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

27th
Jan 2017
Dana Lewis
@danamlewis
Jan 27 2017 00:13
Created openaps/docs#584 to make it easier for @ceben80 and anyone else to give feedback on the xdripAPS instructions for using xdrip to get offline monitoring capabilities in OpenAPS. Please review and give it a thumbs up, or add to it if there’s anything missing so we can get this in to the docs for people to iterate on! :)
Thx @colinlennon ^ for writing it up!
chrisharris751
@chrisharris751
Jan 27 2017 00:30

@colinlennon Colin that all looks great. One issue that xdrip G5 user will run into is xdrip resetting bluetooth when it has trouble connecting for a reading. This causes the bluetooth tether to drop and requires manual re connection. One solution is the collector running on an android wear watch with xdrip+ (this is my current setup and works great) and the other thing I have found that helps is an app called BT AutoTether. It tries to auto reconnect if the tether is dropped. Its not always successful but it helps.

Thanks again for this feature, I've just spent 4 days in the country with no internet access and kept my rig looping the whole time! Cheers

Tim Street
@tim2000s
Jan 27 2017 00:37
Hmm. I think that depends on phone and android setup @chrisharris751 I ran it for a week on a cheap Chinese phone and had no issues with the G5 dropping and losing connection.
danamlewis @danamlewis breaks out the bingo card: “comment on the PR!” ;)
Tim Street
@tim2000s
Jan 27 2017 00:45
:+1:
chrisharris751
@chrisharris751
Jan 27 2017 00:45
How do I comment on the PR?
Dana Lewis
@danamlewis
Jan 27 2017 00:46
@chrisharris751 click on openaps/docs#584 and it should take you to the web page with the PR
Scott Leibrand
@scottleibrand
Jan 27 2017 00:47
This message was deleted
Dana Lewis
@danamlewis
Jan 27 2017 00:48
:+1:
chrisharris751
@chrisharris751
Jan 27 2017 00:48
Done
zekeat09
@zekeat09
Jan 27 2017 00:51
Hi, i need some advise for openAPS. I am started to get a compatible pump. My son is using enlite 2 CGM and i found one Paradigm 715 pump for sell. I was told 715 pump is non sensor pump. I am not sure i still can get this pump for OpenAPS using Medtronic CGM. In my country, only Medtronic CGM is available.
Can anyone advise ?
Dana Lewis
@danamlewis
Jan 27 2017 00:52
@zekeat09 welcome! yes, you can use a medtronic cgm but you’ll need an x22 or x23 that is both OpenAPS compatible and has Medtronic CGM sensor support on it.
Matthias
@ceben80
Jan 27 2017 01:18
Great @danamlewis thx for implementing xdrip offline in rtd :+1:
Dana Lewis
@danamlewis
Jan 27 2017 01:19
…I did the easy part of copying and pasting it and opening the PR :D can’t do the hard part of testing this one, though, so thanks to everyone who can!
Matthias
@ceben80
Jan 27 2017 01:19
N8 :wink:
Scott Leibrand
@scottleibrand
Jan 27 2017 01:34
Every time I see that I read "Nate" (short for Nathan). :)
Chris Oattes
@cjo20
Jan 27 2017 01:35
@jmcrawford The different versions of the edison are whether they're supplied as 1 or a multipack, and the amount of power given to the wifi/BT radio (less power = less range but less heat). The hardware itself is otherwise identical afaik
Scott Leibrand
@scottleibrand
Jan 27 2017 01:36
@cjo20 have you confirmed whether that wifi/BT power limitation is actually a limit in our use case (whether it actually reduces power use)?
Chris Oattes
@cjo20
Jan 27 2017 01:46
I've not looked for any documentation for the power consumption difference
Scott Leibrand
@scottleibrand
Jan 27 2017 01:47
is it more about peak power radiated vs. average?
Chris Oattes
@cjo20
Jan 27 2017 01:53
Noone seems to know the difference
Just that LPXX has lower power wifi than SPXX
Scott Leibrand
@scottleibrand
Jan 27 2017 01:55
yeah, supposedly. :-) I'll believe it when I see details or data.
Chris Oattes
@cjo20
Jan 27 2017 01:55
They all have the same datasheet, and it doesn't mention anything about it
Apparently with no radios, the edison itself draws 13mW, edison + BT is 21.5mW, edison + wifi is 35mW
Scott Leibrand
@scottleibrand
Jan 27 2017 01:59
we see battery life that would indicate average draw of > 100mW total
Chris Oattes
@cjo20
Jan 27 2017 01:59
That's with the edison on standby
Scott Leibrand
@scottleibrand
Jan 27 2017 02:00
ok, so the other 70mW+ is CPU usage?
Chris Oattes
@cjo20
Jan 27 2017 02:00
I guess so
Scott Leibrand
@scottleibrand
Jan 27 2017 02:00
and 900MHz I guess
Chris Oattes
@cjo20
Jan 27 2017 02:01
yeah. 2Ah battery @ 18 hours would be 111mA constant draw. 111mA @ 3.7v is 410mW
so if you shut the wifi off and only used BT, you'd save about 3% total power draw
Scott Leibrand
@scottleibrand
Jan 27 2017 02:03
oh, sorry, I was thinking mA not mW. if standby is only 35mW = ~10mA, then we're at more like 90% being CPU + 900MHz
Chris Oattes
@cjo20
Jan 27 2017 02:04
The main advantage for BT over Wifi is phone battery life though
Scott Leibrand
@scottleibrand
Jan 27 2017 02:05
yeah. I wonder how much we could do with a concerted effort to reduce CPU usage on the Edison
Chris Oattes
@cjo20
Jan 27 2017 02:05
Gut feeling says a lot
Scott Leibrand
@scottleibrand
Jan 27 2017 02:05
maybe first worth looking at how much is CPU vs. 900MHz comms somehow
because we can also reduce pump comms if it's worthwhile
Chris Oattes
@cjo20
Jan 27 2017 02:06
you'd use a hall effect current sensor on a TI stick connected via serial, most likely
Then you can see how much current it uses while transmitting
or just use a normal multimeter
Chris Oattes
@cjo20
Jan 27 2017 02:12
best thing we can do for battery power is to reduce the time it takes for the loop to run, It hink
Brian Rabinovitz
@bdr1177
Jan 27 2017 02:15
For the "reason" in suggested.json it says: "reason":"COB: undefined, Dev: -26, BGI: -1.44, ISF: 240, Target: 150; Eventual BG 106 < 140, setting -0.45U/hr, but 24m left and 0 ~ req 0U/hr: no action required"}
I understand that it is leaving temp basal at 0 because eventual bg is projected below target, but what does "~ req" mean?
Dana Lewis
@danamlewis
Jan 27 2017 02:16
short for required
Chris Oattes
@cjo20
Jan 27 2017 02:16
approximately equal to required
Dana Lewis
@danamlewis
Jan 27 2017 02:16
you can’t actually set a negative temp, so 0 is lowest it’ll go
Brian Rabinovitz
@bdr1177
Jan 27 2017 02:21
That was my assumption but I guess I just don't understand the grammar. -0.45 is what is required. 0 is what it's doing because a negative temp isn't possible. So what is "~ req" doing in that part of the explanation.
Chris Oattes
@cjo20
Jan 27 2017 02:21
Eventual BG 106 < 140 You're going to go low
Dana Lewis
@danamlewis
Jan 27 2017 02:22
it can’t do -0.45, so 0 is what it’ll set, and it’s approximately close enough to the required 0u/hr
Chris Oattes
@cjo20
Jan 27 2017 02:22
setting -0.45U/hr You need -0.45u/hr
Dana Lewis
@danamlewis
Jan 27 2017 02:22
so it’s going to leave the 0 temp for now
Chris Oattes
@cjo20
Jan 27 2017 02:22
but 24m left and 0 ~ req 0U/hr You've got 24 minutes of your current temp basal left, and the temp basal it's about to set (0u/hr) is approximately equal to the current basal rate (0U/hr)
Brian Rabinovitz
@bdr1177
Jan 27 2017 02:25
So "~ req" basically translates to "approximately close enough"?
Chris Oattes
@cjo20
Jan 27 2017 02:25
Well, the "~" means "approximately equal", and "req" is short for required
Brian Rabinovitz
@bdr1177
Jan 27 2017 02:26
Yes, makes perfect sense now.
Matt Walker
@walker0
Jan 27 2017 02:26
@scottleibrand @cjo20 The CC1110 processor uses ~8/9mA constantly if it isn't being put into sleep mode and the radio uses anywhere between 15-30 depending on whether it is in transmit or receive mode and the frequency. Unless intel has started supporting proper sleep modes for the edison since I last checked then you are probably better off going after CC1110 power consumption initially.
Dana Lewis
@danamlewis
Jan 27 2017 02:26
@bdr1177 this page might be good to read, and maybe PR if that’s not clear: http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-3/Understand-determine-basal.html
Brian Rabinovitz
@bdr1177
Jan 27 2017 02:27
I knew you what it was doing, just didn't understand that part of the explanation.
It's a little weird to me to use the abbreviation "req" when the whole word "required " is used a few words later. Just made me think that "~ req" had some special esoteric meaning. But maybe that's just my confusion.
Chris Oattes
@cjo20
Jan 27 2017 02:46
@walker0 even if it doesn't enter idle states, spending less time running instructions on the edison will reduce power consumption
Matt Walker
@walker0
Jan 27 2017 02:48
Have you noticed much difference in testing? I haven't seen power usage on the edison fluctuate that much, I should grab it and the fluke out again and do some testing to see how much of an impact it has.
Chris Oattes
@cjo20
Jan 27 2017 02:49
I don't have a nice way of testing power draw right now
Matt Walker
@walker0
Jan 27 2017 02:50
I'm happy to run some real world tests and report the results if there is anything specific you would like to know.
Chris Oattes
@cjo20
Jan 27 2017 02:50
power draw while there isn't anything running vs power draw when running, say, autosens would be interesting
Matt Walker
@walker0
Jan 27 2017 02:52
Alright, I'll grab my edison and test it later.
Dana Lewis
@danamlewis
Jan 27 2017 02:54
(if anyone has half an hour or so tonight and wants to help me alpha test something, please PM me)
Matt Walker
@walker0
Jan 27 2017 02:56
@cjo20 this covers a fair amount of the situations too http://geek-mag.com/posts/262398/
Chris Oattes
@cjo20
Jan 27 2017 03:19
@walker0 managed to sort of get something working (10A scale on a rubbish multimeter). Idle at command prompt via SSH is 0.06-0.07A. Running autosens is 0.16-0.17A
so thats in the order of 100mA between idle and a single core loade
Matt Walker
@walker0
Jan 27 2017 03:20
is that with autosens pulling from nightscout?
Chris Oattes
@cjo20
Jan 27 2017 03:21
no. That's pure CPU load.
Matt Walker
@walker0
Jan 27 2017 03:22
alright, interesting. I'll try mine out later today. I should get similar results..
Chris Oattes
@cjo20
Jan 27 2017 03:23
Loading up openaps was ~110mA, once it got in to processing properly, it was 160mA
Matt Walker
@walker0
Jan 27 2017 03:23
I guess that is in line with a ~110mA average load
Chris Oattes
@cjo20
Jan 27 2017 03:25
I suspect that autosens will be peak load
Chris Oattes
@cjo20
Jan 27 2017 03:40
@scottleibrand If those rough readings I did are right, then it'll probably be worth replacing openaps calls for things that can easily be scripted (mainly calls to nightscout.sh, or determine-basal etc)
JELCRAWFORD
@JELCRAWFORD
Jan 27 2017 05:07
IMG_0309.PNG
just when we thought our NS display had little left to surprise us with...at a loss of words (besides curse ones)....
Dana Lewis
@danamlewis
Jan 27 2017 05:25
@JELCRAWFORD wow..are you on mmol? actual bg is 19.2?
there’s obviously something wonky going on, wonder what BG openaps got that made it decode it as 39. remind me - are you on medtronic or dexcom sensor?
let me know if you want to troubleshoot. (and if you don’t want to deal with the headache now, you can always turn off the rig and go back to manual mode to deal with the pressing BGs)
JELCRAWFORD
@JELCRAWFORD
Jan 27 2017 05:33
well, it hasn't been a very good week, failed site, wacky bg's (yes mmol), so many changes even the pancreas can't keep up and I've 'perhaps' had an incident where resorted to rage basal changing in the middle of the night :) - so actual bg is probably somewhere closer to 15.5(279), but nothing better than a selection to choose from - 53.7 raw (I'm not even sure if that is meant to be mmol or not!), 19.2(345) or 39 (2.1) - it's a choose your own adventure, but I doubt it will involve much sleep for any of the grownups anytime soon. The above started with a compression to the dex(he's sleeping) - will go in here very shortly to re-calibrate now that we rolled him off the sensor and gave it a bit of time..... Hopefully if he doesn't lay on it again it will even out with something closer to what is the truth... In the end, perhaps NS is just giving us a visual summary for the crazy week! will certainly post if there is any troubleshooting to be done...thx!
Dana Lewis
@danamlewis
Jan 27 2017 05:35
Ah - the compression is probably where the 39 came from - that makes a tad more sense now
Boo for a terrible week :worried: hope it smooths back to more normal soon!
JELCRAWFORD
@JELCRAWFORD
Jan 27 2017 05:36
15.1 on poke.... not fantastic, but not the crazy madness of above either!
Dana Lewis
@danamlewis
Jan 27 2017 05:48
:+1:
JELCRAWFORD
@JELCRAWFORD
Jan 27 2017 05:59
@danamlewis - I'm assuming thats the openhuman stuff? doing it now...
Dana Lewis
@danamlewis
Jan 27 2017 05:59
yup
Tim Street
@tim2000s
Jan 27 2017 06:15
Anyone else tested the "How to lose your hci0 port" game?
Simply rm -rf <openaps directory>
When you used oref0-setup --btmac=AA:BB:CC:DD:EE:FF
then reinstall using the above.
It seems the the hci0 port for bluetooth falls off the face of the planet
*Recommend this is only done on a dev rig as you'll need to flash the edison to get it back.
I spent a good 24 hours trying to work out a way to recover my bluetooth port over the last couple of days while doing this and got nowhere.
Dana Lewis
@danamlewis
Jan 27 2017 07:05
HOUSTON WE HAVE BLUETOOTH TETHERING. Gahhhhhhhh. Freshly flashed edison did the trick, was able to reproduce and succeed by two different people and two different phones. finaaaalllllyyyyyy.
@Bender1061 @Aaybob ^ ::does a dance::
now - I used a hybrid between the existing docs, and the open PR…so need to try it on yet another rig to flesh out what the steps are (for iphone). so maybe tomorrow :) but happy that got it going at least once! progress :)
Dave Ewall
@daveewall
Jan 27 2017 07:07
Make that 3. I just re-flashed my edison on an Explorer board yesterday and it worked, but was struggling with it for a few days before that.
Dana Lewis
@danamlewis
Jan 27 2017 07:07
nice! @daveewall do you know what you did that worked?
JELCRAWFORD
@JELCRAWFORD
Jan 27 2017 07:08
@danamlewis @daveewall Yay! :)
Didn't do anything special aside from running the setup script.
Dana Lewis
@danamlewis
Jan 27 2017 07:09
ok
Dave Ewall
@daveewall
Jan 27 2017 07:10
I specified both --blemac and --btmac with the same mac address, not sure if that made a difference.
Dana Lewis
@danamlewis
Jan 27 2017 07:10
i’ll try with another fresh flash and make better note of my steps to see if there was anything from the existing master copy that was missing that i replaced with the open PR version
Dave Ewall
@daveewall
Jan 27 2017 07:12
Along the way, I had situations like the bluetoothctl not finding the controller, then another time it would pair and connect to the iPhone and not get an IP. After reflashing it and starting from scratch, it all seemed to work properly.
Tim Street
@tim2000s
Jan 27 2017 07:12
I've got a freshly flashed explorer that I can test this evening @danamlewis Which method should I use?
(are you iphone or android?, @tim2000s ?)
Tim Street
@tim2000s
Jan 27 2017 07:14
(Also seem to have the same issue on the MDT CGM again - pumphistory locked out - looks like it's linked to multiple sensor alarms occurring on the same page)
I have both @danamlewis but mostly use iPhone till new android turns up
Dana Lewis
@danamlewis
Jan 27 2017 07:15
k
Tim Street
@tim2000s
Jan 27 2017 07:17
Then a bunch of temp basals clears it.
(Provided by@me or by oref0 in this case)
Dana Lewis
@danamlewis
Jan 27 2017 07:18
yea
Tim Street
@tim2000s
Jan 27 2017 07:31
@daveewall We're you ever able to recover the Bluetooth controller when it got lost or was an edi flash the only way?
Dave Ewall
@daveewall
Jan 27 2017 07:37
I tried really hard to get the bluetooth working again, but nothing helped. Reinstalled bluez from apt, recompiled from source that oref0 downloaded, tried downloading a newer tarball and compiling that, --experimental flags, etc. Everything I tried didn't make a difference. No idea why it stopped working.
Flash fixed it right away.
Matt Walker
@walker0
Jan 27 2017 07:47
@danamlewis it sounds like a fair number of people are intimidated by gitter and would like an openaps Facebook page. I personally dislike Facebook for technical troubleshooting but I'd be happy to help if it is something that you think is a good idea.
Dana Lewis
@danamlewis
Jan 27 2017 07:51
@Kdisimone & I were talking about changed the description of the Looped group to be more inclusive of all the DIY looping options
Kind of like CGMITC as first stop, with the plan to redirect people to Gitter for the advanced technical pieces..
(There's already tons of OpenAPsers in there, too, so I think that likely makes most sense?)
Matt Walker
@walker0
Jan 27 2017 07:54
Sounds fine by me as long as we can keep the support clear so that people don't get confused by the differences, it does seem that a lot of people joining aren't clear about the fact that loop and openaps are distinct options that each have their own benefits.
Dana Lewis
@danamlewis
Jan 27 2017 07:54
:+1:
(Will get @ELUTE input tomorrow re ^ and go from there)
Tim Street
@tim2000s
Jan 27 2017 08:02
I did an interview with Stacey Simms that's coming out probably in the first week of February as a podcast. Hopefully that will provide a good place to point people at as talking about the differences.
Dana Lewis
@danamlewis
Jan 27 2017 08:08
:+1:
John Sjolund
@sjolundjohn
Jan 27 2017 09:49

Very very cool with autotune - just got it running. Question, I have multiple ISF and Carb ratios, they are all pretty very close to each other. Any thoughts experience on moving to a single ISF/Carb and then using autotune to try to get it as accurate as possible.

I am not 100% convinced that trying to optimize multiple ISF/Carb is the way to go if you have autotune and AMA?

Any experience or thoughts?

chrisharris751
@chrisharris751
Jan 27 2017 10:35
Evening All, I've still got an issue where my treatments don't show us in NS. My set up is edison/explorer rig with ble tether and xdrip offline. The bolus from a manual bolus shows up, it's just the bolus wizard calculator (from the pump) that doesn't show up in NS. I've triple checked all my time zones/settings and added the required config variables into heroku. I'm at a loss and really don't know where to go from here? thanks
Tim Street
@tim2000s
Jan 27 2017 11:38
@chrisharris751 Are your TbRs showing in NS? Have you checked your pumphistory.json file to see what is in there?
chrisharris751
@chrisharris751
Jan 27 2017 12:48
@tim2000s Yes tbrs are showing and everthing looks good in pumphistory.json the timestamps are all correct.
chrisharris751
@chrisharris751
Jan 27 2017 12:55
It's super annoying, I need the entries for autotune!
alternateal
@alternateal
Jan 27 2017 15:24
So this may just be me - trying to setup bluetooth on edison with explorer. I use bluetoothctl and it then locks the keyboard up - am I missing something??
Aaron Michelson
@Aaybob
Jan 27 2017 15:48
@alternateal does the cursor change to a blue color? What happens if you bluetoothctl and then type quit <Enter>?
Have you installed the latest dev? @alternateal
Tim Street
@tim2000s
Jan 27 2017 15:52
@alternateal I've seen that behaviour too. Just had to Ctrl-C
Aaron Michelson
@Aaybob
Jan 27 2017 16:04
@daveewall did you flash edison with the latest jubilinux where before was an older version?
Chris Oattes
@cjo20
Jan 27 2017 16:07
@alternateal are you connected via ssh or uart?
Aaron Michelson
@Aaybob
Jan 27 2017 16:19
@bdr1177 my RPi2 died so I got an RPi3 and the CareLink range became useable room wide. I also had to schedule "Preflight" code to run when loop was idle.
JaysonEwer
@JaysonEwer
Jan 27 2017 16:28
@danamlewis @scottleibrand I FINALLY figured out what the problem was with my NS site! Whew, that was frustrating. I just had to go to my NS Admin Tools page and "Clean Mongo status database">click on "Delete all documents". That's the first time I've had to do that in over a year of looping. Just barely over a year. 🙂. Is it possible to programmatically check that and clean it if necessary? I suppose I can ask that in the NS channel.
Tim Street
@tim2000s
Jan 27 2017 16:40
Quick question to those looping with MDT CGM. Do you find that the pump history cycle errors more regularly than when you use Dexcom so that you're effectively only looping four times per hour? This is what seems to be happening for me.
sdneufer
@sdneufer
Jan 27 2017 17:02
@walker0 @danamlewis I do a lot of volunteer stuff with non-technical people and come from a little more of an older tech background (I edit my pi files using vi). I see people generally will use what they need to get what they want, however Facebook is less intimidating to most than gitter for asking for help. The main issue with Facebook is people worry about privacy, ads, or generally do not like it and refuse to get an account. I would say that I personally find it easier to search through the cgminacloud facebook stuff than this chat group.
John Benjamin
@Bender1061
Jan 27 2017 17:02
@tim2000s I use MDT, never used Dex, I think the issue is a mmtune needs to be done, so the first thing that really "Fails" to call the mmtune is the history, it's failing on getting the cgm data, but nothing there is failing that would start an mmtune. So I'm thinking we need a rewrite on how the MDT cgm data is gotten, so that if it fails earlier it will just fix the mmtune right away. I think this is also why sometimes you can get a 10 to 15 mins delay of cgm data being uploaded to nightscout.
@danamlewis Wahooo!! Bluetooth Tether is Awesome!
Chris Oattes
@cjo20
Jan 27 2017 17:04
@tim2000s are you using a WW pump? Are you using the WW script before mmtune?
And it's not that pumphistory fails to call mmtune. If you're using oref0-setup, whenever anything fails it randomly chooses whether to mmtune or not. It should be re-tuning more often than 4 times an hour as far as I can tell. Pumphistory is something that is most likely to fail if there are interference problems because of the amount of data it needs to get from the pump, most of the commands only require a small packet of data. 24h pump history can be multiple pages of data, and if there's a problem with one of them the command fails.
alternateal
@alternateal
Jan 27 2017 17:08
@Aaybob yes it was the latest dev - not sure I saw it turn blue
@cjo20 I was on ssh at the time - should I be on uart instead?
Chris Oattes
@cjo20
Jan 27 2017 17:09
Also, @tim2000s what sort of case do you have it in? If it's in a case with a wire running in the wrong place it'll interfere with the transmissions.
@alternateal I was having problems with the terminal hanging on one but not the other. I can't remember which way round it was though.
alternateal
@alternateal
Jan 27 2017 17:10
@cjo20 - just reflashing so I am clean and will try it both ways around
Tim Street
@tim2000s
Jan 27 2017 17:11
@cjo20 I suspect the issue is that I have corrupt data on one of those pages. I had an issue yesterday that killed the ability to extract data. I suspect that is still causing a few issues.
John Benjamin
@Bender1061
Jan 27 2017 17:11
I always ssh into the uart, that one always works for me.
Tim Street
@tim2000s
Jan 27 2017 17:13
And if you're extracting multiple pages including the CGM data there's obviously a lot more to break.
Chris Oattes
@cjo20
Jan 27 2017 17:14
This is going to be pedantic, but You don't ssh in to the uart. The uart gives you a terminal, which you can use an ssh client to access. It doesn't actually use ssh though.
Dana Lewis
@danamlewis
Jan 27 2017 17:15
@sneufer yes, lots of pros and cons to all the channels. One big pro of Gitter, which we will always keep as a result (regardless of any other channels added on for other things) is the transparency and publicly available content.
@JaysonEwer Ahhhhh! That makes sense now.
Aaron Michelson
@Aaybob
Jan 27 2017 17:19
@alternateal If mine doen's turn blue it won't take any keyboard commands. Control C then terminates out to try again starting with the sudo /usr/local/bin/bluetoothd --experimental & command
@alternateal I used this page to get going: https://wiki.debian.org/BluetoothUser
Dana Lewis
@danamlewis
Jan 27 2017 17:27
@ohbrandon any luck on the pill in NS? If not at this point I'd check what @JaysonEwer did (clean mongo status database. Although that shouldn't be your issue), and if that scratches out, ask in CGMITC on Fbk for other things to try
apabari
@apabari
Jan 27 2017 18:44
Hello! Wondering, can anyone share an example of a wpa_supplicant.conf they could share with me? For some reason I can't get my rig to connect to my phone wifi, if I have two wifi networks listed in the file. It only ever connects if I list one wifi connection.
Scott Leibrand
@scottleibrand
Jan 27 2017 18:45
@JaysonEwer bummer you had to do that. That means one year of historical data that won't ever be donated for research. :worried:
apabari
@apabari
Jan 27 2017 18:45

```ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
network={
ssid="MyHomeWifi"
psk="MyPassword"
}

network={
ssid="iPhone"
psk="iPhonePassword"
}

```

this is what I have today...
Scott Leibrand
@scottleibrand
Jan 27 2017 18:46
@apabari are those the literal ssid's, or did you obfuscate them? Specifically the second one that breaks it...
apabari
@apabari
Jan 27 2017 18:48
The second one is the literal SSID
Tim Street
@tim2000s
Jan 27 2017 18:48
It's device status though @scottleibrand Whilst not ideal it's not treatment or blood glucose data?
apabari
@apabari
Jan 27 2017 18:49
I also changed the passwords before pasting into the chat...
Jarred Yaw
@jyaw
Jan 27 2017 18:50
Is there supposed to be some kind of Edison battery pill in NS with oref0-setup? I've seen what appeared to be some status pills on other's NS pics, but I can't figure out how to enable that.
Scott Leibrand
@scottleibrand
Jan 27 2017 18:57
@tim2000s true. So not nearly as bad.
@apabari only thing I can see different is that extra newline between networks. Maybe test with duplicating your working one and changing a non-special character to make sure it's not something with your SSID or PSK characters?
Tim Street
@tim2000s
Jan 27 2017 18:59
@scottleibrand a question about the stuck data pages in the MDT pump. Did you find they took ages to clear and the pumphistory cycle kept erroring after you'd cleared it with the multiple TBR trick? Any tips for clearing it?
There've been no changes on my rig and it's gone from cycling fairly reliably every 5 or so minutes to massive gapping. ( The worldwide script is operating before mmtune)
Scott Leibrand
@scottleibrand
Jan 27 2017 19:03
That doesn't sound like data pages any more. Sounds more like your cc1110 getting into a stuck state.
I would go run things manually and see exactly what works and what doesn't.
katie disimone
@Kdisimone
Jan 27 2017 19:17
Retina15_—_screen_◂_sudo_—_80×24.png
@scottleibrand just noting a doc deviation perhaps. in the setup for edison on mac, both my edison board flashes have said there is no “edison” user.
Scott Leibrand
@scottleibrand
Jan 27 2017 19:18
Oh, you're not on jubilinix. You're still on yocto.
apabari
@apabari
Jan 27 2017 19:18
strange... I still can't make this work. has anyone else had any issues connecting to multiple wifi networks from the Edison?
katie disimone
@Kdisimone
Jan 27 2017 19:19
oh…that’s what’s missing
Scott Leibrand
@scottleibrand
Jan 27 2017 19:19
@Kdisimone you'll need to re-do the flash.
Tim Street
@tim2000s
Jan 27 2017 19:25
Okay. So manually, I try to run pumphistory, it fails. I mmtune, it works.
Scott Leibrand
@scottleibrand
Jan 27 2017 19:25
Maybe you need to check the mmtune fallback logic then.
It might still be a bit off for MDT CGM?
Dana Lewis
@danamlewis
Jan 27 2017 19:26
Hey all - responding to the request for people who’ve been wanting to donate their data to research - we’re using the OpenHumans platform to create an “OpenAPS Data Commons” to enable that sharing - and to faciliate more research around closed looping. I just created the docs to guide you through the uploading data process if you want to donate your data: http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-4/data-commons-data-donation.html
katie disimone
@Kdisimone
Jan 27 2017 19:29
@scottleibrand what about this message?
Retina15_—_-bash_—_80×24.png
a previous attempt that that command had said “mac OS already provides this software and installing another version in parallel can cause all kinds of trouble."
I think it’s from an old homebrew install…must have left a couple files behind?
Scott Leibrand
@scottleibrand
Jan 27 2017 19:44
Probably fine. What actually fails when you flash?
alternateal
@alternateal
Jan 27 2017 19:48
@apabari nope - I have 4 different networks in my /etc/wpa_supplicant/wpa_supplicant.conf file and it will switch between them
Tim Street
@tim2000s
Jan 27 2017 19:59
Good call @scottleibrand - I was wondering about running the WW script and mmtune on every loop.
katie disimone
@Kdisimone
Jan 27 2017 20:09
@scottleibrand not sure what it was…flashing seeming to go fine this time around.
Chris Oattes
@cjo20
Jan 27 2017 20:11
@tim2000s you should run the WW script every time you are about to run mmtune. You shouldn't need to run mmtune every loop, it shouldn't be necessary and it takes forever to run.
katie disimone
@Kdisimone
Jan 27 2017 20:11
i didn’t get the prompt for control-D that first time..that should’ve been a hint
katie disimone
@Kdisimone
Jan 27 2017 21:20
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@       WARNING: POSSIBLE DNS SPOOFING DETECTED!          @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
The RSA host key for edisonhost.local has changed,
and the key for the corresponding IP address 192.168.1.27
is unknown. This could either mean that
DNS SPOOFING is happening or the IP address for the host
and its host key have changed at the same time.
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@    WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!     @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY!
Someone could be eavesdropping on you right now (man-in-the-middle attack)!
It is also possible that a host key has just been changed.
The fingerprint for the RSA key sent by the remote host is
SHA256:GpipfWcoLIrxOLMFaTCtdOSkFb3odFQT4g9MziEKsoU.
Please contact your system administrator.
Add correct host key in /Users/Retina15/.ssh/known_hosts to get rid of this message.
Offending RSA key in /Users/Retina15/.ssh/known_hosts:4
RSA host key for edisonhost.local has changed and you have requested strict checking.
Host key verification failed.
Retina15s-MacBook-Pro:~ Retina15$
new error. Got to the point of finishing the supplicant list. rebooted and went to ssh in.
@scottleibrand
Scott Leibrand
@scottleibrand
Jan 27 2017 21:27
That's because your new jubilinix is "pretending" to be the same device as yocto was earlier. You can delete the old key.
katie disimone
@Kdisimone
Jan 27 2017 21:27
how do i do that? sorry…don’t know
(and thank you for the explanation too.)
Scott Leibrand
@scottleibrand
Jan 27 2017 21:31
Easiest to just delete the known_hosts file.
Or edit it to remove the referenced line.
Tim Street
@tim2000s
Jan 27 2017 21:44
@cjo20 I'm starting to question my explorer board. I've two and one ran with no issues at all. The second (which I'm using while I rebuild the first) seems much less reliable on the pump comms. I'm also seeing issues with pulling the dirty CGM data.
Scott Leibrand
@scottleibrand
Jan 27 2017 21:46
try swapping the edisons onto the opposite explorer boards and see which device the problem follows?
Tim Street
@tim2000s
Jan 27 2017 21:47
Yup. Needs a spot of unscrewing :smile:
fezulin
@fezulin
Jan 27 2017 21:49
Hi, I am having trouble while setting up my Intel Edison for my loop. (Target is Edison + TI, while Explorer Board did not yet arrive.)
I flashed jubilinux and started configuration according to this documentation: https://github.com/oskarpearson/mmeowlink/wiki/Prepare-the-Edison-for-OpenAPS
After I tried to add the configuration for a second WLAN, I did a reboot.
The Edison did not come up again. Edison does not connect to WLAN anymore, and if I connect with putty (serial, 115200), I get the following output. It never gets further with booting.
******************************
PSH KERNEL VERSION: b0182929
                WR: 20104000
******************************

SCU IPC: 0x800000d0  0xfffce92c

PSH miaHOB version: TNG.B0.VVBD.0000000c

microkernel built 10:14:01 Jan 13 2015

******* PSH loader *******
PCM page cache size = 192 KB
Cache Constraint = 0 Pages
Arming IPC driver ..
Adding page store pool ..
PagestoreAddr(IMR Start Address) = 0x04899000
pageStoreSize(IMR Size)          = 0x00080000

*** Ready to receive application ***


U-Boot 2014.04 (Feb 09 2015 - 15:40:31)

       Watchdog enabled
DRAM:  980.6 MiB
MMC:   tangier_sdhci: 0
In:    serial
Out:   serial
Err:   serial
Hit any key to stop autoboot:  0
Target:blank
Partitioning already done...
Flashing already done...
GADGET DRIVER: usb_dnl_dfu
reading vmlinuz
3732144 bytes read in 98 ms (36.3 MiB/s)
Valid Boot Flag
Setup Size = 0x00003c00
Magic signature found
Using boot protocol version 2.0c
Linux kernel version 3.10.17-poky-edison+ (robin@robin-i7) #6 SMP PREEMPT Wed Mar 23 21:47:59 EDT 2016
Building boot_params at 0x00090000
Loading bzImage at address 00100000 (3716784 bytes)
Magic signature found
Kernel command line: "rootwait root=PARTUUID=012b3303-34ac-284d-99b4-34e03a2335f4 rootfstype=ext4 console=ttyMFD2 earlyprintk=ttyMFD2,keep loglevel=4 g_multi.ethernet_config=rndis systemd.unit=multi-user.target hardware_id=00 g_multi.iSerialNumber=8d4fcc68bebf415dfd649dc51f100cee g_multi.dev_addr=02:00:86:10:0c:ee"

Starting kernel ...

[    0.484891] pca953x 1-0020: failed reading register
[    0.490019] pca953x 1-0021: failed reading register
[    0.490351] pca953x 1-0022: failed reading register
[    0.500156] pca953x 1-0023: failed reading register
[    1.140241] mmc0: error -84 whilst initialising MMC card
[    1.317237] mmc0: error -84 whilst initialising MMC card
[    1.662446] mmc0: error -84 whilst initialising MMC card
[    2.070488] mmc0: error -84 whilst initialising MMC card
[    2.167626] pmic_ccsm pmic_ccsm: Error reading battery profile from battid frmwrk
[    2.176393] pmic_ccsm pmic_ccsm: Battery Over heat exception
Any idea, what could help me?
Scott Leibrand
@scottleibrand
Jan 27 2017 21:50
try unplugging the battery and see if it'll boot?
fezulin
@fezulin
Jan 27 2017 21:51
I tried with / without battery. I even tried with / without TI. Always the same behaviour.
I am connected to PC with two USB-cables.
alternateal
@alternateal
Jan 27 2017 21:54
Ok, so still not cracked Bluetooth - but explorer board up and running :smile:
Dana Lewis
@danamlewis
Jan 27 2017 21:54
:confetti_ball:
taylorfowler
@taylorfowler
Jan 27 2017 22:03
@fezulin You need to stop the auto boot where it says "Press any key to stop autoboot" Take a look at http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-0/setup-edison.html#troubleshooting letter "c"
Oh nm it's prrobably somethign else
fezulin
@fezulin
Jan 27 2017 22:06
Thank you! I just tried what is described there.
Actually I get an "multi-user" - so the problem seems to be something else. :worried:

       Watchdog enabled
DRAM:  980.6 MiB
MMC:   tangier_sdhci: 0
In:    serial
Out:   serial
Err:   serial
Hit any key to stop autoboot:  0
boot > printenv bootargs_target
bootargs_target=multi-user
Tim Street
@tim2000s
Jan 27 2017 22:09
Right then edis have been rehomed. Let's see what happens with the original explorer board.
fezulin
@fezulin
Jan 27 2017 22:10
What does "rehomed" meen?
katie disimone
@Kdisimone
Jan 27 2017 22:10
@scottleibrand thank you. fixed it
fezulin
@fezulin
Jan 27 2017 22:11
@tim2000s Do you mean the same Edi might behave different if plugged on the Explorer Board?
Tim Street
@tim2000s
Jan 27 2017 22:11
@fezulin I'm bug fixing a different issue.
fezulin
@fezulin
Jan 27 2017 22:12
Ah, ok.
Scott Leibrand
@scottleibrand
Jan 27 2017 22:14
@fezulin did you try the troubleshooting section of the flashing edison page?
fezulin
@fezulin
Jan 27 2017 22:20

@scottleibrand I wanted to answer "yes", I did. But after your suggestion, I re-read the chapter. I did not yet try

sudo ./flashall.sh

in the boot> prompt.
Thank you, I will try this.

Well... seems not to work the way it should:
boot > sudo ./flashall.sh
Unknown command 'sudo' - try 'help'
Scott Leibrand
@scottleibrand
Jan 27 2017 22:23
that doesn't run at boot>
the flashall runs in your other terminal window, on your local box
at boot > you'll do things like run do_flash or whatever the troubleshooting docs say
fezulin
@fezulin
Jan 27 2017 22:27
Oops. I will try it again.
I misunderstood point c.) of this document...
https://github.com/oskarpearson/mmeowlink/wiki/Prepare-the-Edison-for-OpenAPS
Sorry. As you may guess: I am not that familiar with Linux.
Tim Street
@tim2000s
Jan 27 2017 22:30
Well then @scottleibrand & @cjo20, it would appear that since my sensor ended last night, and caused a few issues, two things have gone on. One, the explorer board comms make a difference. Board 1 is not generating pumphistory issues.
  1. There are some issues with the glucose data extraction. My loop seems to be completing the dirty CGM extraction but isn't able to interpret and understand it. Except occasionally. Need to spend some time looking at it more deeply.
I think I may revert to a Dexcom rig while I troubleshoot it.
Dana Lewis
@danamlewis
Jan 27 2017 22:41
(if anyone is flashing an Edison today/next few days from a Mac and wants to test something for me, let me know)
Tim Street
@tim2000s
Jan 27 2017 22:52
Will be doing @danamlewis Let me know
fezulin
@fezulin
Jan 27 2017 22:55
@scottleibrand I had so much hope... flashall started well. But at one point, the Edison started to reboot. And the process stoped at the same place again. I did not touch or enter anything - so the edi is trying to reboot every 90 seconds - allways stopping at the same place. :worried:
Scott Leibrand
@scottleibrand
Jan 27 2017 22:56
you might need to do the "press any key to stop boot" and do stuff from there to reflash it. all of this should be documented in the troubleshooting section
fezulin
@fezulin
Jan 27 2017 22:58
@scottleibrand Thank you very much, for giving me hope! :smile:
I will re-read the chapter and try whatever is suggested.
Actually, I will have to do so at another day. I am just too tired - I would likely make mistakes.
So - thank you again and good night.
Scott Leibrand
@scottleibrand
Jan 27 2017 22:59
take care
fwiw, it's really difficult (perhaps impossible) to "brick" an Edison beyond repair when attempting to flash it. you just have to figure out the magic sequence of steps to re-flash it and try again. :)
Tim Street
@tim2000s
Jan 27 2017 23:02
My technique is to set up the screen command via mac or Linux, enter my password than hard turn the explorer board off and hard turn it on. When the little redlight that's not the main one flashes I hammer enter. That seems to get me into boot mode every time.
Jarred Yaw
@jyaw
Jan 27 2017 23:02
nevermind re my ns batter question earlier, I had upbat disabled in my config settings (doh!)...
fezulin
@fezulin
Jan 27 2017 23:04
@scottleibrand That is really good to know!
Tim Street
@tim2000s
Jan 27 2017 23:44
@danamlewis @scottleibrand @cjo20 I'm now certain it's an explorer board issue. Same Edison flipped over to the other board and I'm seeing regular CGM updates (in line with what I was expecting) and regular TBR updates. The whole loop is working as expected. I think it may be a case that some of the issues we are seeing are that not all explorers are made equal. I'm just not sure how to test this.
Chris Oattes
@cjo20
Jan 27 2017 23:46
@tim2000s Have you changed what sort of case it is in?
Scott Leibrand
@scottleibrand
Jan 27 2017 23:46
I think I already asked, but are you using @cjo20's new timing stuff from the latest (unreleased) mmeowlink master?
katie disimone
@Kdisimone
Jan 27 2017 23:47

I have a basic (?) question about the function of the edison for wifi networks and how it queries them, connects, disconnects, etc. Can someone help me by describing how often it looks for network? Does it look for a new network if it’s already connected? I’ve been reading the WiFi troubleshooting pages and wondering how it all plays together. Is all the suggested changes on that page still current? http://openaps.readthedocs.io/en/latest/docs/Resources/wifi.html

Also is there any differences between master vs dev. for this topic? (I thought I’d read something about oref0-online only being in dev??? but i could be mistaken or ignorant or both)

Finally, it seems like a lot of people are prioritizing their wifi networks. Can someone explain how that works and maybe whether we could add it to docs (if it isn’t already in there?) It sort of relates to that troubleshooting question of “I forgot to switch back to home wifi and it runs up my data plan"

Scott Leibrand
@scottleibrand
Jan 27 2017 23:50
@Kdisimone in addition to any tribal knowledge this group might be able to share, I would do some reading up on wpa_supplicant. this is definitely not an openaps-specific thing, so most of the knowledge is dispersed over the Internet rather than collected here...
oref0-online was designed for BT tethering, so it's only in dev. we might use it for helping with wifi switching too, though, as some people have been doing manually.
ejoe132
@ejoe132
Jan 27 2017 23:50
Need help getting my edison set up, I cant seem to get the libusb-1.0.dll download, follow the link in the walkthrough and just does a google search finidng nothing, but if I cut it down to just the http...stuff it finds the dfu-uti but not the other one
Tim Street
@tim2000s
Jan 27 2017 23:51
@cjo20 no. Both have been in Tic Tac boxes and the new cross over went into the tictac box where I was having issues, just to be sure.
@scottleibrand I'm running the mmeowlink 0.10.1 mmeowlink in both cases.
Scott Leibrand
@scottleibrand
Jan 27 2017 23:53
the fixes are unreleased, so I don't know if the version number is helpful here
Tim Street
@tim2000s
Jan 27 2017 23:53
Sorry, I mean, I'm running that on my current rig and it's what I was running previously.
Scott Leibrand
@scottleibrand
Jan 27 2017 23:53
did you specifically clone and install mmeowlink?
Tim Street
@tim2000s
Jan 27 2017 23:53
Yes.
(Same was true on the previous board that's midnrebuild)
Scott Leibrand
@scottleibrand
Jan 27 2017 23:54
k. sounds like we need to figure out how to reliably reproduce the problem at the lowest level we can (with manual commands) to figure out if it's an Explorer problem, and if so what kind.
katie disimone
@Kdisimone
Jan 27 2017 23:55
@scottleibrand OR ANYONE :smile: can you help me understand oref0-online, too? I am brand new to this and don’t grasp where that comes in or how it works. Sorry.
Scott Leibrand
@scottleibrand
Jan 27 2017 23:55
it doesn't come in yet, unless you manually set it up
ejoe132
@ejoe132
Jan 27 2017 23:55
not sure what im doing wrong go into cmd prompt and get to point for flashall.bat then it says dfu-uti is not recognized as an internal or external command
Chris Oattes
@cjo20
Jan 27 2017 23:55
I'm not convinced that it isn't a problem with something like a wire being in the wrong place when it's in the case
@Kdisimone oref0-online checks whether you have a network connection. If you don't, it tries to restart the wifi. If that doesn't help, it looks for a bluetooth connection.
Scott Leibrand
@scottleibrand
Jan 27 2017 23:56
@Kdisimone but, if you're having problems with losing Internet connection, you can use oref0-online to check if you have Internet, and if not, try to reset the wifi or optionally turn on bluetooth tethering
heh jinx
Chris Oattes
@cjo20
Jan 27 2017 23:56
I got there first this time :P
Scott Leibrand
@scottleibrand
Jan 27 2017 23:57
;-p
Dana Lewis
@danamlewis
Jan 27 2017 23:58
(we don’t have any documentaiton on the things @Kdisimone, so she’s asking so we can get a definition/explanation and PR it into the docs)
katie disimone
@Kdisimone
Jan 27 2017 23:58

AWESOME…that helps. Maybe my better question is this:

I have a home wifi and mifi. If my kid comes home from school, i’d like it to get off the mifi and back on the home wifi. Reverse when she goes to school ( off mifi and onto the school wifi). Will I need to do something “special” to do that? Supplicant list?

Chris Oattes
@cjo20
Jan 27 2017 23:59
@ejoe132 on windows? Have you copied and pasted the dfu-util in to the correct place?
@Kdisimone As far as I know, it'll stick with a network connection once it's found it. I turn on my iphone hotspot when I leave the house, and turn it off again when I get home.