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

23rd
May 2017
Scott Leibrand
@scottleibrand
May 23 2017 00:00
that might have been the 640G SmartGuard algorithm we were talking about
garykidd
@garykidd
May 23 2017 00:03
copy that
why i think of random one off conversations from a year or more ago---i have no idea
i'm 99% convinced that oref0 is much better than 670g and 99.99% that oref1 will be better yet.
glenrob1940
@glenrob1940
May 23 2017 00:07
Checking the log I got a no JSON could be decoded error message. My crontab has the correct nightscout info, but for some reason my .profile had an older url. I'm not sure of the relationship and how to fix it. I updated my .profile and think the best course is to re-run oref0-setup
Dana Lewis
@danamlewis
May 23 2017 00:07
hehe. likely that people will see improvements w/ oref1 over oref0. had 120g for lunch…for the longest time didn’t go over 120mg/dl..now I’m “so” high, riding flat in the mid-120s ;)
glenrob1940
@glenrob1940
May 23 2017 00:08
By the way should I just go to oref1? I've been mostly using loop at this time while I was waiting on a new explorer board.
Dana Lewis
@danamlewis
May 23 2017 00:09
how long have you been using openAPS with oref0?
glenrob1940
@glenrob1940
May 23 2017 00:10
Since November - used it exclusively till March and then went to loop because a fair amount of the time I needed to be offline and I couldn't get OTG to work on an early explorer board, plus my Rpi's were to bulky to carry
Dana Lewis
@danamlewis
May 23 2017 00:11
(k, just want to make sure people aren’t just brand new and switching without knowing how things look on oref0, to better test oref1 properly)
garykidd
@garykidd
May 23 2017 00:19
wonder what my BG curve will look like for https://glutenberg.ca/en/products?product=blonde&category=regular-products ....good too
rsilvers129
@rsilvers129
May 23 2017 00:34
@danamlewis You had 120 grams of carb for lunch and it kept you around 120?
Dana Lewis
@danamlewis
May 23 2017 00:34
@rsilvers129 was actually well below that, flat in the 90-100s for the first 3 hours; last hour is when it went up to 120s. yup :)
Scott Leibrand
@scottleibrand
May 23 2017 00:36
It wasn't just oref1: @danamlewis also knows how to bolus properly. But oref1 took care of the tail after the initial bolus started to wear off. :smile:
Dana Lewis
@danamlewis
May 23 2017 00:43
oh yea, in the world where we created UAM I should clarify :D 120 carbs entered, intiial meal bolus done
rsilvers129
@rsilvers129
May 23 2017 00:43
@danamlewis @scottleibrand Yes. Pre-bolus. But the system allows you to do that. The loop I am on now does not allow a low-temp-target at all, because, well, that would be setting it below the min target stated to the FDA. 120g carb would be great. Two Filet-o-Fish and an Apple Pie at McDonalds is 110 gram carb. I have never eaten low-carb, but about 10 days ago I started to after seeing that the loop I am on could not handle much carb and I concluded that I want to get a low A1c more than I wanted carb (I love carb), but I can give stuff up. I have been doing about 80 a day on average recently. But it would be interesting if I could get some of that back.
Dana Lewis
@danamlewis
May 23 2017 00:43
@rsilvers129 no pre-bolus. bolus as I put carbs in mouth.
rsilvers129
@rsilvers129
May 23 2017 00:44
@danamlewis You pre-announce though, which brings you down to 80?
Dana Lewis
@danamlewis
May 23 2017 00:44
I didn’t even do an eating soon mode today, cause I walked to pick up my lunch
but otherwise would’ve let the rig do a target of 80 heading into the meal, yes
but that’s not pre-announce, ie entering carbs in advance
rsilvers129
@rsilvers129
May 23 2017 00:45
Yes. It is just setting a target of 80 or so for 60 minutes or so.
Scott Leibrand
@scottleibrand
May 23 2017 00:45
I think SMB was able to do a single 0.7U microbolus between when she entered carbs and did her manual bolus.
glenrob1940
@glenrob1940
May 23 2017 00:51
My attempt at oref1 setup failed
""
oref0@0.5.0-dev global-install: npm install && sudo npm install -g && sudo npm link && sudo npm link oref0
npm ERR! Exit status 1
npm ERR!
npm ERR! Failed at the oref0@0.5.0-dev global-install script.
npm ERR! This is most likely a problem with the oref0 package,
npm ERR! not with npm itself.
''
Do I need to do something earlier in oref0 setup than this? (i did the dev line before this successfully in the setup)
Ignore my previous, I will start from the last portions of the explorer/edison board setup (after flashing) and try again.
Martin Haeberli
@mhaeberli
May 23 2017 01:18
@scottleibrand @glenrob1940 beating a dead horse, but @glenrob1940 ’s stale BG sounds like same issue I’m having...
glenrob1940
@glenrob1940
May 23 2017 01:24
I found that the url in my .profile was wrong - corrected it - and sourced it. I also got the G5 connected with OTG (plus 5V to the UART port). The stale data went away. Classic poor troubleshooting - did 2 things, so I don't know which fixed it.
John Benjamin
@Bender1061
May 23 2017 01:29
@scottleibrand put in PR 478 to fix that divide by Zed error. been working for me for the past hour.
Martin Haeberli
@mhaeberli
May 23 2017 01:39
@glenrob1940 thx. So you have a dexcom receiver w an OTG cable into the Edison?
glenrob1940
@glenrob1940
May 23 2017 01:40
yes
Let me qualify the yes. I had it connected with an OTG cable (long enough to see it with lsusb plus about 10 minutes). It is no longer connected and I am getting info from my nightscout url now.
peterfulvi
@peterfulvi
May 23 2017 01:43
does a 723 pump with a carellink chip openaps compatable is that what would be considered pc connect
Scott Leibrand
@scottleibrand
May 23 2017 01:44
@peterfulvi that question doesn't parse for me. Can you clarify?
John Benjamin
@Bender1061
May 23 2017 01:44
beat me too it
Scott Leibrand
@scottleibrand
May 23 2017 01:45
@Bender1061 wait, I thought Zed was Canadian for Zee, not for zero. And thanks. ;-)
John Benjamin
@Bender1061
May 23 2017 01:46
lol, I went to school at LSSU, one third of the students were Canadians, so I picked up a few things from them.
glenrob1940
@glenrob1940
May 23 2017 01:49
After starting over (basically to just after 'flashing', and having killed cron) I now have oref1 running. Clearly, I had missed something in the setup (or perhaps, things failed because an instance of oref0 was running during first attempt.
Aaron Michelson
@Aaybob
May 23 2017 01:53
@peterfulvi I think PC Connect is a menu item in the pump. Older pumps don't have that menu item. Its presence has been used to determine openaps compatibilty.
Scott Leibrand
@scottleibrand
May 23 2017 01:56
Yes, that is correct. If it's present under utilities, connect devices, then that pump's firmware is too new to loop.
rsilvers129
@rsilvers129
May 23 2017 01:58
I wanted to confirm.. My carb ratios, basals, and sensititivity lives in my pump? And OpenAPS pulls it from the radio connection? And do I also need to correctly set those in NightScout or not? And when Autotune shows I may want to change them, I change them on my pump, and then also on Nightscout?
Dana Lewis
@danamlewis
May 23 2017 01:59
correct
garykidd
@garykidd
May 23 2017 01:59
OMG Manchester. P+sitive thoughts...
Dana Lewis
@danamlewis
May 23 2017 01:59
(changing on NS is just for your sanity, doesn’t impact OpenAPS operations)
rsilvers129
@rsilvers129
May 23 2017 01:59
thanks
Martin Haeberli
@mhaeberli
May 23 2017 02:00
Do you add an external battery; if so, which one?
glenrob1940
@glenrob1940
May 23 2017 02:18
@mhaeberli Two factors - one, it has to be 5 volts (for OTG to work) and two, the size is a personal choice based on how long you want a charge to last. I don't have information on charge for edison/explorer, my experience (limited) is with a Raspberry Pi which requires quite a bit more current than the edison/explorer
peterfulvi
@peterfulvi
May 23 2017 02:19
so let me reprase my question woul a minimed 723 with firmware 2.3 be compatable. The person selling it said it has a carelink chip in which you could connect to a pc with. I did not know if this was related to pc connect.
Dana Lewis
@danamlewis
May 23 2017 02:26
@peterfulvi unrelated
rsilvers129
@rsilvers129
May 23 2017 02:35
Ran the Explorertest.py test suite and my second Explorer fails the test. My other one passes. So there is a problem with the board that is keeping the radio from working.
Scott Leibrand
@scottleibrand
May 23 2017 02:37
Yep, that sounds pretty conclusive. I'd contact ERD.
rsilvers129
@rsilvers129
May 23 2017 02:38
I emailed.
rsilvers129
@rsilvers129
May 23 2017 03:05
If you enable autotune, is it using those changes daily (within the range that you configure it for), or do you have to go into your pump and update it to what the autotune log suggests?
Dana Lewis
@danamlewis
May 23 2017 03:05
daily.
(capped by percentage from what’s in the pump, so if you’re constantly hitting those limits you may eventually want to update your pump)
rsilvers129
@rsilvers129
May 23 2017 03:07
I dont see that precentage in preferences.json
Dana Lewis
@danamlewis
May 23 2017 03:07
correct
rsilvers129
@rsilvers129
May 23 2017 03:07
Harcoded?
hardcoded
rsilvers129
@rsilvers129
May 23 2017 03:14
Is there a doc with suggest pushover notices? I am not sure which ones to put in.
Martin Haeberli
@mhaeberli
May 23 2017 03:19
incompetent me - trying to find out where to add to openaps loop list - Eden did just ask to join the google group
Scott Leibrand
@scottleibrand
May 23 2017 03:19
the safety caps for autotune are now inherited from the autosens min and max in preferences.json
rsilvers129
@rsilvers129
May 23 2017 03:20
Ok. Good.
Martin Haeberli
@mhaeberli
May 23 2017 03:20
found it!
Martin Haeberli
@mhaeberli
May 23 2017 03:29
she added herself to loopers : :clap: thx @danamlewis @scottleibrand
Dana Lewis
@danamlewis
May 23 2017 03:35
:+1:
rsilvers129
@rsilvers129
May 23 2017 03:52
Does the pump's basal type have to be set to units or percentage?
katie disimone
@Kdisimone
May 23 2017 03:52
@rsilvers129 units
rsilvers129
@rsilvers129
May 23 2017 03:55
thanks
I got NightScout pushover working. But how do I set up carbsReq as a pushover when it says I should eat?
Martin Haeberli
@mhaeberli
May 23 2017 03:59
a) report invoke cgm/ns-glucose.json never completes; when interrupted shows json parse errors; b) openaps get-ns-glucose completes with json parse errors; in any event, cgm/ns-glucose.json is empty
katie disimone
@Kdisimone
May 23 2017 03:59
it's not Nightscout that does the pushover. It's the rig that does the pushover
So you actually don't have to have pushover setup in NS
Martin Haeberli
@mhaeberli
May 23 2017 03:59
openaps get-ns-glucose sometimes updates cgm/ns-glucose.json to non-empty; other times, it leaves it empty
i’ve gotten away from the need, I think, to poll the NS api
but still trying to understand the ‘stuck’ state that I see
on another note, network is flapping up and down - i recall the suggestion to delete oref0-online from crontab
then single step
but I’m trying to understand why the network fails so frequently in the first place
rsilvers129
@rsilvers129
May 23 2017 04:01
How do I turn on pushover in the rig?
katie disimone
@Kdisimone
May 23 2017 04:01
@rsilvers129 if you look at the oref0/bin/oref0-setup.sh around line 108-114, you'll see the tags you need to add to setup to enable pushover alerts from rig
Martin Haeberli
@mhaeberli
May 23 2017 04:01
link down if wlan0 may call cfg80211_disconnected. event : 16, reason=2 from 88:ad:43:05:f8:88
DaisyHK
@laurathiessen
May 23 2017 04:01
I have my edison board all set up but I do not have my OpenAPS rig done yet.... so should I leave the edison board plugged in until I will use it or can I let the battery run dry without an issue when i recharge again? I am leaving on a work trip so i will set up my OpenAPS rig in about 2 weeks
katie disimone
@Kdisimone
May 23 2017 04:02
@laurathiessen just hold the power button down until the red light turns off. It will be fine to just plug it in to charge battery when you get home and finish off your install then
DaisyHK
@laurathiessen
May 23 2017 04:05
oh ok... i will plug it in to make sure the battery is fully charged before I do that thanks!!
rsilvers129
@rsilvers129
May 23 2017 04:07
@Kdisimone Ok, but don't I have to create an app for the rig on pushover.net's site? Or put the API key somewhere?
Martin Haeberli
@mhaeberli
May 23 2017 04:10
[ 2089.279164] wifi_platform_set_power = 0
[ 2096.616439] 
[ 2096.616439] Dongle Host Driver, version 1.141.59 (r)
[ 2096.616439] Compiled in /home/robin/src/edison-linux-helper/edison-bcm43340 on Apr  4 2017 at 22:42:52
[ 2096.616623] wl_android_wifi_on in
[ 2096.647783] wifi_platform_set_power = 1
rsilvers129
@rsilvers129
May 23 2017 04:18
Also I read someone say they have a pushover if their rig doesn't respond within 10 minutes. How does one do that?
rsilvers129
@rsilvers129
May 23 2017 04:27
I am getting a pump battery low pushover from Nightscout every few minutes. Is there a way to just get that once?
katie disimone
@Kdisimone
May 23 2017 04:48

@rsilvers129 I ended up turning off pushover on my NS site…it just alarmed too much for my tastes. BUT, the answer is that the repeat alerts in pushover depend on several factors. Yellow warn pushover alerts repeat at one interval, and red urgent pushover alerts repeat at a shorter interval (I can’t remember what they are off the top of my head…something like 15 min and 7 min respectively?)

BUT, in order to really snooze them, you need to snooze the alert in NS and acknowledge the pushover alert on your phone. simply snoozing it in NS will not stop/snooze the pushover alert.

So, if you want the pump battery to snooze for awhile, push the snooze in NS and then acknowledge the pushover. that will buy you up to 15 minutes or so (sorry, can’t remember the exact time)

@rsilvers129 you can have a pushover through NS by setting up the variable for the number of minutes that OpenAPS hasn’t looped in this variable in your heroku settings…OPENAPS_URGENT and OPENAPS_WARN. That would alert you if the rig has gone stale (but won’t necessarily tell you why of course)
@rsilvers129 yes, setting up separate API tokens for NS and the rig is a good idea. That way, if you decide you’d like to temporarily ever end one of them, but keep the other’s pushovers going, then you can just silence that API token alone.
@rsilvers129 the rig’s pushover is set to not realarm any sooner than 15 minutes from the last pushover message. And it only sends a pushover when there’s carbsReq or InsulinReq
Martin Haeberli
@mhaeberli
May 23 2017 04:59
@scottleibrand i killall -g openaps - trying to kill and debug, among others, oref0-online , but wifi still flapping up and down ...
looks to be the same on two different edisons
Scott Leibrand
@scottleibrand
May 23 2017 05:00
you can't killall -g openaps any more in dev if you're using SMB
I usually do killall -g oref0-pump-loop
Martin Haeberli
@mhaeberli
May 23 2017 05:01
ok i’ll try that
Scott Leibrand
@scottleibrand
May 23 2017 05:02
if you don't install with microbolus, it will still do things the old way. it's only if you're running oref0-pump-loop that you have to kill that instead
katie disimone
@Kdisimone
May 23 2017 05:02
@mhaeberli you’re way more advanced on your understanding of kernels, but just thought I’d share I see the same sort of message in my kernel logs…if that helps you narrow your search for the trouble
May 22 16:39:36 edison3 kernel:  [    7.307018] wifi_platform_set_power = 0
May 22 16:39:36 edison3 kernel:  [    7.308098] wifi_platform_bus_enumerate device present 0
May 22 16:39:36 edison3 kernel:  [    7.352126] systemd-tmpfile (122) used greatest stack depth: 6276 bytes left
May 22 16:39:36 edison3 kernel:  [    7.366916] g_multi gadget: using random host ethernet address
May 22 16:39:36 edison3 kernel:  [    7.371915] usb0: MAC 02:00:86:d1:86:24
May 22 16:39:36 edison3 kernel:  [    7.371936] usb0: HOST MAC fa:c2:e9:ae:10:13
May 22 16:39:36 edison3 kernel:  [    7.389850] g_multi gadget: Mass Storage Function, version: 2009/09/11
May 22 16:39:36 edison3 kernel:  [    7.389875] g_multi gadget: Number of LUNs=1
May 22 16:39:36 edison3 kernel:  [    7.389898]  lun0: LUN: file: /dev/mmcblk0p9
May 22 16:39:36 edison3 kernel:  [    7.390104] g_multi gadget: Multifunction Composite Gadget
May 22 16:39:36 edison3 kernel:  [    7.390119] g_multi gadget: g_multi ready
May 22 16:39:36 edison3 kernel:  [    7.391378] systemd-modules (107) used greatest stack depth: 5520 bytes left
May 22 16:39:36 edison3 kernel:  [    9.442884] EXT4-fs (mmcblk0p10): mounted filesystem with ordered data mode. Opts: (null)
May 22 16:39:36 edison3 kernel:  [   11.141759] 
May 22 16:39:36 edison3 kernel:  [   11.141759] Dongle Host Driver, version 1.141.59 (r)
May 22 16:39:36 edison3 kernel:  [   11.141759] Compiled in /home/robin/src/edison-linux/EDISON/bcm43340 on Mar 23 2016 at 21:48:12
May 22 16:39:36 edison3 kernel:  [   11.141791] wl_android_wifi_on in
May 22 16:39:36 edison3 kernel:  [   11.141807] wifi_platform_set_power = 1
May 22 16:39:36 edison3 kernel:  [   11.704377] F1 signature OK, socitype:0x1 chip:0xa94c rev:0x2 pkg:0x0
May 22 16:39:36 edison3 kernel:  [   11.705856] DHD: dongle ram size is set to 524288(orig 524288) at 0x0
May 22 16:39:36 edison3 kernel:  [   11.708027] dhdsdio_download_firmware: firmware path=/lib/firmware/fw_bcmdhd.bin, nvram path=/lib/firmware/bcmdhd.cal
May 22 16:39:36 edison3 kernel:  [   11.892144] sdioh_request_buffer: [1] doing memory copy buf=f6215000, len=2037
May 22 16:39:36 edison3 kernel:  [   11.897314] dhdsdio_write_vars: Download, Upload and compare of NVRAM succeeded.
May 22 16:39:36 edison3 kernel:  [   12.058245] dhd_bus_init: enable 0x06, ready 0x06 (waited 0us)
May 22 16:39:36 edison3 kernel:  [   12.060987] wifi_platform_get_mac_addr
May 22 16:39:36 edison3 kernel:  [   12.061021] wifi_get_mac_addr_intel: unable to open /config/wifi/mac.txt
May 22 16:39:36 edison3 kernel:  [   12.066264] Firmware up: op_mode=0x0015, MAC=fc:db:b3:9c:7a:1e
May 22 16:39:36 edison3 kernel:  [   12.086443] Firmware version = wl0: Aug 15 2014 15:09:17 version 6.20.190.2 (r497061) FWID 01-bb9bbd07
May 22 16:39:36 edison3 kernel:  [   12.087348] dhd_preinit_ioctls wl ampdu_hostreorder failed -23
May 22 16:39:36 edison3 kernel:  [   12.176019] CFG80211-ERROR) wl_update_wiphybands : bw_cap failed, -23
May 22 16:39:36 edison3 kernel:  [   12.429726] wpa_supplicant (272) used greatest stack depth: 5080 bytes left
May 22 16:39:36 edison3 kernel:  [   16.854017] CFG80211-ERROR) wl_cfg80211_connect : Connectting withe4:f4:c6:04:fd:2e channel (153) ssid "NETGEAR05-5G", len (12)
May 22 16:39:36 edison3 kernel:  [   16.854017] 
May 22 16:39:36 edison3 kernel:  [   16.941549] wl_bss_connect_done succeeded with e4:f4:c6:04:fd:2e
Scott Leibrand
@scottleibrand
May 23 2017 05:02
if you want to disable oref0-online, you can just comment it out in cron and wait a minute: it doesn't take long to finish
Martin Haeberli
@mhaeberli
May 23 2017 05:03
@scottleibrand ok, thx
@Kdisimone so I’m getting set_power = 0, set_power = 1 every few minutes … are you? or do you just get one at the beginning of time as you show above
every 60 seconds or so, set_power = 0
katie disimone
@Kdisimone
May 23 2017 05:05
Mine isn’t cycling, no. These logs are the last ones I have and they are about when we turned the rig on this afternoon and it’s been going solid since. So, it would appear that this string of logs was just a startup sequence and then once it connected (and stayed connected…it’s solid)
@mhaeberli do you see any messages in your logs about avahi-daemon and local?
Martin Haeberli
@mhaeberli
May 23 2017 05:06
no avahi ; no local
at least not in dmesg ; let me look at papertrail
no avahi ; local as in ‘my local ip address'
katie disimone
@Kdisimone
May 23 2017 05:08
@mhaeberli and I have no idea about the details, but those numbers in the brackets are like a counter…mine start at 0.00000 when the rig starts up and then climb until the wifi gets stable and connected which is about 2 seconds on the log time, but lots of lines of logging. :smile: So, your log number is a LOT higher than mine…definitely a lot of cycling then
Martin Haeberli
@mhaeberli
May 23 2017 05:08
yep, seconds since boot most likely
katie disimone
@Kdisimone
May 23 2017 05:15
that makes sense
Martin Haeberli
@mhaeberli
May 23 2017 05:16
@scottleibrand - ok , I commented out oref0-online from cron; did an ifdown wlan0 and ifup wlan0 and it has been stably up now for a few minutes
before it was going offline every minute
so where can I figure out how / what oref0-online is deciding?
katie disimone
@Kdisimone
May 23 2017 05:17
@mhaeberli this is the message that was hanging up three of my friend’s rigs…seemed to be issues with their router setups.
May 22 10:20:51 edisonjoe networking:  Warning: Stopping avahi-daemon.service, but it can still be activated by:
May 22 10:20:51 edisonjoe networking:  avahi-daemon.socket
May 22 10:20:51 edisonjoe avahi:  Avahi detected that your currently configured local DNS server serves
May 22 10:20:51 edisonjoe avahi:  a domain .local. This is inherently incompatible with Avahi and thus
May 22 10:20:51 edisonjoe avahi:  Avahi disabled itself. If you want to use Avahi in this network, please
May 22 10:20:51 edisonjoe avahi:  contact your administrator and convince him to use a different DNS domain,
May 22 10:20:51 edisonjoe avahi:  since .local should be used exclusively for Zeroconf technology.
May 22 10:20:51 edisonjoe avahi:  For more information, see http://avahi.org/wiki/AvahiAndUnicastDotLocal
May 22 10:21:49 edisonjoe systemd:  Starting Avahi mDNS/DNS-SD Stack...
May 22 10:21:49 edisonjoe avahi-daemon:  Found user 'avahi' (UID 104) and group 'avahi' (GID 107).
May 22 10:21:49 edisonjoe avahi-daemon:  Successfully dropped root privileges.
May 22 10:21:49 edisonjoe avahi-daemon:  avahi-daemon 0.6.31 starting up.
May 22 10:21:49 edisonjoe systemd:  Started Avahi mDNS/DNS-SD Stack.
May 22 10:21:49 edisonjoe avahi-daemon:  Successfully called chroot().
May 22 10:21:49 edisonjoe avahi-daemon:  Successfully dropped remaining capabilities.
May 22 10:21:49 edisonjoe avahi-daemon:  No service file found in /etc/avahi/services.
May 22 10:21:49 edisonjoe avahi-daemon:  socket() failed: Address family not supported by protocol
May 22 10:21:49 edisonjoe avahi-daemon:  Failed to create IPv6 socket, proceeding in IPv4 only mode
May 22 10:21:49 edisonjoe avahi-daemon:  socket() failed: Address family not supported by protocol
May 22 10:21:49 edisonjoe avahi-daemon:  Joining mDNS multicast group on interface usb0.IPv4 with address 192.168.2.15.
May 22 10:21:49 edisonjoe avahi-daemon:  New relevant interface usb0.IPv4 for mDNS.
May 22 10:21:49 edisonjoe avahi-daemon:  Network interface enumeration completed.
May 22 10:21:49 edisonjoe avahi-daemon:  Registering new address record for 192.168.2.15 on usb0.IPv4.
May 22 10:21:49 edisonjoe avahi-daemon:  Registering HINFO record with values 'I686'/'LINUX'.
May 22 10:21:49 edisonjoe avahi-daemon:  Withdrawing address record for 192.168.2.15 on usb0.
May 22 10:21:49 edisonjoe avahi-daemon:  Leaving mDNS multicast group on interface usb0.IPv4 with address 192.168.2.15.
May 22 10:21:49 edisonjoe avahi-daemon:  Interface usb0.IPv4 no longer relevant for mDNS.
May 22 10:21:49 edisonjoe avahi-daemon:  Joining mDNS multicast group on interface usb0.IPv4 with address 192.168.2.15.
May 22 10:21:49 edisonjoe avahi-daemon:  New relevant interface usb0.IPv4 for mDNS.
May 22 10:21:49 edisonjoe avahi-daemon:  Registering new address record for 192.168.2.15 on usb0.IPv4.
May 22 10:21:49 edisonjoe avahi-daemon:  Joining mDNS multicast group on interface wlan0.IPv4 with address 192.168.1.82.
May 22 10:21:49 edisonjoe avahi-daemon:  New relevant interface wlan0.IPv4 for mDNS.
May 22 10:21:49 edisonjoe avahi-daemon:  Registering new address record for 192.168.1.82 on wlan0.IPv4.
May 22 10:21:54 edisonjoe avahi-daemon:  Withdrawing address record for 192.168.1.82 on wlan0.
that avahi message at the top shows up about every 20 minutes, but the rig won’t stay connected inbetween the messages either. just keeps cycling
Scott Leibrand
@scottleibrand
May 23 2017 05:19
@mhaeberli you can run through the oref0-online bash script line by line if you want and see which checks succeed vs. fail, and what the script would have done in that case
Martin Haeberli
@mhaeberli
May 23 2017 05:20
@scottleibrand thx - is there some kind of log file that shows the decisions ?
Scott Leibrand
@scottleibrand
May 23 2017 05:20
network.log shows everything that oref0-online prints out
katie disimone
@Kdisimone
May 23 2017 05:22
while I’m on… @scottleibrand any ideas about how to fix a ISF over 255 to work with openaps? I know someone has done it previously, but I can’t find that person nor can i figure it out from gitter search. Definitely confirmed that’s what’s causing the rig to not loop.
Martin Haeberli
@mhaeberli
May 23 2017 05:23
based on network.log, it seems to be thrashing between BT tether and my wifi network
Martin Haeberli
@mhaeberli
May 23 2017 05:40
so it looks like my failure mode is that curl -s -m 15 icanhazip.com | egrep "^[12]*[0-9]*[0-9]\.[12]*[0-9]*[0-9]\.[12]*[0-9]*[0-9]\.[12]*[0-9]*[0-9]$"
fails sometimes, even when the network is working, which results in a flap
Martin Haeberli
@mhaeberli
May 23 2017 05:52
@scottleibrand time for me to re-read troubleshooting, but basically, when oref0-online is in the cron, it flaps every minute, even when the local BT is unavailable ...
At Mon May 22 22:51:02 PDT 2017, my wifi network name is AirportNetwork, and my public IP is: 
Error, cycling networking 
Stopping networking (via systemctl): networking.service.
Starting networking (via systemctl): networking.service.
and getting new wlan0 IP

At Mon May 22 22:51:23 PDT 2017, my wifi network name is AirportNetwork, and my public IP is: 
At Mon May 22 22:51:23 PDT 2017 my public IP is: 
Error, connecting BT to 90:60:F1:B2:E1:BB, getting bnep0 IP

At Mon May 22 22:51:29 PDT 2017, my wifi network name is AirportNetwork, and my public IP is: 2601:0646:c001:be3e:92b6:86ff:fe0c:5239
Finished oref0-online.

Starting oref0-online.
At Mon May 22 22:52:02 PDT 2017 my local IP is: 10.0.0.209

At Mon May 22 22:52:02 PDT 2017, my wifi network name is AirportNetwork, and my public IP is: 
Error, cycling networking 
Stopping networking (via systemctl): networking.service.
Starting networking (via systemctl): networking.service.
and getting new wlan0 IP

At Mon May 22 22:52:20 PDT 2017, my wifi network name is AirportNetwork, and my public IP is: 
At Mon May 22 22:52:21 PDT 2017 my public IP is: 
Error, connecting BT to 90:60:F1:B2:E1:BB, getting bnep0 IP

At Mon May 22 22:52:27 PDT 2017, my wifi network name is AirportNetwork, and my public IP is: 73.223.4.244
Finished oref0-online.
David Goelzhaeuser
@idoodler
May 23 2017 06:16
@mhaeberli I have had the same issue, I ended up using the dev branch
Martin Haeberli
@mhaeberli
May 23 2017 06:17
ok - i admit i was running these tests on master, but having the same experience on dev, I think
@idoodler but thanks - motivates me to re-check
Scott Leibrand
@scottleibrand
May 23 2017 06:30
@mhaeberli I would run just curl -s -m 15 icanhazip.com a bunch of times without all the grep stuff and see what you get
Martin Haeberli
@mhaeberli
May 23 2017 06:30
k thx
confirmed -edison1 was master; edison0 was dev ; re-trying dev oref0-online on otherwise master edison1 now, but looks like it is still ‘flapping'
maybe not flapping
new time constant with dev - 120 seconds or so ...
Scott Leibrand
@scottleibrand
May 23 2017 06:35
check the curl
and I think the curl command on dev is different
Martin Haeberli
@mhaeberli
May 23 2017 06:35
it is , ill try that one, but the ‘old’ curl returns variously empty, an ipv4 address, and ipv6 address
Scott Leibrand
@scottleibrand
May 23 2017 06:36
with this kind of troubleshooting, it's best to break it down into step by step and see what is happening at the point where it's not doing what you want.
the curl -4 in dev takes care of the v6 address not matching the regex
Martin Haeberli
@mhaeberli
May 23 2017 06:36
yees thx
yes thx
new curl curl --compressed -4 -s -m 15 icanhazip.com
Scott Leibrand
@scottleibrand
May 23 2017 06:37
but if you're getting an empty response to your curl, that might be the root cause of your issues: you might not actually have reliable Internet connectivity on this wifi network
Martin Haeberli
@mhaeberli
May 23 2017 06:39
 root@etghopenaps1:~# date; curl --compressed -4 -s -m 15 icanhazip.com[ 5730.598963] link down if wlan0 may call cfg80211_disconnected. event : 16, reason=2 from 88:ad:43:05:f8:88
[ 5730.610839] wl_android_wifi_off in
[ 5730.621686] wifi_platform_set_power = 0

Mon May 22 23:38:04 PDT 2017
root@etghopenaps1:~# date; curl --compressed -4 -s -m 15 icanhazip.com
Mon May 22 23:38:06 PDT 2017
root@etghopenaps1:~# [ 5737.447720] 
[ 5737.447720] Dongle Host Driver, version 1.141.59 (r)
[ 5737.447720] Compiled in /home/robin/src/edison-linux-helper/edison-bcm43340 on Apr  4 2017 at 22:42:52
[ 5737.447903] wl_android_wifi_on in
[ 5737.467672] wifi_platform_set_power = 1
[ 5738.033871] F1 signature OK, socitype:0x1 chip:0xa94c rev:0x2 pkg:0x0
[ 5738.042223] DHD: dongle ram size is set to 524288(orig 524288) at 0x0
[ 5738.052191] dhdsdio_download_firmware: firmware path=/lib/firmware/fw_bcmdhd.bin, nvram path=/lib/firmware/bcmdhd.cal
[ 5738.210377] sdioh_request_buffer: [1] doing memory copy buf=f4834000, len=2037
[ 5738.222755] dhdsdio_write_vars: Download, Upload and compare of NVRAM succeeded.
[ 5738.393565] dhd_bus_init: enable 0x06, ready 0x06 (waited 0us)
[ 5738.403953] wifi_platform_get_mac_addr
[ 5738.409784] wifi_get_mac_addr_intel: unable to open /c...
so it looks like oref0-online made the decision to take the link down; i’ll try to add some logging next time I try it ...
Scott Leibrand
@scottleibrand
May 23 2017 06:46
you're too far down the rabbit hole I think
I would try to reproduce producing the curl output that doesn't match the regex (doesn't show an IPv4 address)
if that fails, it's gonna do all the other stuff you've noticed, by design
Martin Haeberli
@mhaeberli
May 23 2017 06:54
@scottleibrand :-) re rabbit hole … speaking of which - for public consumption or not - I myself would understand oref0-online much better if it were in python vs bash (running exec/shell within) - has that been done / considered? I’m thinking it would be nice to make slightly ‘softer’ decisions, like it has to fail three times in a row to be deemed “off”, and maybe succeed two out of three to be deemed “on”, or something like that...
Scott Leibrand
@scottleibrand
May 23 2017 06:57
I find bash far easier to understand than Python, because it's much easier to transition from the command line to a script and back. That's probably mostly down to familiarity, though. you could do variables within bash if you wanted to, and accomplish the 2/3 thing if you wanted. but the logic here is: if you have flaky wifi, switch to bluetooth tethering until the wifi is happy again.
if you don't have BT tethering set up, and want to just keep using the flaky wifi without trying to "fix" it, you might want to just disable oref0-online
Martin Haeberli
@mhaeberli
May 23 2017 06:58
thx - good logic ; we do have BT tethering, but in fact i have disabled oref0-online because its decisions at the moment, dev or master, were causing more trouble than it’s worth; will have to re -enable by hand in the am. would be nice to generalize ‘flaky’ to x%
Scott Leibrand
@scottleibrand
May 23 2017 07:03
remember, we're talking about the simplest possible HTTP query: just making a single-packet GET request to icanhazip.com and asking for a single-packet response (with the client's public IP). If that fails, your wifi has serious problems, and we should probably be using Bluetooth instead if we want to be able to get BG data and upload treatments.
Martin Haeberli
@mhaeberli
May 23 2017 07:09
good points … http over tcp should anyway have retries / retransmits
rabbit holes aside, i suspect something a little deeper
but still looking
Paul Dickens
@thebookins
May 23 2017 07:11
@PieterGit Is there a "dummies guide" for updating the kernel anywhere? I had a read of the link you sent (https://edison.internet-share.com/wiki/Using_a_vanilla_Linux_kernel_with_Intel_Edison) but this seems more about re-tracing the steps that Andy Shev went through to create https://github.com/andy-shev/linux/tree/eds. I also note that U-boot needs to be updated? I did manage to update the kernel using https://github.com/esialb/edison-linux-helper, would that be any help to get to a 4.x version?
PieterGit
@PieterGit
May 23 2017 07:14
No, @esialb stuff is 3.x based. And yes you need to update u-boot to boot recent 4.x kernels
Martin Haeberli
@mhaeberli
May 23 2017 07:15
running a test overnight pinging poor 8.8.8.8; i’ll take a look in the am and see what packet loss - right now, doing this with oref0-online disabled (with it enabled, it died after a few minutes; so far, it has been running longer than that with 0 packets lost…)
PieterGit
@PieterGit
May 23 2017 07:15
There is quite some info in the threads but I agree it is not completely dummy proof
PieterGit
@PieterGit
May 23 2017 07:52
@mhaeberli you could also check your routing rules with route -n. I had some cases where the routing rules were messed up.
Martin Haeberli
@mhaeberli
May 23 2017 07:58
K thx
'''Temp refresherror, retrying''' ? In loop log
Martin Haeberli
@mhaeberli
May 23 2017 08:11
Probably pump battery
Paul Dickens
@thebookins
May 23 2017 08:16
thanks @PieterGit , I'll go back to the threads and give it a go. It looks like the link layer with with explorer board isn't working yet?
PieterGit
@PieterGit
May 23 2017 08:22
I started with replacing mraa with a pythoon gpio library. That's already in dev. I had pump rig communication once. But I stopped once it turned out a hardware issue with older explorer boards and usb otg.
cameronrenwick
@cameronrenwick
May 23 2017 12:15
when ppl are using the IFTTT "eating soon", just out of curiosity, how soon is soon (after button push?)
samihusseingit
@samihusseingit
May 23 2017 13:13
I will answer based on our own short experience. 30 minutes is minimum as it takes the insulin some time to be active
If your basal profile is active, and you don't have flucations (Rollercoaster) like our son, then it is great if you can do it 1 hour in advanced...but I believe 30 minutes is minimum :)
basal profile is (correct) I meant
meaning it keeps you steady
Martin Haeberli
@mhaeberli
May 23 2017 15:04
@scottleibrand - fyi, with oref0-online disabled, it was stable all night; still have to do stats, but curl every second, ping continuously, very low packet loss, 18 curl failures out of 19422 tries. as soon as i enabled oref0-online (within a minute or two), it failed / disconnected / reconnected, and resumed doing that on some periodicity. I’ll now try to step through oref0-online to see what triggered the flap ...
rsilvers129
@rsilvers129
May 23 2017 15:15
@cameronrenwick If you are already in the 80-100 range, and have insulin on board, then you shouldn't have to wait at all.
rsilvers129
@rsilvers129
May 23 2017 15:31
Anyone have a suggestion for what battery voltage to alert on that will give about 12 hours before an alkaline stops working?
peterfulvi
@peterfulvi
May 23 2017 15:31
@danamlewis ,@Scott Leibrand Let me get something straight about pump compatibility. So some of these pumps with the right version firmware can be incompatible beacause they might have pc connect regardless of the firmware version. So you have to be extra cautious even if you have the so called right firmware version. The pump i saw posted was a 723 with a firmware version of 2.3a, but the owner said it had pc connect in the utilities menu. Im glad i was cautious and did not buy a lemon.
tynbendad
@tynbendad
May 23 2017 15:43
@burmart_twitter fyi, just ran across this page (again), where @tim2000s has a good pic comparing size of edison rig to raspberry & rileylink http://www.diabettech.com/looping-a-guide/
Ryan Dalisky
@rdalisky
May 23 2017 16:17
That was a unicorn, get it!
Martin Haeberli
@mhaeberli
May 23 2017 16:19
@rsilvers129 re battery voltage … not a direct answer, but we have had very good luck / longevity using Energizer ultimate lithium in our pumps. some people i think have characterized the curve; it’s pretty steep at the end, i gather, so you may not want to cut it that fine (12 hours)...
rsilvers129
@rsilvers129
May 23 2017 16:20
I will buy Lithium for sure once I run out of alkaline,
Martin Haeberli
@mhaeberli
May 23 2017 16:21
i gather alkaline are well characterized
Dana Lewis
@danamlewis
May 23 2017 16:21
@rdalisky no - PC connect means it DOES NOT WORK. @peterfulvi.
rsilvers129
@rsilvers129
May 23 2017 16:29
Anyone have Hamshield’s physical address? They are slow to respond to email and their phone goes to voicemail, so if anyone has a package from them with an address on it, that would help me get their address sooner than waiting for them to get back to me.
rsilvers129
@rsilvers129
May 23 2017 16:39
If I remove the pump am I supposed to suspend it to let the rig know? Or does one use ITTT to let it know?
Arlene Samowich
@webwiztn
May 23 2017 16:39
@rsilvers129 The return address on my box is HamShield, 419 19th Ave E, Unit 20, Seattle, WA 98112-5342
rsilvers129
@rsilvers129
May 23 2017 16:41
Thanks.
Lorenzo Sandini
@lsandini
May 23 2017 16:42
I have a very interesting image to post in here, how do I do that ? (feeling dumb)
Insp 035.jpg
wasn't too difficult after all
Lorenzo Sandini
@lsandini
May 23 2017 16:50
This is a do-it-yourself insulin pump from 1984 (!). Connection to the Dexcom transmitter was added later. It works on a PIC16F690 CPU with a 32768 Hz frequency. It has a Faulhaber motor and the gearbox is handmade. The current draw was 100 uA, but when the Pololu Wixel was added and BT capacity, it increased a lot. The connection to a Dexcom sensor was added 3 years ago and the protocol was self written by the inventor, who remained silent and anonymous for a while.
Martin Haeberli
@mhaeberli
May 23 2017 16:53
@rsilvers129 HamShield 419 19th Ave E, Apt 20, Seattle, WA 98112-5342
rsilvers129
@rsilvers129
May 23 2017 17:06
@lsandini That is a great photo. Looks like a modern FDM 3D printed enclosure.
Scott Leibrand
@scottleibrand
May 23 2017 17:11
@peterfulvi if there exists a pump with normally-loopable firmware that also has PC Connect, that is something we've never seen before, and might be worth investigating further. most likely it's just an inattentive (or less likely unscrupulous) seller who is getting one of those two things wrong, though.
to be sure, you might have them send you a video showing them scrolling all the way down to the bottom of the main status screen to show the firmware, and then going into utilities, connect devices to show whether it has PC Connect.
dan lightner
@danimaniac
May 23 2017 17:18
@tynbendad Only my 2000mAh case has the "power button"...You have to pry up gently on the mechanism which then pushes the Explorer Board power button on the opposite side of the board. My 2500mAh and 18650 case have no such feature.
dan lightner
@danimaniac
May 23 2017 17:35
@burmart_twitter for size comparison here is my Edison, Explorer Board, and 2000mAh battery inside one of my cases. This one is mine next to some common items. https://goo.gl/photos/Z682o1qMvfSqxniN8
rsilvers129
@rsilvers129
May 23 2017 17:37
Why are the suggested max_iob in the 0 to 1 range? Why wouldn't someone want to let the loop keep running if they had a a higher carb meal?
Dana Lewis
@danamlewis
May 23 2017 17:39
@rsilvers129 where are you referring to that you’re looking at that?
rsilvers129
@rsilvers129
May 23 2017 17:51
Having trouble finding what I saw. I think it said to keep it at zero and then maybe go to 1.5 - 2x your max basal rate. Also it defaults to zero, and a default setting should be what is most typical.
Dana Lewis
@danamlewis
May 23 2017 17:51
it doesn’t say keep it at zero ;it talks about it defaulting to zero and a rule of thumb for changing it. http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-3/beyond-low-glucose-suspend.html
let me know if you see differently on any other pages
rsilvers129
@rsilvers129
May 23 2017 17:53
Says "This will default to zero. After several days or weeks, depending on your comfort level, you may choose to adjust this number. (Remember in the future if you re-run the setup scripts, it will default back to zero so you will come in here to adjust the max iob, as it is an OpenAPS-specific setting)." What is the rule of thumb for changing it?
Dana Lewis
@danamlewis
May 23 2017 17:54
"A good rule of thumb is for max iob to be no more than 3 times your highest basal rate. Keep in mind you can start conservatively and change this number over time as you evaluate further how the system works for you.” near the top of the page
rsilvers129
@rsilvers129
May 23 2017 17:55
That is what I saw.
Thanks
Is there a reason why or why not this should not be a much higher number to allow for looping when you eat a higher-carb meal?
Dana Lewis
@danamlewis
May 23 2017 17:59
YDMV
dan lightner
@danimaniac
May 23 2017 18:02
@rsilvers129 A max IOB of 0 is, in effect, only a predictive low glucose suspend system. It quits delivering insulin to minimize the chances of a low, but will do nothing to bring down a high. Increasing the max IOB above 0 allows the system to address highs. The further you increase the max IOB, the more aggressively the system will address those highs, but the higher max IOB also allows (sometimes) you to drop further than you or the system intended. That's the way I see it...mine is set to 4...my highest normal basal rate is 1.4. YMMV
ericmtx
@ericmtx
May 23 2017 18:06
I just saw on NS my daughter update her pump notification to vibrate while at school - SMB has been delivering multiple microboluses (nice!) but she noted that yesterday during class it went off every few minutes. ping, ping, ping.
rsilvers129
@rsilvers129
May 23 2017 18:09
I see. I had no idea that a max_iob default would basically turn OpenAPS into a Medtronic 535. It seems like people don't want to talk about this setting to avoid someone having a dangerous low, but I am not sure that is the best way to address it because it is avoiding useful information to let someone decide for themselves. Maybe a lot more information like what Dan just said would be good. I don't want to be the one to try to add that to the docs as it is a sensitive subject. What I would suggest for a way to approch keeping new users from having dangerous lows when starting out is to, rather than disable normal features of OpenAPS, start them on all of the useful features, but tell them to start with a target of 150, then lower it by 10 every X days, if they have no lows for two nights in a row.
live4sw
@live4sw
May 23 2017 18:10
@danimaniac FWIW, I have a max IOB of 1.5 with my highest basal being 0.85, but most of the day I'm at 0.6, so that's pretty similar to you. Definitely YMMV/YDMV
ericmtx
@ericmtx
May 23 2017 18:11
I disagree - I think it's the safest way to start people. Especially with how easy a new rig is to set up (I remember the weeks it took originally compared to the minutes it takes now). As people get into using and monitoring, AND fully reading the docs to understand, this aspect of dialing in the system to make it more aggressive is a natural next step. And one that people would either read and enable or come here to ask. Good safety valve for novices.
live4sw
@live4sw
May 23 2017 18:11
@rsilvers129 I think it is clear enough if one reads the docs. I think a lot of the docs are intentionally designed such that you have to read it pretty thoroughly before you can affirmatively do something beyond low glucose suspend.
The existence of the setup script made it a lot easier to set up a rig and so it makes it even more important for people to thoroughly read and understand the docs before going further.
rsilvers129
@rsilvers129
May 23 2017 18:12
The number of people who build an OpenAPS to get a low-gluclose suspend system is probably zero.
live4sw
@live4sw
May 23 2017 18:13
Everything is a building block, and it's important to thoroughly understand what's going on before going to the next step. But I'm sure that @danamlewis would be happy to consider any edits you propose, it's possible that you're right it isn't clear enough.
rsilvers129
@rsilvers129
May 23 2017 18:16
So there are two ways to go. One is to hide info and let people possibly discover it over time , or the other is to give the info and potential consequences up front, and let people decide what they want to do.
@ericmtx A new rig is taking me about 90 minutes now. First one was probably 12 hours.
live4sw
@live4sw
May 23 2017 18:17
My first rig took me two weekends, but that was in Feb 2016 when every couple of hours I'd be on gitter bothering someone for help
ericmtx
@ericmtx
May 23 2017 18:20
@live4sw I think I was that same timeframe with my daughter. There was a lot of asking her to sit near the PC while I tested (a LOT). It took me a few weeks to get everything working as expected and then more time trying to add autosens and meal handling as that was added. Felt like an install wizard this past week when doing the new one - so simple.
rsilvers129
@rsilvers129
May 23 2017 18:20
@live4sw Andrew, yes. That does make it pretty clear. I was looking for info on that setting and didn't find it because I searched for "max_iob" using google and then again in chrome, and the page has it as "max iob" (without the underscore), so I couldn't find it. I just proposed a pull where max iob is changed to max_iob so when people had some experience and want to find where it is mentioned, they can find it with search. I don't think it needs any other changes. I was only confused because I could not find max_iob info by searching.
ericmtx
@ericmtx
May 23 2017 18:21
And the docs have improved tremendously since the beginning. Kudos to everyone who had a hand in improving them.
rsilvers129
@rsilvers129
May 23 2017 18:24
@ericmtx Flashing Edison is under 15 minutes, even on my first time.
katie disimone
@Kdisimone
May 23 2017 18:26
May 23 11:19:09 caraedison pump-loop.log:  {"requested":{"duration":30,"rate":0.4,"temp":"absolute"},"bg":278,"temp":"percent","snoozeBG":413,"timestamp":"2017-05-23T11:19:05.752919","predBGs":{"IOB":[278,273,265,254,241,226,208,187,164,139,111,81,50,39]},"rate":85,"reason":"COB: undefined, Dev: 102, BGI: -21, ISF: 350, Target: 130; Eventual BG -419 < 130, bolus snooze: eventual BG range -419-413; setting current basal of 0.4 as temp. Setting neutral temp basal of 0.4U/hr","eventualBG":-419,"recieved":false,"duration":0,"tick":-4,"IOB":2.282}
May 23 11:19:29 caraedison pump-loop.log:  glucose.json newer than pumphistory: Refreshed pumphistory
May 23 11:19:35 caraedison pump-loop.log:  {"carbs":0,"reason":"not enough glucose data to calculate carb absorption"}
May 23 11:19:35 caraedison pump-loop.log:  {"ratio":1,"reason":"not enough glucose data to calculate autosens”}
what am I missing for how the BG is seemingly being read from NS (278) but not being used in the loop (“not enough glucose data to calculate carb absorption")?
piss…i see it now.
temp: percent
JELCRAWFORD
@JELCRAWFORD
May 23 2017 18:29
@Kdisimone - boo on the "percent" - easy miss though...
tynbendad
@tynbendad
May 23 2017 18:30
fwiw, @rsilvers129 i tend to agree only because i hate changing that one setting every time i reinstall
@danimaniac but you can get to the reset button thru the vent holes on the 2500 case?
Martin Haeberli
@mhaeberli
May 23 2017 18:56
@rsilvers129 - hate to quibble, but our original goal was managing night-time low bg suspend … we had a 723 aka “530” with that feature, but the enlite sensors were not doing it for us. since we have gotten as far as we have, we’re making broader use of it.
dan lightner
@danimaniac
May 23 2017 19:00
@tynbendad Yes, you could access the Edison's power / reset button through the vent in the top of the 2500mAh case's lid.
David
@davidkeddydb
May 23 2017 19:16
Screenshot 2017-05-23 15.15.02.png
I wish I could get the Zero temp basal time to about 1/2 of what it is now as I sort of have results like this more than I would like with SMB
Scott Leibrand
@scottleibrand
May 23 2017 19:26
@davidkeddydb have you adjusted remainingCarbsFraction yet? there's now a branch that exposes that in preferences.json if you'd like to test it. what is your DIA set to?
David
@davidkeddydb
May 23 2017 19:27
My DIA is set at 3. I adjusted CarbsFraction to 1. I was able to find it in the program files :smile: However, if it is in dev or another branch, I would love to test it.
@scottleibrand I thought but haven't done it, to change it to 1.5 and that would 1/2 the zero temp unless there is another safety in the code, but I haven't started poking around yet
Scott Leibrand
@scottleibrand
May 23 2017 19:32
in the remainingCarbsFraction branch, I enforce a maximum of 1. if you want to be more aggressive than that, you'd probably want to do it with your carb ratio instead.
David
@davidkeddydb
May 23 2017 19:37
@scottleibrand I hear you, but I know for me I would literally be better off when eating items that I know the carb count, to just bolus and go off OpenAPS to not zero temp and my post prandial rise would be less. The flip side though is I can't get the UAM when I don't have the exact carb count. I was actually thinking that I could try making my ISF lower and see how that works post prandial, but then the risk of when no food. I know the carb ratio is right.
Dana Lewis
@danamlewis
May 23 2017 19:48
@davidkeddydb one thought - what about trying turning off smb with bolus, and only using temp targets and/or UAM ?
seeing that with one other person who’s super precise with their carb counts, traditional AMA and their bolusing works well for them
then if you decide it’s an imprecise meal and you want SMB, set a temp target (even 99 if your target is 100) and it can SMB during that time frame. otherwise it won’t, because you won’t have SMB from bolus turned on
rsilvers129
@rsilvers129
May 23 2017 19:49
@danimaniac @tynbendad I see. I will add a hole to my enclosures for the reset button. I forgot that is half the reason why I flipped the Explorer in the enclosure to begin with.
I will also revamp the fit so that they are not too tight for most uncalibrated printers.
peterfulvi
@peterfulvi
May 23 2017 20:19
@scottleibrand thanks for the advice I was wondering the same thing also I will do some investigating
tynbendad
@tynbendad
May 23 2017 20:43
@danimaniac @rsilvers129 thank you both, just got my printer delivered today, hope I can manage some usable prints ;)
Martin Haeberli
@mhaeberli
May 23 2017 21:39
@tynbendad off topic - what printer did you get?
@scottleibrand question : could my problems with oref0-online relate to my wifi network name being “Airport Network” - with a “space” between the words?
@scottleibrand trying to reproduce at work with a name of the form: “XXX”, not failing at the moment
Marco
@CaptainBalou
May 23 2017 21:50
@mhaeberli Hi. You havr stable connection at work with wlan? Is your wlan via wlan0? I cannot see any problem on using a space in between. What does you network.log say while beeing offline?
jmcrawford
@jmcrawford
May 23 2017 21:53
@mhaeberli My $.02: spaces are bad. It takes a lot more care in programming to determine what is one argument with a space vs. two arguments separated by a space. If you've got a choice in the matter, eliminate the space.
Martin Haeberli
@mhaeberli
May 23 2017 21:59
yes via wlan0
i’ll (try to) rename when I get home...
jmcrawford
@jmcrawford
May 23 2017 22:03
That goes for file/directory names as well. I'm looking at you, My Documents.
/rant
PieterGit
@PieterGit
May 23 2017 22:04
@scottleibrand : is this what you had in mind for changing API_SECRET in ~/.bash_profile?
https://github.com/PieterGit/oref0/blob/201705_fixes/bin/oref0-setup.sh#L742-L754
Diabetic_Dude
@diabetic_dude_twitter
May 23 2017 22:07
would it be possiable to get something like this working with an edison set up? (it is a battery power indicator) https://www.amazon.com/DROK-Lead-acid-Capacity-Indicator-Electric/dp/B01I30RAC4/ref=sr_1_3?ie=UTF8&qid=1495574442&sr=8-3&keywords=Battery+Capacity+Monitor
Martin Haeberli
@mhaeberli
May 23 2017 22:09
@diabetic_dude_twitter for what it’s worth, openaps already pushes battery level to a pill in Nightscout ...
Diabetic_Dude
@diabetic_dude_twitter
May 23 2017 22:10
I thought that was for the pump itself
dougtakeuti
@dougtakeuti
May 23 2017 22:10
Hi there, I've been out of this gitter for a while and now that the school year is almost over it is time for me to catch up with the latest developments. I just got two new Explorer Boards and two EDI2.LPON Edisons to make MDT Enlite rigs. I'll be working from scratch and want to eventually implement SMB. How can I help testing/verifying features or docs? Right now rig temperature is definitely something I'll be watching as this is critical here in Brazil.
Martin Haeberli
@mhaeberli
May 23 2017 22:10
the main caveats are: as far as I know, it can’t tell when it’s charging or not; and … when it’s fully charged, it shows 66% on the charger, but when you disconnect the charger, it will show full or near full
definitely there is one for openaps ...
for the pump, the pump’s battery VOLTAGE is easily shown in an openaps pill
some magic may exist (I have seen comments about this) to scale pump battery voltage to percent - it can get tricky because the curve is different between alkaline and lithium-ion batteries
jmcrawford
@jmcrawford
May 23 2017 22:12
A device like that also draws some current which will reduce time between charges
Diabetic_Dude
@diabetic_dude_twitter
May 23 2017 22:12
I am not worried about the pump only the APS... I also agree about the power draw being a big set back
I need to watch that pill more and see what I get. I thought is was for the darn pump. (duh no wonder it was always off)
thanks I will see what I get tonight, I am at 60 and cant charge for another 7 hours.
see you all tomorrow and thanks again
jmcrawford
@jmcrawford
May 23 2017 22:15
Pump battery level shows in the pump pill, if you've got that enabled on your NS site.
Diabetic_Dude
@diabetic_dude_twitter
May 23 2017 22:18
@jmcrawford I have pump enabled in my config
what needs to be enabled to show the openAPs battery
Martin Haeberli
@mhaeberli
May 23 2017 22:18
@diabetic_dude_twitter openaps pill, I think...
but you have to enable the pill at the web interface ...
Diabetic_Dude
@diabetic_dude_twitter
May 23 2017 22:19
ok I have that also
checking that now
I am just going to watch and test tonight
I am pretty sure I already have it enabled.
Dana Lewis
@danamlewis
May 23 2017 22:29
@dougtakeuti welcome back! :) It would be helpful if you can test the dev branch with your current set up of features and +1 that it doesn’t break for non-oref1 looping. then when you’re ready to enable SMB and toggle on those features, it’ll be easy to run again with the flag and add those and test those features, too.
dougtakeuti
@dougtakeuti
May 23 2017 22:30
@danamlewis ok then! :+1:
rsilvers129
@rsilvers129
May 23 2017 22:31
I see pump voltage. How can I see rig voltage on NightScout?
Martin Haeberli
@mhaeberli
May 23 2017 22:33
rig.png
@rsilvers129 see above
the 66% corresponds to a fully charged rig that is still plugged in
rsilvers129
@rsilvers129
May 23 2017 22:34
Oh. I have that.
Mine shows 62% plugged in.
Martin Haeberli
@mhaeberli
May 23 2017 22:34
suggests it’s still charging
rsilvers129
@rsilvers129
May 23 2017 22:35
It has been plugged in all night
Martin Haeberli
@mhaeberli
May 23 2017 22:35
or try unplugging it, wait a few minutes, and see what it shows
:+1: confirming for others, this is part of the openaps pill
rsilvers129
@rsilvers129
May 23 2017 22:35
I want to know how to keep Nightscout from sending me a battery warning very 15 minutes. I just want it once.
Martin Haeberli
@mhaeberli
May 23 2017 22:35
now mine shows 93%
@rsilvers129 i haven’t dug in that part of the ocean yet :-)
rsilvers129
@rsilvers129
May 23 2017 22:36
Mine shows 95% unplugged
Martin Haeberli
@mhaeberli
May 23 2017 22:36
:+1:
tynbendad
@tynbendad
May 23 2017 22:47
@mhaeberli monoprice
glenrob1940
@glenrob1940
May 23 2017 22:49
A question on setup. If you build the lookout version of loop and use xdrip in setup, what is the behavior if you also give your nightscout url API stuff?
Dana Lewis
@danamlewis
May 23 2017 22:55
“lookout”?
Martin Haeberli
@mhaeberli
May 23 2017 23:01
@tynbendad :+1: thx
rsilvers129
@rsilvers129
May 23 2017 23:02
Pretty neat. I entered my meal into MyNetDiary, and I noticed that Loop (which I am running on my phone for no reason) saw that I looged it, and uploaded the carb info to NightScout as a bolus with the grams of carb.
tynbendad
@tynbendad
May 23 2017 23:02
@,
@danamlewis believe he is referring to @thebookins lookout loop-for-BG-for-openaps app. I'm also curious what happens when phone/rig switch from WiFi to BT tether to no coverage/etc with that setup, haven't built it yet.
rsilvers129
@rsilvers129
May 23 2017 23:05
When we use Bolus Wizard, I assume we should never enter our BG into the pump, right?
tynbendad
@tynbendad
May 23 2017 23:05
@rsilvers129 that is Apple's health kit integration. If really using loop it is recommended to turn that off so you don't get double entries...
John Benjamin
@Bender1061
May 23 2017 23:06
@scottleibrand would you rather in openaps/oref0#478 have me make a function that makes a variable that fills for all of the cgm types and call that function at the beginning of the loop, and use that data. Technically we only need for the MDT but if I make this it can be used for if we need anything special in the future. or do you want me to just make a variable that is used for the MDT only?
tynbendad
@tynbendad
May 23 2017 23:06
Yes, no BG in pump, correction is already in your IOB
rsilvers129
@rsilvers129
May 23 2017 23:09
It would seem real easy for Loop to pull HR out of apple heath and have activity data.
David
@davidkeddydb
May 23 2017 23:20
@danamlewis 100% I also may try changing my DIA and see what that does as I am currently at 3 and I would be curious to see how it reacts if I go to 4. I am happy to experiment with all of it and see what happens.
Dana Lewis
@danamlewis
May 23 2017 23:20
@davidkeddydb :+1: I’d just test one thing at a time ;) so you know which does what
Martin Haeberli
@mhaeberli
May 23 2017 23:26
trying to check out latest dev to test with - but oref0-runagain says something about: * master (I already cloned, and git checkout dev) - I hope that works
glenrob1940
@glenrob1940
May 23 2017 23:28
It's the one off build of loop in the docs/walkthrough/phase-1 "Local, offline BGs for iPhone users using a separate Loop app"
I'm currently using it (xdrip/loop) and just hit enter when setup asked for nightscout info.
Dana Lewis
@danamlewis
May 23 2017 23:33
yes, I’ve just never heard it called “lookout"
Dana Lewis
@danamlewis
May 23 2017 23:42
@mhaeberli git branch to check what branch you’re on
Martin Haeberli
@mhaeberli
May 23 2017 23:43
@danamlewis thx - it says dev so I’m good to go...