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

31st
Oct 2017
LilDucky
@LilDucky
Oct 31 2017 00:02
Thanks
jrkf
@jrkf
Oct 31 2017 00:45
KeyError: 'mmtune'
From pump-loop.log
Do you know how to fix it?
What is mean?
Laura Whitted
@NAGW322
Oct 31 2017 01:28
I'm back to getting this when I check the logs:
.....................................................................................................................................................................................................................................................................................................................................................
Starting pump-loop at Mon Oct 30 21:09:02 EDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Listening for 30s silence before mmtuning: .................................................................................................................................................................................................
Starting pump-loop at Mon Oct 30 21:19:02 EDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Listening for 30s silence before mmtuning: ..................................
Starting pump-loop at Mon Oct 30 21:22:03 EDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Listening for 30s silence before mmtuning: ........................................................................................................
Any advice on what to do?
David
@davidkeddydb
Oct 31 2017 01:30
what type of rig are you running? Explorer board and edison or something else?
Laura Whitted
@NAGW322
Oct 31 2017 01:30
explorer board and edison, yes
David
@davidkeddydb
Oct 31 2017 01:31
when you did the set up script, did you say yes to Edison? Is your pump battery fully charged?
Laura Whitted
@NAGW322
Oct 31 2017 01:32
I don't remember answering a y/n question re Edison. Pump battery is fully charged, as is rig and uploader
David
@davidkeddydb
Oct 31 2017 01:34
ok, when in doubt, just rerun the set up script. Start from scratch and there is a question about if you are using an explorer board. I can't remember if it now defaults to yes or not. cd && ~/src/oref0/bin/oref0-setup.sh to rerun set up
Laura Whitted
@NAGW322
Oct 31 2017 01:36
Dumb question, but I have a ton of cheap-o Sam's Club batteries. I've noticed they start out with less life than energizer. Is there any chance it's junk batteries??
David
@davidkeddydb
Oct 31 2017 01:37
Could be if you are less than full battery charge showing. If it is full though, unlikely
Laura Whitted
@NAGW322
Oct 31 2017 01:38
shows full on the pump screen. there's nothing showing up online in the pump pill
David
@davidkeddydb
Oct 31 2017 01:39
I would run the set up again and double check your pump serial number also when you run it
Laura Whitted
@NAGW322
Oct 31 2017 01:39
will do
David
@davidkeddydb
Oct 31 2017 01:39
Let us know when you are looping
Ebgineer
@Ebgineer
Oct 31 2017 01:45
anyone know if the medtronic 722 uses the same insets and reservoirs as their current pumps?
Laura Whitted
@NAGW322
Oct 31 2017 01:46
@davidkeddydb i'm using g5 with xdrip+, so I answer xdrip on the cgm type, correct?
David
@davidkeddydb
Oct 31 2017 01:48
@NAGW322 Yes
alimhassam
@alimhassam
Oct 31 2017 01:48
@Ebgineer yes same reservoirs
Ebgineer
@Ebgineer
Oct 31 2017 01:50
cool, thanks. I am considering picking up a used 722. My insurance should cover supplies for the new pumps.
I've built and loaded my edison/explorer rig but need a compatible pump
LilDucky
@LilDucky
Oct 31 2017 02:01
@Ebgineer to clarify, the new 600 series Medtronic pumps use the same reservoirs and insets as the older 500/700 series pumps.
Ebgineer
@Ebgineer
Oct 31 2017 02:25
Since I'm using an Animus Ping and they are being discontinued, I'll be able to switch. Probably to a 670G. I'm not keen on their proprietary closed loop with Guardian G3. I'd rather run Dexcom G4 in my loop in direct cabling/offline configuration.
Ebgineer
@Ebgineer
Oct 31 2017 02:30
If the 640 protocol hack applies to the 670 or the 670 is also eventually figured out, I'd have either 722 or 670G available to loop at that point.
Laura Whitted
@NAGW322
Oct 31 2017 02:40
@davidkeddydb I'm still getting the same thing...
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Listening for 30s silence before mmtuning: ................................
Starting pump-loop at Mon Oct 30 22:01:03 EDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Listening for 30s silence before mmtuning: .......................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
Starting pump-loop at Mon Oct 30 22:32:02 EDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
mmeowlink.exceptions.CommsException: Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?
Listening for 30s silence before mmtuning: ..
alimhassam
@alimhassam
Oct 31 2017 02:41
@NAGW322 is this an intermittent issue, did it ever successfully complete a loop, or does that always happen from the beginnig?
Laura Whitted
@NAGW322
Oct 31 2017 02:41
it is intermittent
alimhassam
@alimhassam
Oct 31 2017 02:42
and do you have a us pump or a non us pump?
Laura Whitted
@NAGW322
Oct 31 2017 02:42
us
alimhassam
@alimhassam
Oct 31 2017 02:43
ok, i'm asking because i've been having similar issues.
Laura Whitted
@NAGW322
Oct 31 2017 02:44
I had this error several times before. I'm not even sure what corrected it previously. It's been doing this for 2 days now. :/
alimhassam
@alimhassam
Oct 31 2017 02:44
i'm trying to investigate the issue on my side i'm not sure it's the same issue, but it might be.
i'm in a very similar boat, it works for 1h or 2 and then it stops, the longest i got it working was for around 5h yesterday.
Laura Whitted
@NAGW322
Oct 31 2017 02:46
that sounds similar to what mine was doing for 2 days before it just stopped looping completely.
before that I had it working for approx a week without any real issues.
alimhassam
@alimhassam
Oct 31 2017 02:47
yeah i'm pretty sure there is some incorrect interaction somewhere which is causing this
i'll share my improvements if they work, i thought it was due to the fact that i have a non US pump, but it might be something else.
otherwise the only think that sometimes fixes it at least temporarly is to run ~/myopenaps/oref0-runagain.sh and restart but i think you just did that right?
Tyler Duncan
@yduncan
Oct 31 2017 02:50
I was pretty disgruntled with this whole thing for a long time. I was frustrated with the docs and for some reason couldn't figure it out. I've still got lots that doesn't work, but watching the logs just now, I'm looping. Its set a temp basal for me and I'm pretty excited. Thanks for the help everyone.
Dana Lewis
@danamlewis
Oct 31 2017 02:53
Congrats on persevering @yduncan! If you have any feedback to improve the docs where you found them challenging, would love to have you PR them
@alimhassam what error message do you get? Rerunning setup is a bandaid.
alimhassam
@alimhassam
Oct 31 2017 02:54
similar to @NAGW322 lots of intermittent radio comm issues
Laura Whitted
@NAGW322
Oct 31 2017 02:54
@alimhassam , yes, I have run setup again.
it didn't change a thing.
Dana Lewis
@danamlewis
Oct 31 2017 02:55
@nagw322 working intermittently, or not at all now? Should self resolve but if not, you can search the docs for ccprog and reflash ccprog
Also, make sure your edison is firmly seating on the explorer board, that can also cause errors if it comes loose
Laura Whitted
@NAGW322
Oct 31 2017 02:56
It was working intermittently, but hasn't at all for the past two days. :'( The reflashing thing intimidates me. :/ I've also tried to install IFTTT app on my cell. Could I have messed something up doing that?
it's not very hard, i did that yesterday
but it didn't resolve my issues.
might be worth a try
Laura Whitted
@NAGW322
Oct 31 2017 02:59
I've also been having more difficulty getting the 3d printed case on/off when I need to open it. And the rig feels extremely hot.
alimhassam
@alimhassam
Oct 31 2017 02:59
i don't think IFTTT on your cell would have messed that up
Laura Whitted
@NAGW322
Oct 31 2017 03:00
even though I made changes in config variables on heroku?
alimhassam
@alimhassam
Oct 31 2017 03:01
i think it would be quite unlikely
are you on the dev branch or master?
Laura Whitted
@NAGW322
Oct 31 2017 03:02
that's good to know. eh, I don't know about the branch
Laura Whitted
@NAGW322
Oct 31 2017 03:10
Aaaand now I've lost cgm data. :-1:
@alimhassam can you tell me how to check if i'm on the dev branch or master? What would those mean?
alimhassam
@alimhassam
Oct 31 2017 03:11
very experimental, and probably just applies to me, but i have a couple of changes i did to try to either fix/reduce/investaget the issue, but i did that off the dev branch.
this section: Optional: To get on “dev” branch to test even more recently added new stuff
here is my branch, only small changes i did today, not sure if anything will help, or if it's just a coincidence that it restarted working better for me with these changes, so use at your own risk: https://github.com/alimhassam/oref0/tree/debugging-radio-issues
Laura Whitted
@NAGW322
Oct 31 2017 03:15
thanks. :) I don't know enough to feel comfortable changing it tonight, but I'll try to read & learn soon!
alimhassam
@alimhassam
Oct 31 2017 03:16
ok, i'll see how long my loop will last this time, and i'll keep the group updated
Laura Whitted
@NAGW322
Oct 31 2017 03:17
well - @danamlewis @alimhassam, I decided to power off the rig and let it cool down before I powered on. I'm back to showing as looping, but still no cgm data in the past 40 minutes.
my son wears the rig in a spibelt style running belt. is there a way to keep the rig from overheating? maybe i'm just not used to it, but it feels very hot to the touch most of the time.
alimhassam
@alimhassam
Oct 31 2017 03:19
:+1: that's good news!
i'm very new to this also, so i can't help for this question...
Laura Whitted
@NAGW322
Oct 31 2017 03:21
any advice on here for getting cgm data back? I haven't gotten a reading in 44 minutes. :worried: Or does loop continue working using past cgm data for some time?
alimhassam
@alimhassam
Oct 31 2017 03:21
do you have the cgm data anywhere?
is it in ns but not available on the rig?
is it in xdrip?
Laura Whitted
@NAGW322
Oct 31 2017 03:23
no - signal missed 46 minutes ago in xdrip. there's barely 2 feet between the uploader and my son's arm with the sensor.
alimhassam
@alimhassam
Oct 31 2017 03:24
hmm, that's a different issue, and probably won't be able to retrieve these 46 min of data
you're using the g5? how is the battery?
how old is the sensor? might be better to ask on the xdrip group
Laura Whitted
@NAGW322
Oct 31 2017 03:26
battery shows 216 - ok
sensor is 9d14h old
it has been extremely accurate and no missed readings at all until now.
alimhassam
@alimhassam
Oct 31 2017 03:28
i don't really know :( i know that the g5 the range is quite small if i forget my phone a bit too far i miss readings.
(compared to the g4)
mariano42
@mariano42
Oct 31 2017 03:29
I am getting that "unknown" message in the OpenAps pill on NS. Docs mentioned a dev branch from August with a fix, does anyone know if that is in master now?
Laura Whitted
@NAGW322
Oct 31 2017 03:49
is there a certain number of bg readings needed for OpenAPS to begin working?
WooHoo!! CGM and OpenAPS working again!! :sparkles: Feels like Christmas! <3
alimhassam
@alimhassam
Oct 31 2017 05:28
Hi All, i think mmtune doesn't work on my pump maybe, or something in the mmtune process is breaking my loop. Does anyone have expertise to help me confirm that?
it's only ever giving rssi of -99. The communication works for a while but seems to fail at the the first run of openaps mmtune. After nothing works anymore.
does that mean it's always using the same frequency, in which case can i remove that part of the loop?
alimhassam
@alimhassam
Oct 31 2017 05:31
yeah, did that, doesn't fix it unfortunately :(
Martin Haeberli
@mhaeberli
Oct 31 2017 05:31
in particular, cd ~/myopenaps && sudo service cron stop && killall -g openaps ; killall -g oref0-pump-loop; openaps mmtune && sudo service cron start
alimhassam
@alimhassam
Oct 31 2017 05:31
yeah, i did that.
it just gives -99
Martin Haeberli
@mhaeberli
Oct 31 2017 05:31
if you never get a number different from -99 then it’s basically not working
alimhassam
@alimhassam
Oct 31 2017 05:31
but the loop was still working
for a couple of hours
Martin Haeberli
@mhaeberli
Oct 31 2017 05:32
you should see something like mmtune: "916.588", 5, -71 waiting for 22 second silence before continuing
alimhassam
@alimhassam
Oct 31 2017 05:32
until the first time i see mmtune: in the logs
Martin Haeberli
@mhaeberli
Oct 31 2017 05:32
after a successful mmtune
do you have another rig?
alimhassam
@alimhassam
Oct 31 2017 05:32
right, but in my case, it gets to that part, everything just fails
no i don't :(
and i have a non us pump
so i think that might be related
just to confirm, for working setups, do you get any record with numbers different than -99 in cat ~/myopenaps/monitor/mmtune.json
and also, is usedDefault set to true of false for you?
(in that same file)
Martin Haeberli
@mhaeberli
Oct 31 2017 05:42
if you never get anything but -99 when manually mmtune ing , then I’m skeptical that it worked at all
as I recall, related to youlr question re cat - I think the answer is ‘yes'
I also recall there is a configuration for non-US pumps.
also worth asking - for your Explorer board, I hope you didn’t trim the antenna on the board. because trimming the antenna only helps for US pumps, and hurts for non-us pumps
alimhassam
@alimhassam
Oct 31 2017 05:44
i wasn't aware that trimming the antenna was a thing
Martin Haeberli
@mhaeberli
Oct 31 2017 05:44
it’s DEFINITELY worth re-flashing the Explorer board radio with the RIGHT code using the instructions in troubleshooting - search for cctool
or ccprog - hang on a second
alimhassam
@alimhassam
Oct 31 2017 05:44
i did that yesterday, but still got the error today
after looping for ~1h
the looping break always seems to come with a mmtune in the log, but that might be a concequence, i'm not sure yet what's the root cause.
Martin Haeberli
@mhaeberli
Oct 31 2017 05:45
you might want to check that you are using the right .hex file for programming - I suspect, but don’t know for sure, that you should NOT be using the US version mentioned here: mmtune: "916.588", 5, -71 waiting for 22 second silence before continuing
sorry hang on
wget https://github.com/EnhancedRadioDevices/subg_rfspy/releases/download/v0.8-explorer/spi1_alt2_EDISON_EXPLORER_US_STDLOC.hex
alimhassam
@alimhassam
Oct 31 2017 05:47
ah
i should NOT?
that's the one i used
Martin Haeberli
@mhaeberli
Oct 31 2017 05:47
I suspect maybe you should be using an international version … hang on
alimhassam
@alimhassam
Oct 31 2017 05:48
interesting suggestion
maybe i need to compile the international version
Martin Haeberli
@mhaeberli
Oct 31 2017 05:49
may be - likely it will be called WW vs US
i suspect you can find a pre-built
but I don’t know where
alimhassam
@alimhassam
Oct 31 2017 05:50
ok, it's open source so in theory it should be easy to build, i can't find it pre-built either
Martin Haeberli
@mhaeberli
Oct 31 2017 05:50
wishing you luck and skill
perhaps someone else on this gitter channel can advise as well
alimhassam
@alimhassam
Oct 31 2017 05:51
thanks for taking the time, your suggestion was very welcome.
Martin Haeberli
@mhaeberli
Oct 31 2017 05:51
you’re welcome! I would suggest if you can that you order a backup Edison and Explorer - it gives you security plus another configuration to test if you ever need it.
alimhassam
@alimhassam
Oct 31 2017 05:51
yeah, seems tempting to do
but i think they are sold out
Martin Haeberli
@mhaeberli
Oct 31 2017 05:54
explorer boards look to be in stock.
alimhassam
@alimhassam
Oct 31 2017 05:55
where?
Martin Haeberli
@mhaeberli
Oct 31 2017 05:55
at hamshield
alimhassam
@alimhassam
Oct 31 2017 05:55
ah, i thought it was soldout, maybe just the edison is soldout?
preflashed edison from hamshield IS sold out
amazon and, for example, spark fun https://www.sparkfun.com/products/13024 have at least systems that include an edison board. yes - they are getting harder to buy.
alimhassam
@alimhassam
Oct 31 2017 05:59
right, been through that, i guess once i'm setup and have a stable loop, i'll look into redundancy
Martin Haeberli
@mhaeberli
Oct 31 2017 05:59
since you have a WorldWide - non US pump, you should expect mmtune frequency results in the 868.xxx MHz range, not the 916.xxx MHz range
Ebgineer
@Ebgineer
Oct 31 2017 05:59
eventually we'll have to think about porting to another platform in place of edison
Martin Haeberli
@mhaeberli
Oct 31 2017 06:00
(and in running oref0-setup, you’ll choose WW not US for pump type…)
alimhassam
@alimhassam
Oct 31 2017 06:00
any ideas what the next platform will be?
Martin Haeberli
@mhaeberli
Oct 31 2017 06:01
can you confirm you chose WW not US? It may be that you DONT have to re flash Explorer for WW; I don’t know - perhaps one of our WW users can comment.
alimhassam
@alimhassam
Oct 31 2017 06:01
yeah i did choose WW
Martin Haeberli
@mhaeberli
Oct 31 2017 06:01
i hear some discussions of RPi0 with a custom ‘hat'
but i don’t know for sure
alimhassam
@alimhassam
Oct 31 2017 06:02
and it does work for a bit, but once mmtune kicks in, it stops working, very strange, and this is what i see in my mmtunes.json file
{
  "scanDetails": [
    [
      "868.150", 
      0, 
      -99
    ], 
    [
      "868.174", 
      0, 
      -99
    ], 
    [
      "868.198", 
      0, 
      -99
    ], 
    [
      "868.222", 
      0, 
      -99
    ], 
    [
      "868.246", 
      0, 
      -99
    ], 
    [
      "868.270", 
      0, 
      -99
    ], 
    [
      "868.294", 
      0, 
      -99
    ], 
    [
      "868.318", 
      0, 
      -99
    ], 
    [
      "868.342", 
      0, 
      -99
    ], 
    [
      "868.366", 
      0, 
      -99
    ], 
    [
      "868.390", 
      0, 
      -99
    ], 
    [
      "868.414", 
      0, 
      -99
    ], 
    [
      "868.438", 
      0, 
      -99
    ], 
    [
      "868.462", 
      0, 
      -99
    ], 
    [
      "868.486", 
      0, 
      -99
    ], 
    [
      "868.510", 
      0, 
      -99
    ], 
    [
      "868.534", 
      0, 
      -99
    ], 
    [
      "868.558", 
      0, 
      -99
    ], 
    [
      "868.582", 
      0, 
      -99
    ], 
    [
      "868.606", 
      0, 
      -99
    ], 
    [
      "868.630", 
      0, 
      -99
    ], 
    [
      "868.654", 
      0, 
      -99
    ], 
    [
      "868.678", 
      0, 
      -99
    ], 
    [
      "868.702", 
      0, 
      -99
    ], 
    [
      "868.726", 
      0, 
      -99
    ]
  ], 
  "setFreq": 868.328, 
  "usedDefault": true
}
Martin Haeberli
@mhaeberli
Oct 31 2017 06:02
can you send me an excerpt from your cat mmtune ...
excellent
so when you do a standalone mmtune, with your pump and your Edison/Explorer very near each other (maybe only 5 - 10 centimeters) - what does your terminal output look like?
usedDefault means the tune failed ...
alimhassam
@alimhassam
Oct 31 2017 06:04
let me try
i'm scared of breaking my current working loop, but i'll try for investigation, i need to run this right? cd ~/myopenaps && sudo service cron stop && killall -g openaps ; killall -g oref0-pump-loop; openaps mmtune && sudo service cron start
Martin Haeberli
@mhaeberli
Oct 31 2017 06:06
don’t do the cron start at the end yet
i’d actually recommend breaking this down
you can always reboot at the end, or do the cron start
alimhassam
@alimhassam
Oct 31 2017 06:07
when it's broken reboot doesn't fix it, i need to do an oref0-runagain to fix the issue.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:08
so, for example: cd ~/myopenaps , then sudo service cron stop
@alimhassam very strange
about oref0-runagain to fix the issue
alimhassam
@alimhassam
Oct 31 2017 06:08
anyway, too late a ran it :)
here is the output:
# cd ~/myopenaps && sudo service cron stop && killall -g openaps ; killall -g oref0-pump-loop; openaps mmtune 
+ echo

+ echo The CC111x is located at /dev/spidev5.1
The CC111x is located at /dev/spidev5.1
+ cd /root/src/subg_rfspy/tools
+ case "$2" in
+ ./reset.py /dev/spidev5.1
RileyLink OK
Version: subg_rfspy 0.8
OK
+ sleep 2
+ exit 0
+ echo

+ echo The CC111x is located at /dev/spidev5.1
The CC111x is located at /dev/spidev5.1
+ cd /root/src/subg_rfspy/tools
+ case "$2" in
+ ./change_setting.py /dev/spidev5.1 0x06 0x00
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x0C 0x59
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x0D 0x66
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x0E 0x33
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x0F 0x62
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x10 0x1A
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x11 0x13
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x09 0x24
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x0A 0x2E
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ ./change_setting.py /dev/spidev5.1 0x0B 0x38
RileyLink OK
Version: subg_rfspy 0.8
SUCCESS
+ sleep 0.5
+ exit 0
mmtune: pump://JSON/mmtune/monitor/mmtune.json
waking...
No response...
No response...
No response...
No response...
scanning...
868.150, 5, rssi:-99.0
868.174, 5, rssi:-99.0
868.198, 5, rssi:-99.0
868.222, 5, rssi:-99.0
868.246, 5, rssi:-99.0
868.270, 5, rssi:-99.0
868.294, 5, rssi:-99.0
868.318, 5, rssi:-99.0
868.342, 5, rssi:-99.0
868.366, 5, rssi:-99.0
868.390, 5, rssi:-99.0
868.414, 5, rssi:-99.0
868.438, 5, rssi:-99.0
868.462, 5, rssi:-99.0
868.486, 5, rssi:-99.0
868.510, 5, rssi:-99.0
868.534, 5, rssi:-99.0
868.558, 5, rssi:-99.0
868.582, 5, rssi:-99.0
868.606, 5, rssi:-99.0
868.630, 5, rssi:-99.0
868.654, 5, rssi:-99.0
868.678, 5, rssi:-99.0
868.702, 5, rssi:-99.0
868.726, 5, rssi:-99.0
reporting monitor/mmtune.json
there is a script which has a set -x which makes the output verbose
Martin Haeberli
@mhaeberli
Oct 31 2017 06:10
looks like a failed mmtune here also
alimhassam
@alimhassam
Oct 31 2017 06:10
right
Martin Haeberli
@mhaeberli
Oct 31 2017 06:10
looks more verbose than I was used to
so - i remain surprised that it looked like it was working for even an hour or so
alimhassam
@alimhassam
Oct 31 2017 06:10
it failed, but it was looping sucessfully just couple of seconds earlier
Martin Haeberli
@mhaeberli
Oct 31 2017 06:11
can you show me the output of the l command
that’s a small L
alimhassam
@alimhassam
Oct 31 2017 06:11
thats the tail of it
Starting pump-loop at Tue Oct 31 02:03:03 EDT 2017:
Waiting up to 4 minutes for new BG: .......glucose.json newer than pump_loop_completed
Radio ok. Listening: .No interfering pump comms detected from other rigs (this is a good thing!)
Profile less than 60m old. glucose.json newer than temp_basal.json. Temp refreshed
Temp duration 27m or more
glucose.json newer than pumphistory. : Refreshed pumphistoryRemoved duplicate bolus/carb entries:4
 and meal.json
{"carbs":45,"boluses":7.1,"mealCOB":23,"currentDeviation":7.73,"maxDeviation":13.27,"minDeviationSlope":-2.442}
{"iob":3.101,"activity":0.0303,"bolussnooze":0.682,"basaliob":0.155,"netbasalinsulin":0.25,"hightempinsulin":0.4,"microBolusInsulin":0.1,"microBolusIOB":0.083,"time":"2017-10-31T06:04:54.000Z","lastBolusTime":1509426236000}
{"delta":1,"glucose":144,"short_avgdelta":0.67,"long_avgdelta":3.52}
{"duration":28,"rate":1,"temp":"absolute"}
Autosens adjusting basal from 0.9 to 1; target_bg from 135 to 129; sens from 45 to 41.3 (autosens ratio 1.09)
Carb Impact: 6.9 mg/dL per 5m; CI Duration: 3.2 hours; remaining 4h+ CI: 0 mg/dL per 5m
Accel. Carb Impact: 10 mg/dL per 5m; ACI Duration: 2.2 hours
UAM Impact: 6.9 mg/dL per 5m; UAM Duration: 0.3 hours
minPredBG: 66 minIOBPredBG: 47 minCOBPredBG: 129 minUAMPredBG: 39 avgPredBG: 66 COB: 23 carbs: 45
BG projected to remain above 104 for 240 minutes
bgUndershoot: -7 zeroTempDuration: 240 zeroTempEffect: 150 carbsReq: -51
determine-basal://text/shell/enact/suggested.json
reporting enact/suggested.json
{"temp":"absolute","bg":144,"tick":"+1","eventualBG":136,"snoozeBG":100,"insulinReq":0,"reservoir":null,"deliverAt":"2017-10-31T06:05:35.536Z","minPredBG":999,"predBGs":{"IOB":[144,144,143,141,139,135,131,127,123,118,113,108,102,96,91,87,82,78,74,70,67,64,61,58,56,54,52,50,49,48,47],"aCOB":[144,147,150,152,153,154,154,154,154,155,154,154,154,154,153,153,152,151,151,150,149,147,146,145,143,142,140,138,137,136,135],"COB":[144,144,144,144,143,141,140,138,137,136,134,133,133,132,131,130,130,129,129,129,129,129,129,129,129,129,130,130,130,131,132,133,133,134,135,136],"UAM":[144,142,138,131,123,116,108,101,95,88,82,76,70,65,60,55,51,46,42,39]},"COB":23,"IOB":3.101,"reason":"COB: 23, Dev: 42, BGI: -6.26, ISF: 2.3, Target: 7.2, minPredBG 3.7, IOBpredBG 2.6, COBpredBG 7.6, UAMpredBG 2.2; 7.6-3.7 in range: no temp required, temp 1 ~ req 1U/hr. "}
enact/enacted.json: {"insulinReq":0,"received":true,"bg":143,"reservoir":null,"temp":"absolute","snoozeBG":91,"rate":1,"predBGs":{"COB":[143,142,141,140,138,135,133,130,128,126,124,123,121,120,119,118,118,117,117,117,117,117,118,118,119,120,122,123,124,126,128,130,132,135,137,140,142,144,146,148,150,152],"aCOB":[143,147,150,152,153,154,154,155,155,155,155,155,155,155,155,155,154,154,153,153,152,152,151,150,149,148,147,146,144,143,142,142,141],"UAM":[143,140,133,127,119,112,105,97,90,84,77,71,65,60,55,50,45,41,39],"IOB":[143,142,139,136,133,128,123,118,113,107,102,96,90,85,79,74,70,65,61,57,54,50,47,45,42,40,39]},"minPredBG":999,"IOB":3.201,"reason":"COB: 25, Dev: 30, BGI: -6.03, ISF: 2.3, Target: 7.2, minPredBG 3.6, IOBpredBG 2.2, COBpredBG 8.4, UAMpredBG 2.2; 8.4-3.6 in range: no temp required; setting current basal of 1 as temp. ","COB":25,"eventualBG":152,"timestamp":"2017-10-31T02:02:15.052109","duration":30,"tick":-1,"deliverAt":"2017-10-31T06:02:11.826Z"}
Settings less than 10m old
Edison on battery: 52%. Pumphistory-24 < 90m old
Completed pump-loop at Tue Oct 31 02:05:35 EDT 2017


Starting pump-loop at Tue Oct 31 02:06:03 EDT 2017:
Waiting up to 4 minutes for new BG: .......
until i stopped it
to run the mmtune
Martin Haeberli
@mhaeberli
Oct 31 2017 06:12
hard for me to reconcile loop working as shown in above log with mmtune failing
alimhassam
@alimhassam
Oct 31 2017 06:12
should i restart it?
Martin Haeberli
@mhaeberli
Oct 31 2017 06:13
try the sudo cron start and see if that works ...
alimhassam
@alimhassam
Oct 31 2017 06:13
indeed, very weird from my understanding as well
seems like it might be using the default frequency and it's successful?
maybe a bug in mmtune for WW pumps?
not sure.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:13
unlikely … -99 reflects no signal / very bad signal
maybe - I have no experience w WW pumps. @tmecklem ? @danamlewis @scottleibrand ?
alimhassam
@alimhassam
Oct 31 2017 06:14
started the cron
fingers crossed
usually after running mmtunes it fails
Martin Haeberli
@mhaeberli
Oct 31 2017 06:14
< crossed fingers
btw what branch are you using?
I hope 0.5.5 master just so we’re on the same page
alimhassam
@alimhassam
Oct 31 2017 06:15
dev
had the same problem on master
Martin Haeberli
@mhaeberli
Oct 31 2017 06:15
ok
alimhassam
@alimhassam
Oct 31 2017 06:15
so i switch to dev hoping it would fix it
now i'm adding debugging logs to it, for investigation next time it fails
ok i got the error now
not looping anymore.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:16
is there anyone near by that also has a working edison / openaps ?
:-( re me breaking your loop ...
alimhassam
@alimhassam
Oct 31 2017 06:16
Starting pump-loop at Tue Oct 31 02:06:03 EDT 2017:
Waiting up to 4 minutes for new BG: .......
Starting pump-loop at Tue Oct 31 02:14:02 EDT 2017:
Waiting up to 4 minutes for new BG: .glucose.json newer than pump_loop_completed
Radio ok. Listening: .No interfering pump comms detected from other rigs (this is a good thing!)
Profile less than 60m old. glucose.json newer than temp_basal.json. Temp refreshmmeowlink.exceptions.CommsException: No acknowledgement from pump on wakeup. Is it out of range or is the battery too low?
Error, retrying

Starting pump-loop at Tue Oct 31 02:15:19 EDT 2017:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
Radio ok. Listening: .No interfering pump comms detected from other rigs (this is a good thing!)
Profile less than 60m old. glucose.json newer than temp_basal.json. Temp refresh
that's what i have in my l
CommsException is the issue
Martin Haeberli
@mhaeberli
Oct 31 2017 06:16
well, normally in my experience, after an mmtune fails, it can take 20 - 30 minutes to re-establish pump comms
but that assumes working mmtune
alimhassam
@alimhassam
Oct 31 2017 06:17
ah
ok, lets see, i'll let it go for a while.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:18
i’m going to sign off but wish you well - hope someone else chimes in ...
alimhassam
@alimhassam
Oct 31 2017 06:18
thanks for the help!
i'll try the ww subg_rfspy
Tim Street
@tim2000s
Oct 31 2017 06:20
What happens if you reboot after the failure? I run a WW pump, and I can’t say I’ve seen this issue.
And when did it break?
Martin Haeberli
@mhaeberli
Oct 31 2017 06:20
@tim2000s cheering for you ...
he said he has to re-run oref0-setup to get it to work...
standalone mmtunes always fail.
alimhassam
@alimhassam
Oct 31 2017 06:21
yep
Martin Haeberli
@mhaeberli
Oct 31 2017 06:21
does he have to install a custom WW .hex in the Explorer CC111x, or can he use US
he has configured for WW via oref0-setup
alimhassam
@alimhassam
Oct 31 2017 06:21
@tim2000s can you check what you have in cat monitor/mmtune.json ?
do you use explorer board/edison?
if i reboot after the failure, i get the same issue.
Tim Street
@tim2000s
Oct 31 2017 06:22
I run explorer board Edison yes.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:27
which oref0 branch do you happen to be running?
Tim Street
@tim2000s
Oct 31 2017 06:27
I know mine all run okay as I was using them last night for range. My current frequency is not default.
alimhassam
@alimhassam
Oct 31 2017 06:27
hmm
Tim Street
@tim2000s
Oct 31 2017 06:27
But coming back. When did it break? Or has it never worked?
alimhassam
@alimhassam
Oct 31 2017 06:28
i'm new to looping, started couple of days ago, never worked very well, it works for up to a couple of hours, then fails
not sure what's triggering the failure, but it seems related to when mmtune is being run
but that could be a concequence, not the root cause, so i'm not sure.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:29
but i myself can’t explain alim’s ‘working’ mode, as it looks like standalone mmtunes ALWAYS fail
and per alim reboot never works to restart loop, he has to oref0-setup (or maybe oref0-runagain)
alimhassam
@alimhassam
Oct 31 2017 06:30
it reads my pump, i see entries in NS and everything.
bolus / carbs, my pump history work, it's setting temp basals etc
until it breaks.... :(
Martin Haeberli
@mhaeberli
Oct 31 2017 06:31
Alim - your log entries are consistent with that, BUT, hard to explain mmtune ...
Tim Street
@tim2000s
Oct 31 2017 06:31
I know this sounds dumb, but is the pump really a WW one? The serial number definitely ends WW?
alimhassam
@alimhassam
Oct 31 2017 06:31
usually in the middle of the night, or when i'm the furthest place from being able to fix it.
Martin Haeberli
@mhaeberli
Oct 31 2017 06:31
@tim2000s I was just thinking about that question
alimhassam
@alimhassam
Oct 31 2017 06:31
that's a good question, i'm wondering if all ww pumps are created equal
Martin Haeberli
@mhaeberli
Oct 31 2017 06:31
ok - sending good loop karma your way - have to sleep here...
alimhassam
@alimhassam
Oct 31 2017 06:31
it doesn't end with ww
it's a canadian pump
thanks Martin~
!
Martin Haeberli
@mhaeberli
Oct 31 2017 06:32
would suggest you try the US configuration ...
with a hand mmtune
and see what happens
alimhassam
@alimhassam
Oct 31 2017 06:32
i did a quick try, but maybe i should try again
i think it didn't work, i don't remember
i'll try again.
Tim Street
@tim2000s
Oct 31 2017 06:33
That’s your issue. Canadian pumps need the US radio code.
alimhassam
@alimhassam
Oct 31 2017 06:33
ah
are you sure?
Tim Street
@tim2000s
Oct 31 2017 06:33
About 80%
alimhassam
@alimhassam
Oct 31 2017 06:34
ok, let me try a new setup
can i modify oref0-runagain or do i need to start fresh?
Brian Bonner
@itsn0tm3
Oct 31 2017 06:36
meh, looks like Heroku is having issues
you probably could modify the runagain
change the ww to say --radio_locale='US'
(change WW to US)
Tim Street
@tim2000s
Oct 31 2017 06:43
Yes. Makes sense
alimhassam
@alimhassam
Oct 31 2017 06:54
this is what i get from open aps mmtune with a US Setup, is that better?
# openaps mmtune
mmtune: pump://JSON/mmtune/monitor/mmtune.json
waking...
packet = {'rssi': -67, 'data': bytearray(b'\xa72R\x89\x8d\t\x03722\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00E'), 'sequence': 1}
Woke up pump: {'rssi': -67, 'data': bytearray(b'\xa72R\x89\x8d\t\x03722\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00E'), 'sequence': 1}
scanning...
916.300, 5, rssi:-99.0
916.324, 5, rssi:-99.0
916.348, 5, rssi:-99.0
916.372, 5, rssi:-99.0
916.396, 5, rssi:-99.0
916.420, 5, rssi:-99.0
916.444, 5, rssi:-99.0
916.468, 5, rssi:-99.0
916.492, 5, rssi:-99.0
916.516, 5, rssi:-99.0
916.540, 0, rssi:-68.0
916.564, 0, rssi:-62.0
916.588, 0, rssi:-61.0
916.612, 0, rssi:-61.0
916.636, 0, rssi:-61.0
916.660, 0, rssi:-62.0
916.684, 0, rssi:-72.0
916.708, 5, rssi:-99.0
916.732, 5, rssi:-99.0
916.756, 5, rssi:-99.0
916.780, 5, rssi:-99.0
916.804, 5, rssi:-99.0
916.828, 5, rssi:-99.0
916.852, 5, rssi:-99.0
916.876, 5, rssi:-99.0
reporting monitor/mmtune.json
"916.612", 5, -61
alimhassam
@alimhassam
Oct 31 2017 07:00
Thanks a lot @tim2000s and @mhaeberli i think you're right, that might have bee the root problem, my loop is working now as a us pump, and mmtune works too, so chances are, the loop might be in better shape now.
I really appreciate the help
LilDucky
@LilDucky
Oct 31 2017 07:02
I can positively confirm that Canadian pumps are NA and not WW.
alimhassam
@alimhassam
Oct 31 2017 07:03
lol, i feel dumb now...
i'm not sure why i thought it was WW
Brian Bonner
@itsn0tm3
Oct 31 2017 07:16
Question: Since Heroku is having problems, I started up my old Azure instance since I hadn't shut it down yet, and it's up and running. Is there a problem having 2 NS instances updating the bridge?
(I can't shut down the Heroku instance because apparently the controls are broken and I'm not even sure if Heroku is updating the bridge)
not sure if NS itself has dedupe built in?
LilDucky
@LilDucky
Oct 31 2017 07:50
@alimhassam don’t feel too bad - we all make mistakes and that is how we learn
philipgo
@philipgo
Oct 31 2017 09:24
@alimhassam I am sure you are not the first one who misunderstood this! Maybe you can add this to the docs.
LilDucky
@LilDucky
Oct 31 2017 10:33
@scottleibrand here is an example of where I feel that the loop is not trying to achieve the target BG of 7.5
Starting supermicrobolus pump-loop at Tue Oct 31 21:31:02 AEDT 2017 with 1 second wait_for_silence:
MDT CGM configured; not waiting
Radio ok. Listening: .No interfering pump comms detected from other rigs (this is a good thing!)
Preflight OK.
Attempting to retrieve MDT CGM data from pump
MDT CGM data retrieved
MDT New cgm data reformatted
Profile less than 60m old and valid. Refresh  _date: 'a49f121e11' }
ed pumphistoryWarning: clock input Invalid Date is unzoned; please pass clock-zoned.json instead
 and meal.json
Checking pump clock: "2017-10-31T21:31:25+11:00" is within 1m of current time: Tue Oct 31 21:32:07 AEDT 2017
and that pumphistory is less than 1m old.  Temp refresh  _date: 'a49f121e11' }
ed
Autotune exists! Hoorah! You can use microbolus-related features.
{"carbs":40,"mealCOB":0,"currentDeviation":-4.67,"maxDeviation":4.14,"minDeviation":-1.8,"slopeFromMaxDeviation":-1.333,"slopeFromMinDeviation":0,"allDe                viations":[-2,-1,-1,2,3,4,0,-1],"lastCarbTime":1509435743000}
{"iob":0.109,"activity":0.0103,"basaliob":-1.141,"bolusiob":1.25,"netbasalinsulin":-2.45,"bolusinsulin":6.8,"time":"2017-10-31T10:31:28.000Z","iobWithZe                roTemp":{"iob":0.109,"activity":0.0103,"basaliob":-1.141,"bolusiob":1.25,"netbasalinsulin":-2.45,"bolusinsulin":6.8,"time":"2017-10-31T10:31:28.000Z"},"                lastBolusTime":1509438022000,"lastTemp":{"rate":1.85,"timestamp":"2017-10-31T21:28:39+11:00","started_at":"2017-10-31T10:28:39.000Z","date":150944571900                0,"duration":4.55}}
{"delta":-6,"glucose":172,"short_avgdelta":-5.22,"long_avgdelta":-3.7}
Autosens ratio: 1; Basal unchanged: 1.85; ISF unchanged: 36; CR: 8.167
currenttemp: { duration: 28, rate: 1.85, temp: 'absolute' } lastTempAge: 4 m tempModulus: 2 m
Last carbs 170 minutes ago; remainingCATime: 7.3 hours; 100% carbs absorbed
Carb Impact: -4.1 mg/dL per 5m; CI Duration: 0 hours; remaining CI (~2h peak): 0 mg/dL per 5m
UAM Impact: -4.1 mg/dL per 5m; UAM Duration: 0 hours
minPredBG: 163 minIOBPredBG: 137 minZTGuardBG: 163 avgPredBG: 163 COB: 0 / 40
BG projected to remain above 135 for 240 minutes
naive_eventualBG: 168 bgUndershoot: -80.5 zeroTempDuration: 240 zeroTempEffect: 263 carbsReq: -78
Checking deliverAt: 2017-10-31T10:32:16.447Z is within 1m of current time: Tue Oct 31 21:32:16 AEDT 2017
and that smb-suggested.json is less than 1m old
enact/smb-suggested.json: {"temp":"absolute","bg":172,"tick":-6,"eventualBG":157,"insulinReq":0,"reservoir":"290.55","deliverAt":"2017-10-31T10:32:16.44                7Z","predBGs":{"IOB":[172,166,161,157,153,149,146,143,141,140,138,138,137,137,137,137,137,137,137,137,138,138,138,139,139,140,140,141,141,142,142,142,14                3,143,143,144,144,144,144,144,145],"ZT":[172,170,168,167,166,165,164,164,163,163,164,164]},"COB":0,"IOB":0.109,"reason":"COB: 0, Dev: -11, BGI: -1.85, I                SF: 2.0, Target: 7.5, minPredBG 9.1, minGuardBG 7.6, IOBpredBG 8.1; Eventual BG 8.7 > 7.5 but Min. Delta -6.00 < Exp. Delta -2.8, temp 1.85 ~ req 1.85U/                hr. "}
No smb_enact needed. Temp refreshed: monitor/temp_basal.json: {"duration":28,"rate":1.85,"temp":"absolute"}
No bolus needed (yet). Settings less than 15 minutes old. Edison on battery: 50%. Pumphistory-24 < 30m old
Completed supermicrobolus pump-loop at Tue Oct 31 21:32:20 AEDT 2017


Starting supermicrobolus pump-loop at Tue Oct 31 21:33:02 AEDT 2017 with 4 second wait_for_silence:
MDT CGM configured; not waiting
Radio ok. Listening: .No interfering pump comms detected from other rigs (this is a good thing!)
Preflight OK.
Attempting to retrieve MDT CGM data from pump
MDT CGM data retrieved
No new MDT CGM data to reformat

Profile less than 60m old and valid. Refresh  _date: 'a49f121e11' }
ed pumphistory  _date: 'a49f121e11' }
 and meal.json
Checking pump clock: "2017-10-31T21:33:08+11:00" is within 1m of current time: Tue Oct 31 21:33:51 AEDT 2017
and that pumphistory is less than 1m old.  Temp refresh  _date: 'a49f121e11' }
ed
Autotune exists! Hoorah! You can use microbolus-related features.
{"carbs":40,"mealCOB":0,"currentDeviation":-2.67,"maxDeviation":4.14,"minDeviation":-1.8,"slopeFromMaxDeviation":-0.979,"slopeFromMinDeviation":0,"allDeviations":[-3,-2,-1,-1,2,3,4,0,-1],"lastCarbTime":1509435743000}
{"iob":0.089,"activity":0.0099,"basaliob":-1.122,"bolusiob":1.211,"netbasalinsulin":-2.4,"bolusinsulin":6.8,"time":"2017-10-31T10:33:12.000Z","iobWithZeroTemp":{"iob":0.089,"activity":0.0099,"basaliob":-1.122,"bolusiob":1.211,"netbasalinsulin":-2.4,"bolusinsulin":6.8,"time":"2017-10-31T10:33:12.000Z"},"lastBolusTime":1509438022000,"lastTemp":{"rate":1.85,"timestamp":"2017-10-31T21:28:39+11:00","started_at":"2017-10-31T10:28:39.000Z","date":1509445719000,"duration":6.28}}
{"delta":-6,"glucose":172,"short_avgdelta":-5.22,"long_avgdelta":-3.63}
Autosens ratio: 1; Basal unchanged: 1.85; ISF unchanged: 36; CR: 8.167
currenttemp: { duration: 26, rate: 1.85, temp: 'absolute' } lastTempAge: 5 m tempModulus: 1 m
Last carbs 172 minutes ago; remainingCATime: 7.3 hours; 100% carbs absorbed
Carb Impact: -4.2 mg/dL per 5m; CI Duration: 0 hours; remaining CI (~2h peak): 0 mg/dL per 5m
UAM Impact: -4.2 mg/dL per 5m; UAM Duration: 0 hours
minPredBG: 164 minIOBPredBG: 137 minZTGuardBG: 164 avgPredBG: 164 COB: 0 / 40
BG projected to remain above 135 for 240 minutes
naive_eventualBG: 169 bgUndershoot: -81.5 zeroTempDuration: 240 zeroTempEffect: 263 carbsReq: -78
Checking deliverAt: 2017-10-31T10:34:00.715Z is within 1m of current time: Tue Oct 31 21:34:00 AEDT 2017
and that smb-suggested.json is less than 1m old
enact/smb-suggested.json: {"temp":"absolute","bg":172,"tick":-6,"eventualBG":158,"insulinReq":0,"reservoir":"290.5","deliverAt":"2017-10-31T10:34:00.715Z","predBGs":{"IOB":[172,166,161,157,153,149,146,143,141,140,138,138,137,137,137,137,137,137,137,138,138,138,139,139,140,140,141,141,142,142,142,143,143,143,144,144,144,144,144,145],"ZT":[172,170,169,167,166,165,164,164,164,164,164,165,165]},"COB":0,"IOB":0.089,"reason":"COB: 0, Dev: -11, BGI: -1.78, ISF: 2.0, Target: 7.5, minPredBG 9.1, minGuardBG 7.6, IOBpredBG 8.1; Eventual BG 8.8 > 7.5 but Min. Delta -6.00 < Exp. Delta -2.7, temp 1.85 ~ req 1.85U/hr. "}
No smb_enact needed. Temp refreshed: monitor/temp_basal.json: {"duration":26,"rate":1.85,"temp":"absolute"}
No bolus needed (yet). Settings less than 15 minutes old. Edison on battery: 54%. pumphistory-24h refresh
image.png
{
        "max_iob": 6,
        "max_daily_safety_multiplier": 3,
        "current_basal_safety_multiplier": 4,
        "autosens_max": 1.2,
        "autosens_min": 0.7,
        "rewind_resets_autosens": true,
        "adv_target_adjustments": false,
        "exercise_mode": false,
        "unsuspend_if_no_temp": false,
        "enableSMB_with_COB": true,
        "enableSMB_with_temptarget": true,
        "enableSMB_with_bolus": true,
        "enableUAM": true,
        "curve": "rapid-acting"
}
Dana Lewis
@danamlewis
Oct 31 2017 13:37
@mariano42 yes that fix is in master now
@alimhassam glad you sorted it out!
Scott Leibrand
@scottleibrand
Oct 31 2017 13:40
@LilDucky that one is behaving as expected: setting a neutral temp because BG is falling faster than it needs to to get you back to target. I would scroll back to some time during the high BG period when you think it should've been doing more and see what the predictions and reason were then.
alimhassam
@alimhassam
Oct 31 2017 13:49
Hi all, are there recommendation, or ways of testing/validating Dia value for different types of insulin?
Is the value is specific to individual or are there recommended value per insulin type?
Scott Leibrand
@scottleibrand
Oct 31 2017 13:56
If you are on the exponential curves it will use a DIA of at least 5h. You can raise that to 6h or 7h if you see lows 4-6h after a large meal.
We've talked about a way to autotune DIA and insulin activity peak time but haven't tried implementing it get.
mariano42
@mariano42
Oct 31 2017 14:01
@danamlewis hmm I just updated last night and I am still getting 'unknown'.. do you recall if anyone else has reported that?
Velibor Maric
@vebaba
Oct 31 2017 14:05
@scottleibrand I can't figure out how predicted BG are so low. I entered 6.7IU for 42g of CH with (7 - 0.3 correction for lower BG)
NS-zerotemp.gif
Velibor Maric
@vebaba
Oct 31 2017 14:12
Is it prediction in case I was wrong about CH input? That makes sense then...
Dana Lewis
@danamlewis
Oct 31 2017 14:20
@mariano42 check for future treatments, then mlab size (compress if too big), etc.
alimhassam
@alimhassam
Oct 31 2017 14:32
I'm on the default
"curve": "bilinear"
But maybe I should change that. Trying to understand it and the implications.
Dana Lewis
@danamlewis
Oct 31 2017 14:39
I would try exponential curves.
We are going to move toward making that the default for everyone
ejoe132
@ejoe132
Oct 31 2017 14:41
Hopefully a quick question, rig was running good then quit. Checked this morning on the logs and first thing I saw was said it was out of memory, then it started all this
root@edisonone:~# [ 105.403174] link down if wlan0 may call cfg80211_disconnected. event : 16, reason=2 from 00:27:22:98:fc:26
[ 105.419495] wl_android_wifi_off in
[ 105.427591] wifi_platform_set_power = 0
[ 112.446282]
[ 112.446282] Dongle Host Driver, version 1.141.59 (r)
[ 112.446282] Compiled in /home/robin/src/edison-linux-helper/edison-bcm43340 on Apr 4 2017 at 22:42:52
[ 112.446457] wl_android_wifi_on in
[ 112.450398] wifi_platform_set_power = 1
[ 113.015813] F1 signature OK, socitype:0x1 chip:0xa94c rev:0x2 pkg:0x0
[ 113.023737] DHD: dongle ram size is set to 524288(orig 524288) at 0x0
[ 113.032600] dhdsdio_download_firmware: firmware path=/lib/firmware/fw_bcmdhd.bin, nvram path=/lib/firmware/bcmdhd.cal
[ 113.170678] sdioh_request_buffer: [1] doing memory copy buf=f4e9f000, len=203 and Im not quite sure what its trying to do here, it runs this as soon as I log in, not in the logs. Do a fresh install?
alimhassam
@alimhassam
Oct 31 2017 15:00
@scottleibrand
Oops
Scott Leibrand
@scottleibrand
Oct 31 2017 15:00
@vebaba can you use something other than pasteimg? It's all but unusable for zooming in on iOS (and completely covered in ads to boot). On a computer you can just drag and drop the image directly into Gitter.
@vebaba the top prediction takes into account COB. The bottom one is IOB only
Velibor Maric
@vebaba
Oct 31 2017 15:00
NS-zerotemp.gif
Sorry!
alimhassam
@alimhassam
Oct 31 2017 15:02
So if i change curve to exponential, i need to change dia to 5 in my pump and in ns profile, is that correct?
Scott Leibrand
@scottleibrand
Oct 31 2017 15:02
The COBpredBGs are going low because 30g is not enough to cover 4.5U of insulin. What's your CR?
Velibor Maric
@vebaba
Oct 31 2017 15:03
@scottleibrand 1IU covers 6g UH (autotune says 6.324g
Scott Leibrand
@scottleibrand
Oct 31 2017 15:04
@alimhassam you should, but if you don't it'll just override it to 5h.
Velibor Maric
@vebaba
Oct 31 2017 15:05
@scottleibrand At 7.2mmol, Temp basal kicked in with 1.3U/h which would do exactly what I've expected to do. I believe that 6.7mmol was too low for it in case I got wrong CH input
I had to go offline for half an hour so I gave 0.6 bolus and everything went fine. Back in normal again
ejoe132
@ejoe132
Oct 31 2017 15:09
Now not getting anything to work, activing like it erased all my files. Cant even reset new hostname to begin again. not sure where to go from here
Velibor Maric
@vebaba
Oct 31 2017 15:09
@scottleibrand @danamlewis OAPS rocks! ;)
jaylagorio
@jaylagorio
Oct 31 2017 15:11
@ejoe132 Sounds like you may need to reflash. I had to do that my first go around after a colossal BlueZ install failure.
ejoe132
@ejoe132
Oct 31 2017 15:18
@jaylagorio When Ive done that before I just did the setup script again and it normally solves it, but now when I try to pull the master branch it gives me: cd ~/src/oref0 && git checkout master && git pull && sudo npm install -g oref0
-bash: cd: /root/src/oref0: No such file or directory
Craig Brenner
@cbrenner
Oct 31 2017 15:20
I found this reproducible bug using 0.5.5. With BT tethering enabled on rig but iPhone hotspot shut off, anytime the rig moves out of range from Wi-Fi and attempts to connect to disabled iPhone hotspot, the rig locks up and ceases to work. Always requires a hard reset to reboot rig. We keep the iPhone hotspot on all the time but there are random times when iOS seems to take it upon itself to disable the hotspot. I have read that this may be the result of "low power mode" being activated in iOS. Unfortunately, I cannot provide any logs because the rig is locked up once this happens and there is no papertrail to pull up. Everything looks normal right up to the transition out of Wi-Fi then rig goes silent.
Brian Rabinovitz
@bdr1177
Oct 31 2017 15:21
I'm seeing a lot of zero temping after meal boluses, which is causing post-meal highs. I don't necessarily see anything wrong with the logic that OpenAPS is using in these cases, but the overall effect seems non-optimal. I think I understand the reasons for removing bolus snooze (with SMBs basal/bolus distinction doesn't make sense), but I'm really wondering if this configuration without bolus snooze works for us.
ejoe132
@ejoe132
Oct 31 2017 15:25
@cbrenner This could be what happened to me as it was when I got home, how did you get the rig unlocked up? I did a hard reset but nothing is working now, Try to do a fresh install and get : d ~/src/oref0 && git checkout master && git pull && sudo npm install -g oref0
-bash: cd: /root/src/oref0: No such file or directory
Craig Brenner
@cbrenner
Oct 31 2017 15:28
@ejoe132 Just did a hard reset and rig rebooted back into a working state. Perhaps you have a corrupted system or drive.
ejoe132
@ejoe132
Oct 31 2017 15:31
@cbrenner Thats what Im thinking, just not sure where to go from here
David
@davidkeddydb
Oct 31 2017 15:35
Anyone tell me what "ls" command can I run to find out if the dex is being seen by the explorer board? I just can't remember
David
@davidkeddydb
Oct 31 2017 15:42
never mind. I used dmesg and it got me there
Scott Leibrand
@scottleibrand
Oct 31 2017 15:45
@ejoe132 you need to re-clone oref0 since you (?) deleted it.
lsusb lists the connected USB devices.
@bdr1177 is it predicting a low when it zero temps, or just doing it for SMB mode?
Brian Rabinovitz
@bdr1177
Oct 31 2017 15:53
I think for SMB mode. Here's an example:
Capture3.JPG
There weren't actually any SMBs, but SMB is enabled for low temp target and COB and the COBpredBG was 124. Normal target is 130, but it still set a zero temp.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:01
Here's another example of what's going on right now after a lunch bolus. BG is 217 and rising. Predicted BG is significantly above target and yet the rig has set a 120m zero temp. Don't understand this at all.
Capture4.JPG
Scott Leibrand
@scottleibrand
Oct 31 2017 16:06
It always sets a zero temp with SMB if IOB alone is enough to make you go low. The question is not why it's setting a zero temp. It's why it's not SMBing as well.
Paul Andrel
@stavlor
Oct 31 2017 16:06
right cause IOB predBGs are heading quite low
Scott Leibrand
@scottleibrand
Oct 31 2017 16:07
And the answer looks like "because you need less than 0.2U of insulin, and half of that rounds down to zero"
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:09
@scottleibrand This may all be true and maybe our basals are just too low to use SMB. But we find SMB to be helpful in some situations. And bolus snooze seems like it used to function better for us than no bolus snooze. I understand that IOBpredBG is below target. But the reality is that there is significant COB. This is the exact situation that bolus snooze helped with.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:11
Yeah. I need to think about the best way to disable SMB automatically when it's not needed on tiny basals.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:12
Okay, we'll let me know if you think of something and you want me to test it. Do you want me to open an issue, or are you not at that stage yet?
Dana Lewis
@danamlewis
Oct 31 2017 16:13
not at that stage yet
btw - @bdr1177 how does autotune say your isf is? compared to what's on your pump? ditto for carb ratio?
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:17
@danamlewis That's actually a related issue that I'm having. Despite 4 days of high BGs and what seems like significant resistance, autotune keeps adjusting ISFs higher (from pump ISF of 190 to current autotune ISF of 235). The non-limited autotune ISF from yesterday was actually 345 (adjusted down to 271), which seems ridiculously high given insulin resistance.
On CR, autotune is recommending modest adjustments down for the last couple of days from 1:35 to 1:30.
Dana Lewis
@danamlewis
Oct 31 2017 16:18
what has autosens been saying? has it been picking up the resistance?
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:19
No, autosens has also generally been adjusting ISF higher over the last several days.
I am testing the a52-risk branch, in case that could have anything to do with all of this.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:24
If you've been identifying overnight resistance and autosens is detecting sensitivity, I'd like to see a morning autosens-loop.log run.
If OTOH the resistance is only at mealtimes, it would only show up in lowered CR, not necessarily in higher basals from autosens or lower ISF from autotune.
Overall, ISF is the least important of the three, despite everyone paying it the most attention.
Because it only comes into play for determining how aggressively to correct non-meal excursions. CR is what matters for most high BGs (caused by meals).
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:27
Yes, I think the resistance is meal related that that is why autotune and autosens keep adjusting up. That is why I initially brought up the concern about the meal-time zero temps rather than focusing on autotune and autosens.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:28
Have you adjusted your CR to keep up with autotune (or get ahead of it a bit)? That will help slightly with zero temping too.
And if you need more meal insulin, using the lower CR to calculate a bigger meal bolus will help, whether or not SMB zero temps.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:30
I haven't made that adjustment yet, but would have thought that the autotuned CR would have helped the rig adjust to the modest difference between 1:35 and 1:30 on its own. But I'll go ahead and make that adjustment tonight.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:31
It will help, but not as much as actually dosing based on the updated CR.
if you're doing a full meal bolus, SMB is unlikely to need to top it off until well afterward, so the lower CR won't help as much as it would if you were doing a partial meal bolus.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:32
I've also been using ES, which I thought would also pick up some of this slack, but it hasn't been enough to control the post meal highs.
In fact, I usually do ES and then don't even subtract out the net basal IOB when doing the meal bolus. Still doesn't seem like enough.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:33
There is another new branch you might want to try that has a preference that makes it treat ES like a "resistance mode" setting.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:34
Yes, I saw that. Is there any more need to test a52-risk? I've been testing that since last Friday.
How low does ES have to be set to trigger resistance mode?
Scott Leibrand
@scottleibrand
Oct 31 2017 16:34
If you can note in the PR that you think it's ready to merge we'll do that later.
The amount of resistance is proportional to the target, up to the autosens max.
So anything below about 90 will use the default max of 1.2x basal for IOB calcs etc.
It uses the same formula as exercise mode, so I think 80 would actually come out to 1.5x if you had your max that high.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:37
So ES has to be below 90? Just wondering because we sometimes use either 100 or 99 depending on whether I'm trying to trigger SMB. Our normal target is 130, so 80 sometimes seems too agressive.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:37
Hmm, if you're not even doing 80 for ES now, maybe try that first without any resistance mode?
But if you do 95 with the resistance mode setting it'll do half as much resistance as if you do 90.
It (barely) kicks in at 99 based on a neutral of 100.
Sounds like lowering your target might also be in order.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:39
Yes, this gets back to a conversation you and I have were having many weeks ago about pre-meal lows and post-meal highs followed by lows. But that was before the new insulin curves and a ton of other changes. So many moving parts.
Scott Leibrand
@scottleibrand
Oct 31 2017 16:39
SMB is optimized around a target of 100 or 110, so if you're battling mealtime highs lowering your target toward that might help.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:40
Basically, I was trying to be aggressive enough with ES to avoid the post-meal highs, but not so aggressive as to cause pre-meal and post-high lows.
And we are seeing mealtime highs, but more concerning to me is the way that these highs seem to linger long after the mealtimes.
That is where the zero temps seem to me to be a contributing factor.
Brian Rabinovitz
@bdr1177
Oct 31 2017 16:47
I should say potentially contributing factor, since I'm not really sure about the cause.
Jason Wittmer
@jbwittmer
Oct 31 2017 17:08
I don't know if it's related but we're also on the a52-risk branch and we have also have been seeing what appear to be inappropriately high autotune-set ISF in the setting of what I would perceive to be insulin resistance. We also likely have some confounding problems with basal and CRs as he's growing like the proverbial weed but this seems like too much of a coincidence that we're both seeing same thing?
Scott Leibrand
@scottleibrand
Oct 31 2017 17:35
There are no changes in that branch in autotune: you can review the code diff on the PR. Any autotune misbehavior would be the same in dev.
Jason Wittmer
@jbwittmer
Oct 31 2017 17:40
OK Sorry I hadn't done that. Probably just coincidence then but I'll try to drill down and see if I find something else.
Brian Rabinovitz
@bdr1177
Oct 31 2017 17:48
Here's another example of where this looks like a problem to me. Temp was set to zero, but BG was over 250 and rig wasn't connecting for over 30m, so I had school caregivers cancel temp basal to set it back to neutral. She had an additional 6g and 0.1U and rig connected again and set temp basal back to zero for 90m, even though bg is predicted to stay up around 200. I understand now that there is a safety feature here and the rig is setting zero temps because bg based solely on IOB is predicted to go low, but this just seems like a problem to me when the actual bg based on both IOB and carbs is predicted to be around 200. In this case, I wish that it would just set a neutral temp basal if it is going to do anything. Ninety minutes of zero temp is going to keep her at 200 or above for an extended period of time.
Capture3.JPG
Brian Rabinovitz
@bdr1177
Oct 31 2017 18:02
@scottleibrand Maybe the length of the SMB-related zero temp could be limited to the length of maxSMBBasalMinutes. I don't think this would completely solve the issue we're having, but it would at least prevent a really long zero temp remaining enacted if the rig stops looping shortly after enacting the long zero temp but bg continues to rise.
Scott Leibrand
@scottleibrand
Oct 31 2017 18:03
Sounds like you shouldn't be using SMB at school for now.
Brian Rabinovitz
@bdr1177
Oct 31 2017 18:03
Right. I'll disable it tonight.
Martin Haeberli
@mhaeberli
Oct 31 2017 18:03
@alimhassam - you’re welcome - glad it improved things!
Tyler Duncan
@yduncan
Oct 31 2017 20:10
image.png
Anyone know why I'm not seeing anything from my nightscout site?
What am I missing?
api-secret is correct in my rig, host site is correct in my rig. Why isn't it pushing data up there?
Tyler Duncan
@yduncan
Oct 31 2017 20:23

There isn’t anything to compare.

nightscout:master and yduncan:master are identical.

Seems like all settings are the same.