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

17th
Jun 2017
Zakm0n
@Zakm0n
Jun 17 2017 00:05
This is unreal, i've never had this much trouble with linux based stuff
rsilvers129
@rsilvers129
Jun 17 2017 00:12
Thanks
marafie
@marafie
Jun 17 2017 00:41
Can someone give me some guidance how to calculate them requires insulin when eating protein ? 20 protein = X carbs to enter in the pump
The=them
Dana Lewis
@danamlewis
Jun 17 2017 01:05
You can probably google some suggestions - everyone does it differently (including some don't at all calculate protein or fat)
fidoman
@efidoman
Jun 17 2017 01:14
Yep, I don't usually count anything for protein and fat. Sometimes I'll bolus less for a meal with high protein or fat levels because carb absorption is much slower for me with a fatty meal.
Kate Farnsworth
@ELUTE
Jun 17 2017 01:43
@danamlewis will do as soon as I can
@marafie a typical starting point for a low carb high protein meal is 20 protein = 10 carbs
garykidd
@garykidd
Jun 17 2017 02:12
Fyi. Just wait he'd to android. S8+ nice. Even though it isn't officially supported yet ,,,was able to get dexcom g5 mobile app running perfectly. Unfodtunately,,,no luck with xdrp+
Wait=switched
rsilvers129
@rsilvers129
Jun 17 2017 02:54
Sometimes when I am high and a lot of SMBs are piling up, and when I eventually come down, I go low. Is the fix for that to change my duration of insulin action from 3 to 4? Or verify and raise my ISF? I don't think the ISF is it because autosense or autotune uses what it wants for that anyway.
Cameron Chunn
@applehat
Jun 17 2017 03:06
I've never calculated fat or protine in glucose rises because they have always been super minimal. When I was eating keto before my loop, just my base basal was plenty to keep me from ever going high.
And that died has tons of fat and protine.
That diet*
@rsilvers129 do you have pushover notifications enabled?
Pushover tells me when SMB overshot before I go low, and asks me to eat some carbs.
rsilvers129
@rsilvers129
Jun 17 2017 03:13
IMG_7063.PNG
@applehat Yup. Here is me sleeping through a whole lot of them. I can’t think of a way to have it wake me up. I need something like if carbsReq > ISF/10, then tell Alexa to wake me up. I often see CarbsReq value of 1. I am not sure I need to be alarmed about that.
garykidd
@garykidd
Jun 17 2017 03:20
@danamlewis did I retarget #858 correctly (to master)?
katie disimone
@Kdisimone
Jun 17 2017 03:26

@danamlewis do you remember for the 2016 explorer boards for raw data on G4....was the flakiness that some boards worked, some didn't? Or was it all boards worked sometimes, sometimes didn't?

One of my 2016 boards is picking up raw and I'm wondering if is likely this board may end up being reliable because it just demonstrated the ability to do this...or if raw is hit and miss on all the 2016s

Katie Aldridge
@kcrcgm
Jun 17 2017 03:42
BT tethering problem.JPG
I've run into a problem setting up BT tethering with my phone. I have a Samsung Galaxy S7 (yay it should work) but my provider is Tmobile (boo they've turned off BT tethering). So I downloaded FoxFi, which is supposed to get my phone to do BT tethering and hide it. I worked through the BT tethering section on the docs (http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-4/bluetooth-tethering-edison.html) and successfully paired the Edison to my phone. But then I got to "Testing to make sure it works after you successfully did the above" and got an error when I tried to "establish a Bluetooth Network connection" with my phone. Here's a screenshot of the error. I'm not sure, but I think it's telling me that the Foxfi thing didn't work... Can somebody look at the error and let me know what you think? And any suggestions for what to do? Thanks in advance!
Cameron Chunn
@applehat
Jun 17 2017 03:54
@rsilvers129 does Alexa support push notification yet?
Scott Leibrand
@scottleibrand
Jun 17 2017 04:36
@rsilvers129 how does your autotuned ISF compare to your pump's? Is it being limited by the min/max caps?
Matthias
@ceben80
Jun 17 2017 04:38
@scottleibrand Well, tbh I don't care what people are posting but I can remember in the beginning of rtd, one important point was a comment "do not post your serial", anyway....
Scott Leibrand
@scottleibrand
Jun 17 2017 04:39
yeah, we originally made a big deal of that, until I realized just how easy it is to sniff
Matthias
@ceben80
Jun 17 2017 04:42
saw this video above, great and impressive (bt sniffing)
Scott Leibrand
@scottleibrand
Jun 17 2017 04:43
there is a single mmeolink command that just prints out all the rig-pump comms it sees, including the serial numbers. so if anyone is under any illusions that their rig-pump comms are the slightest bit secure, they need to realize they're not
of course, if someone wanted to hurt you, there are far easier ways to do it than trying to make your pump dose you too much insulin.
Matthias
@ceben80
Jun 17 2017 04:47
then quickly need a huge carb bucket :smile:
today a big BBQ party with tons of :pig2: and carbs :tada: (...and Adrian is coming :star2: )
Scott Leibrand
@scottleibrand
Jun 17 2017 05:01
nom
Matthias
@ceben80
Jun 17 2017 05:02
....SMB stress test :wink:
sefish99
@sefish99
Jun 17 2017 05:13
@rsilvers129 I ended up changing my DIA when I was ending up with a bunch of lows using SMBs. It’s worked amazingly well. I couldn’t believe the difference.
rsilvers129
@rsilvers129
Jun 17 2017 05:14
I just made the change from 3 to 4 now. Just had a low. I will see how it goes.

I just helped a friend do a runagain, and now it is saying
mmeowlink.exceptions.CommsException: No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
Pump is right beside rig. I had him reboot. I had him pull power. Nothing fixes it.

This is on dev.

rsilvers129
@rsilvers129
Jun 17 2017 05:38
I had him change to another Explorer with the same Edison, and now it says mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Cameron Chunn
@applehat
Jun 17 2017 06:04
I fixed that issue by bending my explorer board and jamming it into my case. But uh, that's not actually the fix and don't do that.
Also I ate 2 cupcakes with up front guess-boluses and SMB only let me hit around 160! (I then ate something else and forgot to bolus but that's another story)
(had to make up for that lunch today that I managed to not suck at diabetes on 😅)
sysyc
@sysyc
Jun 17 2017 06:06
snipaste_20170617_135636.png
snipaste_20170617_135652.png
@scottleibrand Why don't show smb and temp on nightscout but displayed on the pill?
Matthias
@ceben80
Jun 17 2017 06:14
@Kdisimone amazing troubleshooting section in rtd :+1:
rsilvers129
@rsilvers129
Jun 17 2017 06:14
Thanks. I am looking at that now. What does it mean when the docs say "Remove and re-add your pump device"
katie disimone
@Kdisimone
Jun 17 2017 06:15
That is leftover from old manual OpenAPS installation directions.
@ceben80 thanks! Still have some more plans for expansion but it's a start
PieterGit
@PieterGit
Jun 17 2017 06:18
@rsilvers129 : stopping loop, killing all pump processes and running openaps mmtune fixes this in most cases. If you have a succesfull mmtune and a manual loop, start the (pump) loop again
rsilvers129
@rsilvers129
Jun 17 2017 06:19
Checking deliverAt: null is within 1m of current time: Sat Jun 17 00:19:10 MDT 2017
date: invalid date ‘null’
Error, retrying
Waiting for 30s silence before mmtuning
rsilvers129
@rsilvers129
Jun 17 2017 06:27
mmtune is working. But I have a null date problem.
Checking deliverAt: null is within 1m of current time: Sat Jun 17 00:26:13 MDT 2017
date: invalid date ‘null’
Error, retrying
Sleeping 6;
It seems to be issue 461. openaps/oref0#461
rsilvers129
@rsilvers129
Jun 17 2017 06:37
The radio error went away for whatever reason. The null date error was fixed when I removed microbolus from the runagain. So it is working again, but never worked with SMB.
Matthias
@ceben80
Jun 17 2017 06:42
Did you enabled SMB in preferences.json?
Frank Best
@efbest
Jun 17 2017 09:33
I had to switch pumps. The loop now hangs trying to refresh 'old pump history'. Rebooting, shutting down, calibrating without success. Any ideas?
Radio ok, Listening: .No pump comms detected from other rigs MDT cgm data retrieved MDT cgm data reformatted Old pumphistory: Refresh Error, retrying
Matthias
@ceben80
Jun 17 2017 09:47
Did you re-run the setup with new serial or simply modify your pump.ini with your new serial? @efbest
Frank Best
@efbest
Jun 17 2017 09:48
re-run the setup with new serial
Matthias
@ceben80
Jun 17 2017 10:10
killall -g openaps && openaps use pump model in your myopenapsdirectory, did you get the pump model?
@efbest but sorry time shortage due to party preps here :wink:
Frank Best
@efbest
Jun 17 2017 10:12
thanks @ceben80 😎
yes '722' @ceben80
Matthias
@ceben80
Jun 17 2017 10:18
that's great, generally the comm is working :+1:
@efbest I think for a fully data retrieval it needs some time due to the fact all rig data belongs to your 754 before...
fidoman
@efidoman
Jun 17 2017 10:42
Anyone know why I get this error periodically in my logs? I've searched through the code and can't find the "Process is running" error message. glucose.json newer than pump_loop_completed. Temp refreshprocess is running and remove the file manually to continue.' Error, retrying Sleeping 10;
Tim Street
@tim2000s
Jun 17 2017 11:40
@JELCRAWFORD or was it someone else? I was unable to update the wpa supplicant packages using apt-get because without the wifi there is no network connection. I'd have to go throuh a manul aprocess to try and move them via the uart serial connection.
That's why i was having the issue on the edison
cameronrenwick
@cameronrenwick
Jun 17 2017 13:08
@kcrcgm I think you are correct when you say that the foxfi does not allow BT tethering. I too have a S7 and am on Rogers wireless in Canada. BT tethering is turned off. I successfully wifi tether tho. I've researched and used several apps but didn't manage to get BT tethering to my rig. I have learned that by rooting the S7 and reinstalling the OS that BT tethering can be switched on... it voids the warranty and has some potential to brick your phone (2 things I'm not super pleased about) so... I've got an old android phone that I'm going to update to nougat OS (same as the S7) and practice on it to make sure the BT tether works... if successful I'll attempt on my S7. I can't do that for about 2 weeks as I've just got too much school going on now to dig into "another" problem
John Benjamin
@Bender1061
Jun 17 2017 13:47

can anyone confirm if they see this error every so often in their pump loop log

pump_loop_completed less than 5m ago. RefreshWarning: clock input Invalid Date is unzoned; please pass clock-zoned.json instead
ed pumphistory

I'm wondering if it's a MDT specific issue or not

does not happen all the time, but it happens enough that I want to fix it
JELCRAWFORD
@JELCRAWFORD
Jun 17 2017 15:03
@tim2000s - might have been @jmcrawford - but not me. Sounds like you got it sorted though :)
ejoe132
@ejoe132
Jun 17 2017 15:13
I'm having a problem since last night, was out biking and battery died, no big deal, got home and plugged it in and it isn't looping. Any ideas? Heres the log... File "/usr/local/lib/python2.7/dist-packages/mmeowlink/link_builder.py", line 16, in build
return SubgRfspyLink(port)
File "/usr/local/lib/python2.7/dist-packages/mmeowlink/vendors/subg_rfspy_link.py", line 55, in init
self.open()
File "/usr/local/lib/python2.7/dist-packages/mmeowlink/vendors/serial_interface.py", line 28, in open
self.check_setup()
File "/usr/local/lib/python2.7/dist-packages/mmeowlink/vendors/subg_rfspy_link.py", line 72, in check_setup
self.serial_rf_spy.sync()
File "/usr/local/lib/python2.7/dist-packages/mmeowlink/vendors/serial_rf_spy.py", line 121, in sync
raise CommsException("Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?")
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Usage: grep [OPTION]... PATTERN [FILE]...
Try 'grep --help' for more information.
Listening: ....................................................................................................Error, retrying
Listening: ...............................
ejoe132
@ejoe132
Jun 17 2017 15:37
checked in cat pump.ini and have this in as the port and radio type, I believe this looks correct.... port = /dev/spidev5.1
radio_type = subg_rfspy
wharper11
@wharper11
Jun 17 2017 15:37
hello. I accidentally hit "unpair" on my android phone and the edison rig is no longer paired on my phone. I am currently on a serial session with the edison and would like to try to re-pair. is there a command to get the rig to link back up to the phone?
wharper11
@wharper11
Jun 17 2017 15:43
is it possible to reconnect using a serial session?
ejoe132
@ejoe132
Jun 17 2017 15:44
@wharper11 Have you been pairing with a wifi hotspot?
wharper11
@wharper11
Jun 17 2017 15:45
we're camping and just need the bt tether
accidentally hit unpair
no, @ejoe132 - bt tether
ejoe132
@ejoe132
Jun 17 2017 15:49
@wharper11 Oh, actually still trying to setup my bt connection myself. I would try serial and see if you can get it working
jdunn42
@jdunn42
Jun 17 2017 15:53
@wharper11 You can run bluetoothctl and pair the devices just like you did initially.
katie disimone
@Kdisimone
Jun 17 2017 16:24
@wharper11 yeah, I think that's bad news...unpairing will require re-pairing using the same series of instructions you originally setup with. Before you do that though, you could try...I'm not sure it will work...simply entering bluetoothctl and then trying the command pair AA:BB:CC: with your BT mac of course. See if it send a pair command again. I doubt it, but worth a shot. If it doesn't work, you'll want to remove AA:BB:CC (equivalent to forgetting a paired device) so that the rig will be able to pair with it again when you go through those commands in the docs again. Similar to pairing a new dexcom receiver, need both devices to forget each other before you're able to re-pair them.
Cameron Chunn
@applehat
Jun 17 2017 16:26
Alternatively, edit /etc/rc.local and enable auto-pair, reboot, and then try pairing from your phone.
(this isn't documented as far as I know - just something in Jubilinux that seems to work for me)
hilarykoch
@hilarykoch
Jun 17 2017 16:41
I had a rig stop running last night after running autotune report. Suddenly "cannot read property 'sensitivities' of undefined” appeared and seems to have halted everyting. I have other rigs, so we’re not without. Just trying to identify what may have happened and how to fix it. ISF is 65 (or maybe 70) and rig is reading pump correctly. Autotune report was completely messed up, so I tried erasing autotune directory, but problem is still there. Anyone with ideas before I rerun script?
Screen Shot 2017-06-17 at 11.06.04 AM.png
francesc0-cgm
@francesc0-cgm
Jun 17 2017 16:55
I have a big problem. I did not used openaps for a while. Today i'm in a bad signal area and i have to use it off line. But i have only could not get subg_rfspy state atc
Tried to rerun oref0 again
Flash the edison chip using ccprog
But same error message
Manual mmtune gives me a spider already in use message
It tells waiting for 30 secs of silence before mmtuning and it doesn't perform it
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:08
Nmp run global-install doesn't work at all
Don't know what happened
marafie
@marafie
Jun 17 2017 17:09
If i wanted check the past basal history on the rig which file should I open?
Scott Leibrand
@scottleibrand
Jun 17 2017 17:15
@francesc0-cgm if you get spidev 5.1 already in use, it means you didn't successfully kill your already-running pump-loop process before running manual mmtune
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:15
Which is the command?
Because it is stucked on waiting for 30 s for silence before mmtune but it doesn't perform mmtunr
Scott Leibrand
@scottleibrand
Jun 17 2017 17:16
the most complete thing you can do is comment out pump-loop in your cron and reboot
@hilarykoch please save a copy of any files you still have, particularly autotune log files, so we can try to figure out what it got stuck on
what likely happened is that the bad autotune output got copied over to your ~/myopenaps/settings directory as well. there are ways to fix that manually, but if you've already deleted some things, it might be easiest to save whatever you have left and then re-run oref0-setup
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:18
Tryimg
Scott Leibrand
@scottleibrand
Jun 17 2017 17:19
@ejoe132 did you run through the troubleshooting doc linked a few hours ago?
@ericfeibelman that is the last line of a multi-line error complaining about a stale lock file. oref0-reset-git has some logic to clean that up, so I wouldn't worry about it unless it's persisting long enough to break your loop.
@Bender1061 yes, that "pass clock-zoned.json" is something specific to the MDT CGM setup. if you can dig in further, that'd be great, since we don't have the setup to troubleshoot that.
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:21
@scottleibrand did it but i have always comms error
Explorer board has gone
Rig is next to the pump
Riley is working using loop
So it could be a problem with the explorer board
Scott Leibrand
@scottleibrand
Jun 17 2017 17:24
@rsilvers129 the date: invalid date ‘null’ thing is usually a symptom of something earlier in the log not running successfully. do you have the full pump-loop.log output for one of those failing runs?
@francesc0-cgm what are your two LEDs next to the USB ports doing?
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:24
Tried also to detach and put again edison on explorer
One red
Sorry led are off
Status led (the big one) is on
rsilvers129
@rsilvers129
Jun 17 2017 17:26
No. Not my rig. But I think it was related to me copying and pasting my Heroku enable variables into his, and a copy error leading to bridge not being there. So he was not getting Dexcom data. I couldn't figure it out from the error messages.
Scott Leibrand
@scottleibrand
Jun 17 2017 17:26
@sysyc you have to wait until the next pumphistory pull and ns-upload before a microbolus or temp basal will be uploaded to NS and displayed. the openaps pill is updated right after it's enacted.
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:26
When i reboot they blink then they get off as usual
Scott Leibrand
@scottleibrand
Jun 17 2017 17:26
@rsilvers129 yeah, "BG data is too old" is usually the real culprit behind the invalid date stuff
@francesc0-cgm ok, that's good. and after a reboot what does your openaps mmtune output look like?
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:28
Could not get subg etcetc
Scott Leibrand
@scottleibrand
Jun 17 2017 17:29
ok, try ccprog again maybe, and note whether it succeeds, and what the LEDs do afterward?
if they are on solid after the ccprog flash, openaps mmtune should make them blink and go off
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:31
After falshing they are off
After erasing they were on
Solid
No way
Same error
After mmtune same commsexception error
Scott Leibrand
@scottleibrand
Jun 17 2017 17:43
not sure. at this point you may either want to start reinstalling things, or try swapping your Edison to a new Explorer board and see which device the problem follows
francesc0-cgm
@francesc0-cgm
Jun 17 2017 17:45
Already waiting for a new board but it is somewhere...
Seems strange because the rig was working and it was sitting in my desk
hilarykoch
@hilarykoch
Jun 17 2017 17:48
@scottleibrand I kind of understand and kind of need a translator. :smile: “comment out pump-loop” means to clear something - just not sure HOW to do that. Same for saving. I understand that I should save the files. Just not sure HOW. (And yes, think once I learn that, I might PR something somewhere, because I get that these are probably key/basic skills.)
Brian Bonner
@itsn0tm3
Jun 17 2017 17:50
need to use reset_spi_serial when seeing the commsexception. I see that all the time on my 2017 boards but never on my 2016 boards
Scott Leibrand
@scottleibrand
Jun 17 2017 17:50
"comment out" means "add the comment character(s) to the beginning of the line." In the case of cron and bash scripts, the comment character is #. (in the case of .js files, they're //)
@itsn0tm3 I think openaps mmtune does reset_spi_serial on every run
Brian Bonner
@itsn0tm3
Jun 17 2017 17:51
it doesn't for me :(
Scott Leibrand
@scottleibrand
Jun 17 2017 17:52
@itsn0tm3 have you checked? openaps alias show mmtune
@hilarykoch and saving a file after you edit it is already documented I think: it's Ctrl-X for nano, or Esc :wq for vi
Brian Bonner
@itsn0tm3
Jun 17 2017 17:52
let me log in and check
root@tron:~/tron# openaps alias show mmtune
mmtune ! bash -c "oref0_init_pump_comms.py; find monitor/ -size +5c | grep -q mmtune && cp monitor/mmtune.json mmtune_old.json; echo {} > monitor/mmtune.json; echo -n \"mmtune: \" && openaps report invoke monitor/mmtune.json; grep -v setFreq monitor/mmtune.json | grep -A2 $(json -a setFreq -f monitor/mmtune.json) | while read line; do echo -n \"$line \"; done"
I just booted up Tron (one of my 2017 boards) and seeing the commsexception
it never recovers.. I always have to log in and issue the killall and then reset spi then it works beautifully after that
hilarykoch
@hilarykoch
Jun 17 2017 17:56
@scottleibrand It’s still all Greek to me. I know I’ve added # to lines when setting up OpenAPS initial. And I remember it being called comment out. But unclear as to how to comment out pumploop to solve this problem. And to my second question, I know when I edit say preferences, I know it saves as I exit (or asks me if I want to save), I understood you to ask me to save a specific autotune logs - also unclear how to do this. Sorry. really trying to understand, just not getting it.
Brian Bonner
@itsn0tm3
Jun 17 2017 17:56
mmtune just ran and seeing the commsexception now
pump://JSON/mmtune/monitor/mmtune.json
Traceback (most recent call last):
(massive cut)
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Usage: grep [OPTION]... PATTERN [FILE]...
Try 'grep --help' for more information.
 Listening: .........
I need to look into the init pumps comm script and figure out where the grep is failing and then fix that.. just haven't had time to do it yet
Scott Leibrand
@scottleibrand
Jun 17 2017 17:58
@itsn0tm3 ah, sounds like you're having trouble with the stuff @PieterGit changed then
@hilarykoch ah, sorry, I thought you meant how to save the file after commenting something out
francesc0-cgm
@francesc0-cgm
Jun 17 2017 18:02
@itsn0tm3 seems to have your same issue. How do you manage it?
Brian Bonner
@itsn0tm3
Jun 17 2017 18:03
not very well ha, I've been travelling quite a bit lately so I haven't had time to really look into it. Right now I just carry around my HooToo and just turn it on if I need to reset a 2017 board and just log in and issue a "killall -g openaps; reset_spi_serial.sh" then it works fine after that as long as I don't reboot the board
Scott Leibrand
@scottleibrand
Jun 17 2017 18:03
@hilarykoch as far as saving a copy of a file, that can be something as simple as displaying the file and copying and pasting it somewhere on your computer for safe keeping. or you can use cp to copy it somewhere that won't get deleted...
@itsn0tm3 are you running with or without SMB?
hilarykoch
@hilarykoch
Jun 17 2017 18:05
@scottleibrand Ok. That makes sense. I can do that. :) Now back to first question then. You said I should, "the most complete thing you can do is comment out pump-loop in your cron and reboot” — How do I comment out pump-loop?
Scott Leibrand
@scottleibrand
Jun 17 2017 18:06
crontab -e to edit your crontab. find the pump-loop line and add # to the beginning of it. save and quit.
@hilarykoch also, I think that was a suggestion for someone else, not for you
francesc0-cgm
@francesc0-cgm
Jun 17 2017 18:07
It was for me
Scott Leibrand
@scottleibrand
Jun 17 2017 18:07
but I'm getting all these convo's confused
commenting out pump-loop is for making sure you don't get spidev5.1 already in use errors when running mmtune manually
francesc0-cgm
@francesc0-cgm
Jun 17 2017 18:07
@itsn0tm3 till now it worked always quite well
I will try this way to fix it
Scott Leibrand
@scottleibrand
Jun 17 2017 18:08
@francesc0-cgm same question for you: are you currently using SMB or not?
the oref0-pump-loop used by SMB does a reset_spi_serial.py inside the mmtune funciton, but it doesn't look like openaps mmtune used by openaps pump-loop does
hilarykoch
@hilarykoch
Jun 17 2017 18:08
@scottleibrand Ha! Ok, I looked back and see that now. This is personally why I prefer FB for troubleshooting - it’s too easy to get confused with multiple convos here. THANK YOU! And seriously want y’all to do an OpenAPS tutorial session at FFL. Any chance of that?
Scott Leibrand
@scottleibrand
Jun 17 2017 18:09
@hilarykoch you'll have to ask @danamlewis about that. I'm not going to be at FFL myself, and can't remember what her plans are
hilarykoch
@hilarykoch
Jun 17 2017 18:10
Gotcha. @danamlewis with sugar on top???
Scott Leibrand
@scottleibrand
Jun 17 2017 18:12
if you want to submit a PR to re-add reset_spi_serial.py 2>/dev/null; prior to the new oref0_init_pump_comms.py, that's the workaround that worked for me in oref0-pump-loop.sh
so that mmtune alias command would start with reset_spi_serial.py 2>/dev/null; oref0_init_pump_comms.py; find monitor/ -size +5c etc...
francesc0-cgm
@francesc0-cgm
Jun 17 2017 18:14
I'm using smb
It is on on oref0 setup and disabled on preferences atm
Scott Leibrand
@scottleibrand
Jun 17 2017 18:15
@francesc0-cgm ok, in your case oref0-pump-loop should already be doing the reset on every mmtune then
(make sure you're seeing supermicrobolus pump-loop in your pump-loop.log to confirm you're using oref0-pump-loop)
francesc0-cgm
@francesc0-cgm
Jun 17 2017 18:16
Yes i have
Also doing a reset manually i have same problem on mmtune
It should be something hardware related
Scott Leibrand
@scottleibrand
Jun 17 2017 18:17
yeah, sounds like it. just to confirm, can you try reset_spi_serial.py and then openaps use pump mmtune ?
@itsn0tm3 please try updating your openaps mmtune alias to include reset_spi_serial.py and see if that fixes your need to do so manually. if so, a PR to do so for everyone else would be much appreciated.
Brian Bonner
@itsn0tm3
Jun 17 2017 18:27
will do
aabramowicz
@aabramowicz
Jun 17 2017 18:27
mmtune.JPG
What file do I remove to get the loop running again? Running Dev
Brian Bonner
@itsn0tm3
Jun 17 2017 18:28
@scottleibrand one interesting thing for me was that it was in the @reboot on cron, but that didn't work so I added a sleep just to see what that would do and it looked like @reboot (sleep 120; killall -g openaps; reset_spi_serial.py) & but that didn't work either
John Benjamin
@Bender1061
Jun 17 2017 18:35
So I'm having an issue trying to get something to work.
I want to take a Time string and turn it into Seconds.
If I do a jq .[0].display_time monitor/glucose.json
I will get something like "2017-06-17T14:17:00-04:00"
If I take that String and do date -d "2017-06-17T14:17:00-04:00" +%s
I get 1497723420 exactly as I want.
But if I try to combine it into one
date -d '(jq .[0].display_time monitor/glucose.json)' +%s
I get 1497672000 which is seconds to the beginning of the day.
I've been trying to figure this out for hours, wondering if someone can get me any help on what I'm doing wrong.
I know it has nothing to do with the timezones, I'm trying to do something specific for the MDT stuff, so I'm using for my example a file I believe everyone has.
Scott Leibrand
@scottleibrand
Jun 17 2017 18:40
Try date -d $(jq whatever)
John Benjamin
@Bender1061
Jun 17 2017 18:41
does the same thing
Scott Leibrand
@scottleibrand
Jun 17 2017 18:41
Your single quotes don't invoke what is inside them. Backtick would, but those are hard to escape etc.
John Benjamin
@Bender1061
Jun 17 2017 18:42
sorry, not the same thing, it says the date is invalid
root@ED209:~/johnopenaps# date -d $(jq .[0].display_time monitor/glucose.json) +%s
date: invalid date ‘"2017-06-17T14:27:00-04:00"’
root@ED209:~/johnopenaps#
```
Scott Leibrand
@scottleibrand
Jun 17 2017 18:43
You need to get rid of the "s
You can use | tr -d '"' for that.
John Benjamin
@Bender1061
Jun 17 2017 18:44
Oh thank you so much!
it worked
I get what I was doing wrong now
Dana Lewis
@danamlewis
Jun 17 2017 18:49
@hilarykoch i land 5:30pm on 4th and leave 5:30pm on 6th..so I'm sure we can find time :D question will be how many others there will want to sit down together too.
hilarykoch
@hilarykoch
Jun 17 2017 18:54
🎈🎉🕺💃 @danamlewis yay!
Brian Bonner
@itsn0tm3
Jun 17 2017 18:54
@scottleibrand Looks like the change back to reset spi does help it recover from that error
ejoe132
@ejoe132
Jun 17 2017 19:16
@itsn0tm3 Is this for the commsexceptioon error? Having trouble with that and not sure what is wrong as my pump.ini is correct for the the serial number
Brian Bonner
@itsn0tm3
Jun 17 2017 19:18
@ejoe132 Yes, a good test is to go into your openaps directory and type in killall -g openaps; reset_spi_serial.py; openaps mmtune
although mmtune for you might re-run the reset spi
ejoe132
@ejoe132
Jun 17 2017 19:19
@itsn0tm3 Ok Ill give that a shot, that just resets the serial number and that may kick it back into working?
Brian Bonner
@itsn0tm3
Jun 17 2017 19:20
no it just resets the connection to the SPI chip on the explorer board
it wont' reset your serial number after all that's in your pump.ini file
ejoe132
@ejoe132
Jun 17 2017 19:23
@itsn0tm3 ok got it, just was trying installing the master branch now so when thats done Ill switch back to the dev and give that a shot
Brian Bonner
@itsn0tm3
Jun 17 2017 19:23
(thumbs up)
ejoe132
@ejoe132
Jun 17 2017 19:39
@itsn0tm3 seems to be running fine on master branch, gonna try the dev again and do as you suggested if it doesnt connect up
PieterGit
@PieterGit
Jun 17 2017 19:45
@itsn0tm3 @scottleibrand : i'm working on refactoring the init pump stuff. I don't the problem you experience is caused by oref0_init_comms.py. I found three cases in which the init/mmtune fails:
  • it seems possible that there can be more than one pump loop proccess running, and that messes the mmtune (i think you experienced this one)
  • there is a bug in the maybe_mmtune that on dev currently mmtunes in 80% instead of 20% of the cases, it seems the pump communication doesn't like that (this is fixed in this PR https://github.com/openaps/oref0/pull/470/commits/eea59a11d4bffb78630e31a8a4d176f5dc2dca7e )
  • there are some cases in which mmtune is not called when it should
@itsn0tm3 : what type of pump do you have WW or US?
ejoe132
@ejoe132
Jun 17 2017 19:59
@PieterGit so getting smb on board, in preferences shouldnt remaingcarbcap be greater than zero for the system to be able to microbolus?
davidklee
@davidklee
Jun 17 2017 20:38
Hi. Can someone help me with what the issue is here? Pump shows bolus for 86 grams of carbs at 12:24p and another 45 grams at 2:35p. These were done while not looping (we don't have the loop set up offline yet). After getting home at 4:00p, loop picks back up and shows there are still 124 carbs on board. here is what seems like the relevant part of the log: {"carbs":131,"boluses":11.8,"mealCOB":124}
{"iob":1.974,"activity":0.0395,"bolussnooze":0.001,"basaliob":0.249,"netbasalinsulin":0.25,"hightempinsulin":0.25,"time":"2017-06-17T20:06:40.000Z"}
{"delta":7,"glucose":220,"short_avgdelta":5.72,"long_avgdelta":4}
{"duration":20,"rate":1.75,"temp":"absolute"}
target_bg unchanged: 120
Carb Impact: 12.9 mg/dL per 5m; CI Duration: 6.6 hours
Accel. Carb Impact: 10 mg/dL per 5m; ACI Duration: 8.5 hours
determine-basal://text/shell/enact/suggested.json
reporting enact/suggested.json.
Zakm0n
@Zakm0n
Jun 17 2017 20:59
@Kdisimone turns out that windows 10 was the problem. I tried installing from a different windows 10 machine and got the same problem, installed Ubuntu on my spare hard drive and it installed first try like magic
Might be wise to advise other windows 10 users to run Ubuntu from
A flash drive if they have trouble flashing
katie disimone
@Kdisimone
Jun 17 2017 21:07
@Zakm0n 👍🏻👍🏻
Martin Haeberli
@mhaeberli
Jun 17 2017 21:28
recent lessons w/G5; will cross-post to intend-to-bolus, cgm-remote-monitor and Loop: - after much hassle, identified the iPhone 6 in question had flaky Bluetooth. New iPhone 7 much better. BUT still had “lost signal”. Until: a) turned off and un-paired from Dexcom G4 Share receiver (and powered same off), and b) stopped Tile app on iPhone. Since then, rock solid. Before then, had intermittent / persistent “lost signal” at 3 feet unobstructed. Y(CGM)MV
Tim Street
@tim2000s
Jun 17 2017 21:41
To be fair @Zakm0n flashing from Linux is by far the easiest and most consistent.
PieterGit
@PieterGit
Jun 17 2017 21:57
@ejoe132 : no, SMB also works with remainingcarbcap==0, but it smb's more early if remainingcarbscap>0 afaict. Please ask @scottleibrand or @danamlewis for details, or see http://openaps.readthedocs.io/en/master/docs/walkthrough/phase-4/oref1-features.html?highlight=remainingCarbsCap#how-to-turn-on-smb-uam . It's used in this piece of the code https://github.com/openaps/oref0/blob/dev/lib/determine-basal/determine-basal.js#L300
Zakm0n
@Zakm0n
Jun 17 2017 22:11
ok, got nightscout set up, but I'm missing the part on how to get my medtronic sensor connected to it
Brian Bonner
@itsn0tm3
Jun 17 2017 22:20
@PieterGit @scottleibrand I'm using US, I'll try rebooting the board and just playing with mmtune (not the alias) and seeing what comes out of that.
PieterGit
@PieterGit
Jun 17 2017 22:21
@itsn0tm3 do you use SMB? You can also disable the pump loop in the cron before rebooting, to make sure the mmtune can access /dev/spidev5.1
i'll tag you when i have rewritten and tested the init logic, so that you can test these improvements
Brian Bonner
@itsn0tm3
Jun 17 2017 22:44
@PieterGit Nope, not on SMB yet. Behind on the latest news so I'm still on oref0
rsilvers129
@rsilvers129
Jun 17 2017 22:51
I will add this to my rig. Sounds like the system could use a new default alias that does killall -g openaps; reset_spi_serial.py; openaps mmtune
Scott Leibrand
@scottleibrand
Jun 17 2017 23:11
@rsilvers129 no reason it should be standalone: it belongs in the mmtune alias, either directly or via the init Python script, so it runs automatically
@davidklee did you have CGM data for whole time after the carb entries?
davidklee
@davidklee
Jun 17 2017 23:18
when this happened today, yes, although it doesn't all show up in nightscout now for the time it was offline (the bolus does). This happened 2 nights ago also, resulting in a low, but I didn't get the logs, and the bolus had happened during the 2 hour dexcom warm up. This is a new pump set up, that's we've only had online these few times.
Scott Leibrand
@scottleibrand
Jun 17 2017 23:23
AMA relies on CGM data to decay carbs. If that's missing it only decays them at the min_5m_carbimpact rate (in mg/dL/5m). If you have a CGM gap you'll need to set a high temp target. In dev you can safely raise min_5m_carbimpact as well.
davidklee
@davidklee
Jun 17 2017 23:27
So I should shut off AMA until all of our radio issues are resolved. That's the other issue we have. The rig and pump only communicate sometimes. When it's not looping, I get a -99. It'll loop, then not loop. I thought it might be the pump, so got it working with another pump, but it's the same issue. So, should I assume it's the radio on the explorer board? Is there anything to check before I order a new one?
davidklee
@davidklee
Jun 17 2017 23:35
does it not decay if it's not looping? CGM data has been good the last hour, but the rig has been off (because of the extra carbs it thought were on board) now back on it doesn't seem to have decayed much of dinner yet either.
katie disimone
@Kdisimone
Jun 17 2017 23:58
@scottleibrand someone else on Master branch today is having these mmtune errors…just fyi