct
Jul 29 2017 21:27
i will take a look

@scottleibrand i have just spotted something unusual. my autotune/profile.pump.json contains

"carb_ratio":12.100000000000001,"carb_ratios":{"units":"grams","raw":"0x01 0x00 0x0b 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00","first":1,"schedule":[{"start":"00:00:00","r":11,"ratio":11,"offset":0,"i":0,"x":0}]}}

but my pump is set to 11g/unit i don't understand how it can be 12.1g/unit i spotted this in the auto tune output....

I cant see any mmtune failures in the logs
Iain
@iainct
Jul 29 2017 21:33
as an example, (but i am more concerned about the profile error right now!)
Starting supermicrobolus pump-loop at Sun Jul 30 06:10:03 ACST 2017 with 6 second wait_for_silence:
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
mmeowlink.exceptions.CommsException: No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
Unsuccessful supermicrobolus pump-loop at Sun Jul 30 06:12:34 ACST 2017

Starting supermicrobolus pump-loop at Sun Jul 30 06:13:02 ACST 2017 with 13 second wait_for_silence:
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
mmeowlink.exceptions.CommsException: No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
Waiting for 30s silence before mmtuning
Radio ok. Listening: .No pump comms detected from other rigs
Listening for 30s silence before mmtuning: .No pump comms detected from other rigs
mmtune: "916.636", 0, -99 waiting for 78 second silence before continuing
Radio ok. Listening: .No pump comms detected from other rigs
Unsuccessful supermicrobolus pump-loop at Sun Jul 30 06:17:18 ACST 2017

Starting supermicrobolus pump-loop at Sun Jul 30 06:18:03 ACST 2017 with 27 second wait_for_silence:
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
mmeowlink.exceptions.CommsException: No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
Unsuccessful supermicrobolus pump-loop at Sun Jul 30 06:19:35 ACST 2017

Starting supermicrobolus pump-loop at Sun Jul 30 06:20:03 ACST 2017 with 24 second wait_for_silence:
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
Preflight OK. Profile less than 60m old. glucose.json newer than pump_loop_completed.
Temp refreshIndexError: bytearray index out of range
Unsuccessful supermicrobolus pump-loop at Sun Jul 30 06:21:01 ACST 2017
cameronrenwick
@cameronrenwick
Jul 29 2017 21:41
ya that -99 means your rig and pump are not communicating. I see you've tried to put the 2 closer, that'll help. How new is your pump battery? I've had problems were I've got 1 or 2 bars on the pump but the battery didn't have enough juice to properly power the radio on the pump
Scott Leibrand
@scottleibrand
Jul 29 2017 21:43
yeah, mmtune: "916.636", 0, -99 is the root problem there
cameronrenwick
@cameronrenwick
Jul 29 2017 21:44
@RedCyclist1 I've fiddled around a bit with the % of insulin bolused on my pump (and I'm using fiasp ftr) but I have found personally I try to enter via the bolus wizard the number of carbs, my current BG and when the bolus wizard suggests x amount insulin I back that off about 30%. From there SMB takes over and it seems to work well. This IS going to be a YDMV thing but my real learning came from entering as close to the actual amount of carbs eaten... fwiw, my 2 cents
Iain
@iainct
Jul 29 2017 21:59
@cameronrenwick & @scottleibrand thanks, what link is nominal?
i have another explorer board i could try...
Scott Leibrand
@scottleibrand
Jul 29 2017 22:00
anything better than -80 should let you loop. better than -70 is good.
Iain
@iainct
Jul 29 2017 22:00
damn i never get that
Scott Leibrand
@scottleibrand
Jul 29 2017 22:01
do you have another Explorer board?
Iain
@iainct
Jul 29 2017 22:01
i do, i will give it a test
Re the pump profile carb delta any ideas?
if i manually copy the settings/pumpprofile.json over autotune/profile.pump.json and re-run autotune the estimate for updated carbratio is growing absurdly 13->15->16 etc....
Scott Leibrand
@scottleibrand
Jul 29 2017 22:03
re-running repeatedly on the same day's data?
(not advisable)
Iain
@iainct
Jul 29 2017 22:05
@scottleibrand i thought i read it used the pump profile as the baseline?
I use multiple rigs all the time so i need a method of running autotune on demand rather than once per day....
elodaille01
@elodaille01
Jul 29 2017 22:13
@scottleibrand please, could it be a good idea to set the 'remaining carbs fraction' different from '1' ?
fidoman
@efidoman
Jul 29 2017 22:25
@scottleibrand - I wasn't aware of this fact .... "if you enter any carbs from the pump, it will only use pump-entered carbs. if there are no pump carbs, it will check NS" .... I have been alternating carb entry throughout every day (sometimes the pump, sometimes the apple watch via NS) and I haven't noticed any real issues with SMB not working according to the carbs I enter. Am I misunderstanding your statement?
Scott Leibrand
@scottleibrand
Jul 29 2017 22:31
@ericfeibelman I think I may have been misremembering the older workflow. from the code, it looks like it always uses NS (including pump carbs that have been uploaded) if it can. that may account for the double-carb-entry bug I'm trying to troubleshoot for Bängtsson's as well: if both the canceled and the real carb entry get uploaded, subsequent ns-meal-carb calculations might include it...
elodaille01
@elodaille01
Jul 29 2017 22:36
@ericfeibelman when you enter your carbs via NS (care portal), how do you bolus ?
Scott Leibrand
@scottleibrand
Jul 29 2017 22:38
one method is to easy bolus from the pump: hit the up arrow as many times as needed, hit Act to select, then count the number of buzzes/beeps and hit Act again if it's right.
you can adjust the easy bolus interval in settings
elodaille01
@elodaille01
Jul 29 2017 22:38
ok that what I thought
LilDucky
@LilDucky
Jul 29 2017 23:35
hi again, needs some assistance troubleshooting a new edison install. had issues with my firewall blocking things but now radio comms are not working.
how do i remove the entire install and start again?
Scott Leibrand
@scottleibrand
Jul 29 2017 23:37
mv ~/myopenaps ~/myopenaps.bak
and re-run oref0-setup
or start over with the package install steps if you think any of those failed
LilDucky
@LilDucky
Jul 29 2017 23:38
had issues with edisonbattery as well.
will try that as well as removing the src and re-cloning it
Scott Leibrand
@scottleibrand
Jul 29 2017 23:41
if your installations are damaged beyond repair, you can also re-flash the Edison.
LilDucky
@LilDucky
Jul 29 2017 23:42
hope not. have removed src and myopenaps, re-pulled and running setup now
wish me luck