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

26th
May 2017
rsilvers129
@rsilvers129
May 26 2017 00:12
Doing a runagain to rebuild everything didn't fix this. The error is still repeating, and the loop is not working.
peterfulvi
@peterfulvi
May 26 2017 00:40
@rsilvers129 I'm in the same boat. I'm​ going to try update the branch and run setup again. Good luck.
rsilvers129
@rsilvers129
May 26 2017 00:52
What is worse - I shut off that rig (on dev), and turned on another one (on master). I did a runagain. It came up with the same error! The error propagated from one rig to another.
PopcornPam
@PopcornPam_twitter
May 26 2017 00:53
Hi all, I finally set up papertrail, but seem to have 2 separate logs being created, neither of which show me much for pump-loop
I just get this May 25 13:17:01 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
and no info like the pill in nightscout shows
rsilvers129
@rsilvers129
May 26 2017 00:54
@peterfulvi when it broke for you, were you running dev? Maybe something about dev corrupted us.
peterfulvi
@peterfulvi
May 26 2017 01:07
That's what I'm thinking. It was dev branch I was running.
Looks like I am back online. I wish I knew what caused it so I could avoid it.
rsilvers129
@rsilvers129
May 26 2017 01:37
I see this in the troubleshooting: "Check to make sure you have a carb ratio set manually in your Medtronic insulin pump, if it is not done, the follwoing display will appear in your pump.log: Could not parse input data: [SyntaxError: /root/myopenaps/monitor/iob.json: Unexpected end of input]"
I did have a carb ratio. But not for every hour of the day. I wonder if the code got confused and thought there was no carb ratio?
peterfulvi
@peterfulvi
May 26 2017 01:46
@rsilvers129 Thanks. Im starting the process of elimination now
rsilvers129
@rsilvers129
May 26 2017 01:53
Did I miss something in the docs that says you need to program a certain number of carb ratios? I had at least one.
peterfulvi
@peterfulvi
May 26 2017 02:04
It always gets to the old pumphistory :refreshthen errors out
Starting pump-loop at Thu May 25 22:00:01 EDT 2017:
Radio ok, Listening: .No pump comms detected from other rigs
MDT cgm data retrieved
MDT cgm data reformatted
Old pumphistory: Refresh
Error, retrying
Im wondering if its a harware issue
feraridurango
@feraridurango
May 26 2017 02:23

/etc/wpa_supplicant/wpa_supplicant.conf" 8L, 111C written

interfaces(5) file used by ifup(8) and ifdown(8)

auto lo
iface lo inet loopback

auto usb0
iface usb0 inet static
address 192.168.2.15
netmask 255.255.255.0

auto wlan0
iface wlan0 inet dhcp
wpa-conf /etc/wpa_supplicant/wpa_supplicant.conf
network={
ssid="ANGLANI"
psk="9548309385"
}
~
"/etc/wpa_supplicant/wpa_supplicant.conf" 33L, 389C written

does this look right?
dan lightner
@danimaniac
May 26 2017 02:55
@rsilvers129 A few of us BT tether 100% of the time. I commemted out my wpa supplicant file. My android phone is pretty good at auto switching between Wi-Fi and cellular data and will share a wifi connection over BT. This way my rig is always BT tethered to the phone and the phone switches connections. 100% happy with that setup. I read somewhere though that an iPhone cannot share a Wi-Fi connection over Bluetooth. Only Android.
rsilvers129
@rsilvers129
May 26 2017 02:56
In case anyone finds this interesting, My rig is drawing 0.5 amps at 5 volts when plugged in.
Eric
@ecc1
May 26 2017 03:24
Is it charging the battery? It sounds too high for normal operation.
rsilvers129
@rsilvers129
May 26 2017 04:14
Yes. It always charges the battery when plugged in.
Scott Leibrand
@scottleibrand
May 26 2017 04:37
the red battery charging light in the corner goes off when the battery is fully charged. is that draw measured while that light is on or off?
JELCRAWFORD
@JELCRAWFORD
May 26 2017 05:34
@danamlewis - Can you please clarify about using bolus wizard - we have DS use BW on pump to calculate insulin. We currently do have him enter BG into the wizard - largely so we can ensure he has poked and we know current BG via poke vs. the dex reading. He knows not to do a correction based on the BW - to give the full insulin recommendation. I've always run with the assumption to use the BW in this way will work fine - implementation of the recommendation is the issue, not the entry into bolus wizard - are we correct about this? Assuming BG by poke is reasonably consistent with Dex reading, he needs to do nothing more. If it is out, we will prompt him by text to calibrate... maybe someday he'll remember to do this himself without reminder :)
Dana Lewis
@danamlewis
May 26 2017 05:35
correct
the number going in is not the issue
JELCRAWFORD
@JELCRAWFORD
May 26 2017 05:35
thx!
Dana Lewis
@danamlewis
May 26 2017 05:35
using the number to double correct for the BG that’s already being addressed by temp basals is the concern. so if you have someone who knows to dial back up or down the dose based on the carbs, that’s totally fine to use
but for anyone who wants to fly through the menus ;), I wouldn’t do the BG as there’s no point, except in cases like yours where there’s a remote caregiver who needs that info
Scott Leibrand
@scottleibrand
May 26 2017 05:37
someone who's flying through the menus is probably also just trusting the CGM data and not doing an extra fingerstick test at every meal. in that case, you don't want to enter the CGM reading into the pump, because it just creates extra work to avoid giving a correction bolus.
JELCRAWFORD
@JELCRAWFORD
May 26 2017 05:39
yup - is as we thought. Sometimes is tricky if a new caregiver who wants him to give the exact amount - so we've debated a change, but right now hanging on by our fingernails for the end of school and are trying not to change anything!
I can't speak for all the pediatric cases out there using CGM, but by our limited knowledge, it seems that the variance between cgm and poke with kids is bigger than adults, so we poke pretty much as often as when on MDI (6-8 times per day I'd say). This variance is better/less on a good sensor session, but still sometimes large...
Scott Leibrand
@scottleibrand
May 26 2017 05:42
makes sense
JELCRAWFORD
@JELCRAWFORD
May 26 2017 05:43
round of bad sensors here this week - we've had bg by cgm at 5.6 and by poke at 3.6 (100.8 - 64.8) - Dexcom told us this week that if BG is above 4.4 by cgm reading, it can be out as much as 20% and still be acceptable. Below 4.4, should not be out more than 10% (79.2)
Scott Leibrand
@scottleibrand
May 26 2017 05:58
yeah, when you're that high, being off that much still gives plenty of time for OpenAPS (or a human) to fix any corrections done based on the inaccurate CGM reading before you're in danger of going low.
JELCRAWFORD
@JELCRAWFORD
May 26 2017 06:03
2nd sensor this week - today DS had two times at school by cgm 5.5 (99) (our Openaps target btw!) but by poke 3.7(66.6). Freaks out the school - he is usually unaware of this - today was no different, so it took him by surprise. New sensor went in Monday eve - hoping we see it even out tomorrow...
Dana Lewis
@danamlewis
May 26 2017 06:03
I swear there is something in the air/water this week - my sensor went nuts this morning.
coincidence, I know.
JELCRAWFORD
@JELCRAWFORD
May 26 2017 06:04
we usually get 20-25 days - last week - 6 days.... this week not looking good. Oh well.... at least Dexcom replaces when we have numbers outside the range.
Full moon maybe? Though I think that was last week or week before..... so thats not an explanation!
Dana Lewis
@danamlewis
May 26 2017 06:09
hehe :)
Matthias
@ceben80
May 26 2017 06:11
annoying in the case all is run well but the basic environment is running out of reliability, isn't it?
JELCRAWFORD
@JELCRAWFORD
May 26 2017 06:28
maddening really :)
Rick Warren
@riwarren
May 26 2017 08:04
Can anyone tell me which chip on a medtronic 522 the firmware would be stored on?
If its the chip I am thinking it is, I may know a way to get the firmware off of it. but need confirmation as to whether I am looking at the correct chip or not.
skubesch
@skubesch
May 26 2017 09:16
Screen Shot 2017-05-26 at 5.13.33 AM.png
Question about switching from Azure to Heroku for Nightscout. The predictive blue dots are now gone so the more recent data point is showing up right in the middle of the y-axis scale making it really hard to see. Is there any fix for this?
rsilvers129
@rsilvers129
May 26 2017 09:23
When you first start looping with OpenAps, how many nights before autotune shows a suggested change? Is it the first midnight after a complete 24 hour cycle?
fidoman
@efidoman
May 26 2017 10:02
@scottleibrand - testing with latest dev this morning and SMB enabled works fine now for me with MDT without my loop modifications. I'm still getting printf error for autotune though.
fidoman
@efidoman
May 26 2017 10:17
well, I spoke too soon again. I will need to add the command back to the loop to force first CGM update: cd ~/myopenaps && openaps first-upload and now dev supermicrobolus loop stopped working again. Going back to my custom loop again
fidoman
@efidoman
May 26 2017 10:44
@Bender1061 - I tried running your command to git clone Bender1061/oref0.git and I'm not sure to tell if it worked or not ... when I run a git pull now, it just shows branch "*dev"
Gregor Schlump
@gregor_schlump_twitter
May 26 2017 10:52
just to make sure: there is no way of running openAPS with a minimed 640G?!
fidoman
@efidoman
May 26 2017 11:13
@Bender1061 - I think I'm running your branch, it is now working without the constant date errors and worked from the start without me having to customize anything yet in the loop. Autosens seems to be working now as well. I'll keep testing and see if it continues to loop without errors. So far so good.
cameronrenwick
@cameronrenwick
May 26 2017 11:21
@JELCRAWFORD I'm not sure if your child uses a xdrip+ or not. One thing I found quite interesting. I used to think the the dexcom value on my pump was the "gold standard" as it "seemed" right to me when I would check it. My finger check often differed from my dexcom and I attributed that to lousy meters. The best meter I found for consistency was the contour next one. Funny thing is when I switched from my animas vibe to the medtronic and started using my xdrip and could see raw data (no amimas/dexcom algorithm) the values my xdrip showed compared to my meter were almost identical and super consistent. It occurred to me that with the pure data from my cgm was much more accurate when it didn't use the dexcom layer. I don't know if others have compared this nor if my findings only apply to me but for the cost/sake of comparison, you may want to try; I was really surprised. To be clear, I used my amimas vibe, xdrip+ and contour next one meter all side by side for a few weeks (before I started looping) just so I was sure about what I was doing. Maybe worth a try as your dexcom/finger test values may be misleading. FWIW, my 2 cents.
Rick Warren
@riwarren
May 26 2017 11:27
@skubesch It's in the settings if you click on the menu in the top right you have to enable the predictive blue dots likely.
rsilvers129
@rsilvers129
May 26 2017 12:56
@cameronrenwick Yes, I also concluded that the Bayer Contour is the best meter. It is the most accurate according to some tests, and is small, and uses very little blood, is low-cost, and the strips are about 20 cents each. One Touch Blue strips are more expensive, less accurate, and use more blood.
Russell Gardner
@RussGG
May 26 2017 13:00
@skubesch There are three dots next to 3HR 6HR 12HR click on those and tick mark both
PopcornPam
@PopcornPam_twitter
May 26 2017 13:30
Question.. what is a lease, and what does it mean if papertrail says "no working lease"?
sdneufer
@sdneufer
May 26 2017 13:35
As to the meter discussion, I have found the contour meters work the best for me. I bring my meters when I go in for blood work and they are closest to my lab blood test consistently. My worst, one touch. So bad I complained and they sent me a new one which was just as inaccurate as the one it replaced. I currently use a contour next one and a contour next as a backup.
JELCRAWFORD
@JELCRAWFORD
May 26 2017 13:59
@cameronrenwick - I don't know if we've attended carefully to the boys raw vs. xdrip data in that way - we do use the raw data as an extra indicator of where things are going though - we'll have to watch on purpose - I feel like your sense that the raw data is more accurate to the finger poke is probably true though, based on reflection. That aside, I think the differences we are seeing this week (and last) are either sensors or our transmitter is getting tired - it's 9 months old, so could be the case. Our current rig is medtronic pump, xdrip+ running on a samsung phone, wixel/bt collecting and transmitting Dexcom CGM data to the phone, Contour next link as a meter. No Dex receiver in the mix.
PopcornPam
@PopcornPam_twitter
May 26 2017 13:59
also.. noticing that my rig can't connect to bluetooth when I am actually talking on my phone.. anyway to fix this?
JELCRAWFORD
@JELCRAWFORD
May 26 2017 14:03
@jbwittmer @danamlewis - I assume you folks know this, but just incase - re the use of Gcal with IfTTT - if your gcal is synced to your Iphone Ical, you can set your prescheduled treatments in ical app and it will upload to your gcal. I set it up as a specific calendar for the boy within Gcal - it's brilliant!
ericmtx
@ericmtx
May 26 2017 15:06
@PopcornPam_twitter that might be a function of your carrier - are you on Sprint?
cameronrenwick
@cameronrenwick
May 26 2017 15:09
@JELCRAWFORD .. that's exactly the setup I have. I wasn't 100% sure when you'd earlier posted about dexcom reading if they were from the dexom receiver thus my msg to you. On your point about the transmitter getting tired, that mirrors my finding. I just replaced my batteries a week ago and it seems to be functioning like a "younger" version. This may be just coincidence as the signal should either work or not (as in dead batteries) but I can honestly say since the change it's been more accurate.
PopcornPam
@PopcornPam_twitter
May 26 2017 15:12
@ericmtx no, tmobile
Been having a lot of connection issues today
I keep getting this on papertrail, and don't understand what it means or how to fix No working leases in persistent database - sleeping
ericmtx
@ericmtx
May 26 2017 15:15
@PopcornPam_twitter OK we're on Tmobile too - I haven't tested the BT connection while she's on the phone, but as a teenager she's never on the phone, just texting.
RedCyclist1
@RedCyclist1
May 26 2017 16:00
We installed the dev build but never got SMB fully working. If I want to go back to oref0 do I need to rerun the setup script or can I just leave it and make sure the json is correct?
JELCRAWFORD
@JELCRAWFORD
May 26 2017 16:04
@cameronrenwick - you did a DIY battery replacement? I've got some extra transmitters to play with, but haven't attempted yet. Maybe should give it a try....
Scott Leibrand
@scottleibrand
May 26 2017 16:06
@popcornpam it's referring to a DHCP lease, which is how the phone or router temporarily gives the rig an IP address to use. No lease, no IP, and theoretically no Internet connectivity, unless it keeps using the old one.
Gregor: correct, we can't set temp basals on the 640G, so it can't be used with OpenAPS.
@rsilvers129 if it's working properly, autotune should suggest changes every time it runs, based on the previous day's data. If you installed it in your loop, that should mean it automatically updates your profile at 12:05am.
rsilvers129
@rsilvers129
May 26 2017 16:10

@scottleibrand I started the system yessterday and my autotune_recommendations says ns-entries.2017-05-25.json
root@pancreas1:~/myopenaps/autotune# more autotune_recommendations.log

Parameter | Current | Autotune

ISF | 28.000 | 28.000
CSF | n/a | 0.000
Carb Ratio | 7.600 | 7.600
Basal Profile | |
00:00 | 0.625 | 0.625
00:30 | |
01:00 | |

Scott Leibrand
@scottleibrand
May 26 2017 16:10
@riwarren you might also want to join medtronic-flash/Lobby
https://gitter.im/medtronic-flash/Lobby
rsilvers129
@rsilvers129
May 26 2017 16:10
So it doesn’t seem to have changed anything in that file
Maybe it needs one more day.
Scott Leibrand
@scottleibrand
May 26 2017 16:11
You have the zero CSF bug we're trying to isolate. There's an oref0 issue for it.
rsilvers129
@rsilvers129
May 26 2017 16:12
@scottleibrand You mean me with autotune?
Scott Leibrand
@scottleibrand
May 26 2017 16:12
@davidkeddydb same for you.
@rsilvers129 yes
rsilvers129
@rsilvers129
May 26 2017 16:13
Ok. No problem. The system is working very well for me. It is on SMB also.
David
@davidkeddydb
May 26 2017 16:13
@scottleibrand Yup, I know. I put it on my spare rig so I can give you all reports of whatever you want to see what is what.
Scott Leibrand
@scottleibrand
May 26 2017 16:13
We need someone who's experiencing it to add some debugging to their autotune script and see why you get the line 72 printf errror
David
@davidkeddydb
May 26 2017 16:13
I will add anything you want. Let's do it. It is not my main rig anyway
rsilvers129
@rsilvers129
May 26 2017 16:14
@scottleibrand I can add anything if needed.
Scott Leibrand
@scottleibrand
May 26 2017 16:14
Not sure I'll be able to debug much more this morning, so hopefully someone else here can help figure it out
Otherwise I can help tonight or this weekend.
rsilvers129
@rsilvers129
May 26 2017 16:20
By the way, for everyone’s reference. My best dB before I cut the antenna trace was -72. My best after cutting the trace was -62. 10 dB is huge. That may not be an average difference though, but I think it is at least 5 dB, which is a big deal.
Scott Leibrand
@scottleibrand
May 26 2017 16:22
How many more feet away can you get and see the same rssi?
rsilvers129
@rsilvers129
May 26 2017 16:28
I will test that when I get my second board back from being repaired.
My rig lost connection last night. It was four feet from me on my nightstand and I was wearing the pump. So everything helps.
Scott Leibrand
@scottleibrand
May 26 2017 16:30
:+1:
PopcornPam
@PopcornPam_twitter
May 26 2017 16:53
@scottleibrand thanks.. is this likely an issue of the phone or the rig?
Scott Leibrand
@scottleibrand
May 26 2017 16:55
Hard to say without some more investigation. Does switching hotspot off and back on fix it?
Also, does it kick the rig offline, or is it just a warning?
PopcornPam
@PopcornPam_twitter
May 26 2017 16:56
each time i hard reset (unplug battery) it comes back on for a few minutes before stopping
I think the rig is online since it gives me thigns like this in papertrail
May 26 12:54:41 localhost dhclient: DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 6
May 26 12:54:45 localhost root: Uploaded upload/latest-treatments.json to https://pippidog.herokuapp.com
May 26 12:54:47 localhost dhclient: DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 10
So it was able to tell nightscout that I gave myself a bolus, but is not looping
rsilvers129
@rsilvers129
May 26 2017 17:00

@scottleibrand I manually ran autotune successfully. Parameter | Current | Autotune

ISF | 28.000 | 33.200
CSF | n/a | 3.674
Carb Ratio | 7.600 | 9.036
00:00 | 0.625 | 0.625
01:00 | | 0.625
02:00 | | 0.625
03:00 | | 0.736
03:30 | 0.675 |
04:00 | | 0.786
05:00 | | 0.844
06:00 | | 0.770
07:00 | | 0.812
07:30 | 0.8 |
katie disimone
@Kdisimone
May 26 2017 17:00

@PopcornPam_twitter I get that message too from the localhost part of my papertrail systems when the rig is not on our (local) home wifi. I believe, from my reading that it is an artifact that doesn’t really affect performance of the wlan0. this page was helpful https://bugs.launchpad.net/ubuntu/+source/dhcp3/+bug/190359

basically, while anna is at school her papertrail system for the rig’s name shows that it properly connects to the school’s wifi, but the localhost part in her papertrail shows that message repeating over and over as it (glitchingly) tries for the local connection (which isn’t active)

cameronrenwick
@cameronrenwick
May 26 2017 17:04
@JELCRAWFORD I have, and it's not that bad a job. If you search back to yesterday I posted the instructions. It takes some tooling and careful work but is definitely do-able. Alternatively, I'd offer to help you out and do the replacement if you'd like. PM me if you're interested.
PopcornPam
@PopcornPam_twitter
May 26 2017 17:04
@Kdisimone , that is good to know.. but I don't understand why it refuses to loop. Seems like it is able to go pump->edison->nightscout, but not the other way around
katie disimone
@Kdisimone
May 26 2017 17:05
@PopcornPam_twitter what’s the pump-loop saying?
PopcornPam
@PopcornPam_twitter
May 26 2017 17:06
That is the other issue, I just set up papertrail yesterday, but when i search pump-loop.logs all i get is stuff like this
May 25 13:10:01 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
May 25 13:11:01 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
May 25 13:12:02 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
May 25 13:13:01 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
May 25 13:16:02 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
May 25 13:17:01 localhost CRON: (root) CMD (cd /root/myopenaps && ( ps aux | grep -v grep | grep -q 'openaps pump-loop' || openaps pump-loop ) 2>&1 | tee -a /var/log/openaps/pump-loop.log)
papertrail also shows it as localhost instead of the name of the rig, which doesn't make sense to me
katie disimone
@Kdisimone
May 26 2017 17:09
@PopcornPam_twitter you have a typo on your OPENAPS_FIELDS and OPENAPS_RETRO_FIELDS lines in your heroku settings. Take out the typo. \tstatus-symbol status-label iob meal-assist rssi see the \t? that shouldn’t be there
@PopcornPam_twitter you don’t appear to have the logs for the rig itself. Somewhere you must have done a step wrong or missed one. All the logs you have posted are for localhost. YOu need the logs for the actual edisonhost name...
And you have a typo in your SHOW_PLUGINS line for openaps pump iob sage cage careportal\n delta direction upbat remove that \n
PopcornPam
@PopcornPam_twitter
May 26 2017 17:12
Where is that setting? on dashboard or the app?
katie disimone
@Kdisimone
May 26 2017 17:13
login to your heroku account. click on your app name, then under settings…click on reveal vars. change them there
PopcornPam
@PopcornPam_twitter
May 26 2017 17:13
wierd.. was no T in heroku, but was a random space
katie disimone
@Kdisimone
May 26 2017 17:14
take out the random spaces…they get translated incorrectly. you also have an extra random space on your enable line, between iob and cage
PopcornPam
@PopcornPam_twitter
May 26 2017 17:15
how can you telll?
katie disimone
@Kdisimone
May 26 2017 17:16
i’m a russian spy
PopcornPam
@PopcornPam_twitter
May 26 2017 17:16
clearly! Were you able to get that from my heroku, or papertrail? I want to learn the way of digital spies
katie disimone
@Kdisimone
May 26 2017 17:17
just kidding. if you add /api/v1/status.json to then end of a ns url you can see if there are any errors in the setup of someone’s site. you had your site url in the logs and i thought i’d double check the settings just in case that was the issue. :)
PopcornPam
@PopcornPam_twitter
May 26 2017 17:17
cool!
so is that the likely culprit of the loop messup?
katie disimone
@Kdisimone
May 26 2017 17:18
site looks fine, you were however missing your openaps pill symbols and those edits will fix that issue
looks like it’s working now
:+1:
happy looping
PopcornPam
@PopcornPam_twitter
May 26 2017 17:18
ya! Still would love to figure out why it wasn't before so I can troubleshoot in the future
katie disimone
@Kdisimone
May 26 2017 17:19
maybe that was it. one of those errors was on careportal and that is what needs to feed back-forth for openaps to loop….so maybe that was the only issue
your papertrail though…that’s a different issue
Do you see only one “system” listed in your papertrail dashboard? localhost? or is the myopenaps2 listed as well?
PopcornPam
@PopcornPam_twitter
May 26 2017 17:21
only localhost
fidoman
@efidoman
May 26 2017 17:21
@rsilvers129 - what command(s) did you use to manually run autotune successfully?
rsilvers129
@rsilvers129
May 26 2017 17:22
@ericfeibelman First, make sure you have the latest version of oref0: npm list -g oref0 | egrep oref0@0.4.[0-9] || (echo Installing latest oref0 package && sudo npm install -g oref0)
Install jq: sudo apt-get install jq
Make two copies of your profile.json, one to be the starting point for autotune, and one to provide the pump baseline for enforcing the min/max limits: cd ~/myopenaps/settings/ && cp profile.json autotune.json && cp profile.json pumpprofile.json
Run oref0-autotune --dir=~/myopenaps --ns-host=https://mynightscout.azurewebsites.net --start-date=YYYY-MM-DD (obviously, sub in your NS url and the start date you want to start with. Try 1 day first before moving on to 1 week and 1 month to better troubleshoot).
fidoman
@efidoman
May 26 2017 17:24
@rsilvers129 - I did that procedure the other day and still got the printf error. Did something change to make it work in latest dev? I updated to latest dev this morning.
rsilvers129
@rsilvers129
May 26 2017 17:24
For start date I put yesterday, if that helps.
JELCRAWFORD
@JELCRAWFORD
May 26 2017 17:28
@cameronrenwick - thanks for the share of experience w/dex data and the batt replacement. I may take you up on that, but first am going to give it a go - as Dana has said before - teach a man to fish..... well in this case a psychologist, So I'll give it a go - I'm sure @jmcrawford will be very pleased to add another of my 'little projects' to the list! It's very helpful to know that a symptom of a tired transmitter may be greater variance between readings and poke. I'm on phone for git right now. Will search for your post from yesterday when home from work tonight. cheers and thx again.
Matthias
@ceben80
May 26 2017 17:28
@scottleibrand as you know the edison voltage cut-off cron line has a threshold of 3050mV, I think this needs to be lift up e.g. 3200mV to shut down the rig (I did this already in my cron job), today my rig reached this threshold and was going down properly, here a screenshot of 4% battery voltage with 3115mV. If you would like, I will make a PR to change the voltage from 3050mV to 3200mV :question: :exclamation:
image.png
fidoman
@efidoman
May 26 2017 17:37

@ericfeibelman - didn't work for me ... I just did a single day. p50deviation: 1.475 p50BGI -2.725 p50ratios: 1.069 Old ISF: 46 fullNewISF: 46 adjustedISF: 46 newISF: 46
totalMealCarbs: 129 totalDeviations: 500.2199999999999 fullNewCSF: 3.88 newCSF: 3.838

Autotune pump profile recommendations:

Recommendations Log File: /root/myopenaps/autotune/autotune_recommendations.log

/usr/local/bin/oref0-autotune-recommends-report: line 72: printf: null: invalid number

Parameter | Current | Autotune

ISF | 46.000 | 46.000
CSF | n/a | 0.000
Carb Ratio | 12.000 | 12.000
00:00 | 0.650 | 0.650
04:00 | 0.800 | 0.800
10:00 | 0.650 | 0.650
15:00 | 0.800 | 0.800
18:00 | 0.650 | 0.650

Dana Lewis
@danamlewis
May 26 2017 17:41
@ericfeibelman we need someone to help dig in - it's calculating correctly but not printing correctly. I created an oref0 issue for it if anyone wants to poke around and help figure it out before Scott or I have more time this weekend to do so
It's something related to dev branch and SMB setups
rsilvers129
@rsilvers129
May 26 2017 17:43
I am running dev and smb and if you need me to insert debugging code, just provide a file and I will overright the existing one.
overwrite I mean
fidoman
@efidoman
May 26 2017 17:54
I am also running dev and SMB and can insert debugging code.
rsilvers129
@rsilvers129
May 26 2017 18:46
Wanted to confirm.. If I remove my pump tp swim or shower, I am supposed to suspend it so that the rig can accuratly know I didn’t get the insulin from that period?
rsilvers129
@rsilvers129
May 26 2017 18:51
@danamlewis Thanks. Makes sense.
skubesch
@skubesch
May 26 2017 20:10
That fixed it, thanks @RussGG!
mrshapir
@mrshapir
May 26 2017 20:34
Hi, I'm trying to set up my Edison and running into an issue. Getting the "cant find a PTY" error regardless of which side of the computer my USB cords are on. Any ideas?
Lynne Butts
@flutelynne_twitter
May 26 2017 20:35
I just received the 2000 battery from robotshop that was linked from the docs, thinking it would be a smaller rig. Does this look right to you?
FullSizeRender.jpg
rsilvers129
@rsilvers129
May 26 2017 20:52
@danamlewis So removing my pump for 30 minutes didn’t go well. I must have had no IOB as I shot up so fast that the DexCom got confused and gave ???. Maybe I should have given myself a small bolus to last the duration of the shower?
IMG_6260.PNG
katie disimone
@Kdisimone
May 26 2017 20:53
@flutelynne_twitter I believe the small 2000 battery is from adafruit. Not robotshop?
@rsilvers129 have you checked that the cgm is right? My daughter gets false jumps from showers.
@mrshapir Restart computer. That will fix the can't find a pty error
rsilvers129
@rsilvers129
May 26 2017 20:54
@Kdisimone I did two blood tests and also look at the Eversense. All were over 200.
IMG_6259.PNG
This was before the shower.
0.87 IOB and 0.0 basal.
Lynne Butts
@flutelynne_twitter
May 26 2017 20:57
@Kdisimone thanks. They are out of stock at adafruit, but maybe worth waiting for?
katie disimone
@Kdisimone
May 26 2017 20:58
I use the 2500. The small size was just not enough of a benefit for us to lose the capacity.
rsilvers129
@rsilvers129
May 26 2017 20:58
You can get this 2500 if you want. It lasts a whole day. https://www.amazon.com/gp/product/B01NAX9XYG/ref=oh_aui_search_detailpage?ie=UTF8&psc=1
If I could only have one, I would get the 2500.
Lynne Butts
@flutelynne_twitter
May 26 2017 21:10
Thanks. I do have another 2500. Just looking for a little smaller.
mrshapir
@mrshapir
May 26 2017 21:22
Thanks @Kdisimone , restarting worked once but when I exited terminal and tried to login again, I'm just getting a white screen forever and not asking for a password
ericmtx
@ericmtx
May 26 2017 21:24
speaking of batteries, anyone know the part numbers to enable wireless charging of the lipo/explorer board combo? I have the battery but having to pull out the rig from the tic-tac case to put the USB charge in each night. I saw a pic of someone using a wireless charging setup inside the tic-tac case with their rig
katie disimone
@Kdisimone
May 26 2017 21:24

Yes. That's the way console/screen mode works. If you lose the connection, it's a reboot to get the USB port cleared again to use screen more again.

So, reboot again. That why most of us use ssh to login to rig once we are done getting it setup.

@mrshapir ^^
rsilvers129
@rsilvers129
May 26 2017 21:33
@ceben80 That battey doesn’t show a connector in the photo. You may want a JST Connector on it.
Dana Lewis
@danamlewis
May 26 2017 21:34
@ericmtx a couple folks tried QI charges (@tim2000s, others) but backed away from it - heat, etc. were issues
Matthias
@ceben80
May 26 2017 21:34
@rsilvers129 np, just soldering a suitable :wink:
mrshapir
@mrshapir
May 26 2017 21:34
@Kdisimone that makes sense. I've tried rebooting the computer a bunch of times and what happens is I try to start screen mode, get to white screen, press enter a few times and wait and still no login :(
Dana Lewis
@danamlewis
May 26 2017 21:35
@rsilvers129 in general, yes, depending on how long you’re going to be off and netIOB going into a long shower/swim, you may decide to bolus up front for missing iob later. however, that one seems to be a faster rise than would be correlated with the basals there, so there might have been some other physiological impact going on
ericmtx
@ericmtx
May 26 2017 21:36
@danamlewis ah good to know. She sleeps with it by her head so I'll look closer and just getting a nice 3dcase printed and make it easier to access the charge port
Dana Lewis
@danamlewis
May 26 2017 21:36
ooh yea, definitely wouldn’t if it’s near the body. a case with access to the charge port is :+1:
rsilvers129
@rsilvers129
May 26 2017 21:50
@danamlewis Agreed that rise is something I am not used to seeing.
rsilvers129
@rsilvers129
May 26 2017 22:03
Dexcom blackouts cause problems for me sometimes. If I knew I was rising I could have mitigated it.
Scott Leibrand
@scottleibrand
May 26 2017 22:18
If you're on G4 you might look into using G4-upload to display raw CGM data in NS (and loop off it).
rsilvers129
@rsilvers129
May 26 2017 22:30
Oh wow. Really? No ??? I have a G4 receivers and some transmitters. Maybe I should.
Dana Lewis
@danamlewis
May 26 2017 22:31
g4-upload + plug in receiver and second battery = raw in NS if you enable raw to always show (or show when noisy), plus raw for OpenAPS to loop from (up til 160 or whatever level we put in the code).
rsilvers129
@rsilvers129
May 26 2017 22:33
That is very interesting. Does raw mean it is uncalibrated?
Scott Leibrand
@scottleibrand
May 26 2017 22:55
no, what's actually displayed is more like "unfiltered". we use the last calibration record from the Dex
the uncalibrated raw data is in there as well, if you want to play with it
so if you start a new sensor, for the first two hours the unfiltered data displayed in NS will be based on the calibration from the last sensor, which is helpful (and in our experience usually safe to loop off of, because it's directionally accurate) but the actual BG levels are often quite inaccurate until your first calibration
rsilvers129
@rsilvers129
May 26 2017 23:01
If your IOB is greater than the max_iob, does that mean the rig turns basal to 0.0?
Scott Leibrand
@scottleibrand
May 26 2017 23:04
I believe it will run neutral basal at that point.
but in practice it's very unusual for your basal IOB to actually get all the way up to max_iob, as it sets lower and lower high temps that won't exceed it.
rsilvers129
@rsilvers129
May 26 2017 23:28
I am thinking of the IOB display on NS, which seems to be bolus IOB. So you are saying max_iob is specifically just for basal IOB? I just don’t understand the max_iob setting.
Dana Lewis
@danamlewis
May 26 2017 23:29
max_IOB is how much IOB will be given by the loop
so netIOB which is the IOB display in NS, is going to be from any boluses you do, plus the IOB OpenAPS does (+ or -)
max_IOB is a setting we introduced into OpenAPS to restrict how much it can do
your actual netIOB is unrelated to max_IOB - max_IOB just impacts how much OpenAPS will do at any point in time depending on what it thinks you need + how much room it has between where it is and the max_IOB cap
keep in mind, like the COB pill, the IOB pill may be out of sync if your rig isn’t uploading. so always best practice to check your logs or check inside the OpenAPS pill to see what you actually have on board
Dana Lewis
@danamlewis
May 26 2017 23:35
@rsilvers129 ^
rsilvers129
@rsilvers129
May 26 2017 23:40
I see. So if I were to take the strategy of having bolus only giving myself 75% of what I think I need for the carb, then I would need to raise max_iob to cover that other 25%. So if the max meal I wanted to be able to deal with was 120 grams of carb, then I would need insulin for the remaining 25% (30 grams). If my carb ratio is 9, then I would need a max_iob of 3.3.
I think there was some pill or some NS variable we are not supposed to turn on because it messes something up.
Dana Lewis
@danamlewis
May 26 2017 23:41
Keep in mind those carbs are going to hit over time - so you might end up getting 3 extra units of insulin from the rig (or more, or less, depdning on how absorption actually goes), but that’s going to be phased in over time.
yea, COB pill messes stuff up. But just be aware for any non-OpenAPS pill showing info, you’re at the mercy of NS working perfectly, and there could always be a delay in showing an update from what the rig is doing, so the logs/OpenAPS pill are always the most correct and up to date with what’s happening
rsilvers129
@rsilvers129
May 26 2017 23:44
But in that scenero, I would be counting on those three extra units, because I only bolused for what I needed minus three units.
Dana Lewis
@danamlewis
May 26 2017 23:45
I wouldn’t derive maxIOB by how much you are expecting it to give for any particular meal, because that’s going to vary a lot
but you certainly will want to experiment with what portion of a meal bolus you want to do vs. what you want OpenAPS to fill in on; and that may vary from basal only (oref0) to SMB (with oref1); and vary also over time with your experiences with low and how sensitive you are to other changes and how much extra insulin you are comfortable having on board at any given time that you haven’t chosen to give yourself manually
you’ll also want to think about maxIOB in context of your basals and correction factor, too
someone with tiny basals is going to have a different approach to maxiob to someone who has 3x or more basals
rsilvers129
@rsilvers129
May 26 2017 23:55
My basals are around 0.65 to 0.9
My ISF is 33
mrshapir
@mrshapir
May 26 2017 23:59
Still can't get my Edison to the login page through terminal on my OSX. I get to the white screen, press enter and wait and nothing happens. Then quit terminal, try again and end up with the cant find PTY error and have to restart computer. Am I doing something wrong?