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

14th
Dec 2017
Chris Oattes
@cjo20
Dec 14 2017 00:03
usually cd myopenaps
Brian Cadre
@bcadre
Dec 14 2017 00:05
@cjo20, @danamlewis, @scottleibrand: Thanks for your help; I figured it out and NS appears to be up and running normally.
Brian Cadre
@bcadre
Dec 14 2017 02:15
@Diadon81 , @cjo20 I am looking forward to your correction of the date errors. I presume this may also be why my SAGE is 12/31/1969 and the CAGE is n/a. But, since I ran openaps first-upload we are successfully looping and NS is capturing data. Thank you @scottleibrand and @danamlewis and everyone here! Let the tweaking begin...
Dana Lewis
@danamlewis
Dec 14 2017 02:26
Sage and cage have to be manually set
So try inputting some data for that if you haven't via care portal
Also happy looping :smile:
Ebgineer
@Ebgineer
Dec 14 2017 02:30
Is there a way to reset the explorer board USB interface via linux shell? Rebooting works, but seems like a hack-y windows style workaround.
Scott Leibrand
@scottleibrand
Dec 14 2017 02:33
You mean reset?
Ebgineer
@Ebgineer
Dec 14 2017 02:40
I'm not sure. when I enter reset at the command line the only thing I see it doing is clearing the screen. lsusb still returns -99. How would I check what reset executes?
I'm looking for a way to force the USB services to shutdown and reinitialize like they do when booting. Haven't found much documentation on the jubilinux USB driver.
Aaron Michelson
@Aaybob
Dec 14 2017 02:54
@garykidd yes we did try other timings. We had loops that took longer. But if we don't step on the previous loop why not loop again to wait for a new glucose input from the CGM. This is a 'polling' method as opposed to an 'interupt' method. If we received an intrerupt we'd get immediate action upon new data. I tried every 15 minutes loop cycle and there was delayed action. Pretty obvious the faster the action the better because of the pharmaceutic delay. And if there is no variation in the blood sugars 15 minute cycle is just as good. But sometimes one wants the rig to act as soon as possible. Sometimes not allways.
Scott Leibrand
@scottleibrand
Dec 14 2017 03:26
@Ebgineer ah, ok. I misunderstood.
Ebgineer
@Ebgineer
Dec 14 2017 03:28
I'm thinking it is unbind/bind in /sys/bus/pci/drivers/xhci_hcd for lspci | grep USB which yields 00:11.0 USB controller: Intel Corporation Device 119e (rev 01)
but the 00:11.0 is returning a No such device error
Paul Dickens
@thebookins
Dec 14 2017 03:36
Is there any way to override autosens from NS? Kid is running high, but stuck on an a sensitivity ratio of 0.85, which doesn't want to budge.
Scott Leibrand
@scottleibrand
Dec 14 2017 03:38
Depends what preferences you have on. There is one that flags resistance with a low temp target.
Paul Dickens
@thebookins
Dec 14 2017 03:45
This one: low_temptarget_lowers_sensitivity ?
We don't have it enabled.
On a related note, I know adv_target_adjustments is not recommended with 0.6.x, but I'm convinced that our kid needs more proportionally to bring him down from a high > 10 or 11.
i.e. a bg-dependent sensitivity.
is there anything equivalent to adv_target_adjustments to play with in 0.6.x?
Scott Leibrand
@scottleibrand
Dec 14 2017 03:51
you can try adv_target_adjustments if the default 0.6.x behavior is insufficient
Scott Leibrand
@scottleibrand
Dec 14 2017 04:03
make sure you have UAM on too
and if you'd like to troubleshoot why the sensitivityRatio of 0.85, take a look at the autosens-loop.log once he gets home
Paul Dickens
@thebookins
Dec 14 2017 04:48
Thanks Scott. UAM is on. He is home now - having a look.
what are we looking for? this long string of ++++----xxxxx etc?
Paul Dickens
@thebookins
Dec 14 2017 04:54
here's the latest autosens run:
Starting oref0-autosens-loop at Thu Dec 14 15:32:02 AEDT 2017:
Calculating sensitivity using 8h of non-exluded data
u(x34g(xx16h=47gxxxxxxxxxxxx17hxxxxxxxxxxxx18h=xxx)u(x15g(xx19h24gxxxxxxxxx20h=20gxxxxxxxxxxxxx)u(x)------22h=--------+---23h-9g(xxxxxx)u(xxxxx0h=x)-----------1h-----------+2h=+------++---3h------+----+4h=--+----+----5h--++++++-+-6h=-----+u(x40g(x79gx7hxxxxxxxxxxxx8h=xxxxxxxxxxx9hxxxxxxxxxx11gx10h=xxxxxxxxxxxx11hx32gxxx34gxxxxxxxx12h=xxxxxxxx0gx)--13h--19g(xxxxxxxxxx14h=xxxxxxx20gxxxxx15h21gxxxx28gx
Using most recent 96 deviations since 2017-12-13T04:32:05.705Z
19% of non-meal deviations positive (>50% = resistance)
73% of non-meal deviations negative (>50% = sensitivity)
Insulin sensitivity detected: ISF adjusted from 99 to 116
Calculating sensitivity using all non-exluded data (up to 24h)
u(x34g(xx16h=47gxxxxxxxxxxxx17hxxxxxxxxxxxx18h=xxx)u(x15g(xx19h24gxxxxxxxxx20h=20gxxxxxxxxxxxxx)u(x)------22h=--------+---23h-9g(xxxxxx)u(xxxxx0h=x)-----------1h-----------+2h=+------++---3h------+----+4h=--+----+----5h--++++++-+-6h=-----+u(x40g(x79gx7hxxxxxxxxxxxx8h=xxxxxxxxxxx9hxxxxxxxxxx11gx10h=xxxxxxxxxxxx11hx32gxxx34gxxxxxxxx12h=xxxxxxxx0gx)--13h--19g(xxxxxxxxxx14h=xxxxxxx20gxxxxx15h21gxxxx28gx
Using most recent 111 deviations since 2017-12-13T04:32:14.348Z
17% of non-meal deviations positive (>50% = resistance)
73% of non-meal deviations negative (>50% = sensitivity)
Insulin sensitivity detected: ISF adjusted from 99 to 115
Using 8h autosens ratio of 0.85 (ISF 116)
Autosens refreshed: {"ratio":0.85}
Completed oref0-autons-loop at Thu Dec 14 15:32:19 AEDT 2017
I'm assuing the 'x's mean that we've no deviations during the day because of carbs getting in the way?
Scott Leibrand
@scottleibrand
Dec 14 2017 04:56
that's a lot of -'s. yeah, x is for excluded
was it low-temping all night?
Paul Dickens
@thebookins
Dec 14 2017 04:57
no funnily-enough, it wasn't
i'll post a pic
image.png
if anything it was high-temping more
Scott Leibrand
@scottleibrand
Dec 14 2017 04:59
what was IOB at a few typical points overnight?
I see you did have to do rescue carbs before midnight after a long slide down...
Paul Dickens
@thebookins
Dec 14 2017 05:01
true, that was unusual, we hardly ever have to do rescue carbs at night
Scott Leibrand
@scottleibrand
Dec 14 2017 05:03
anyhow, you might want to turn on that resistance mode preference if you want your low temp targets to work more strongly (and counteract any autosens-detected sensitivity that you know is no longer there)
Paul Dickens
@thebookins
Dec 14 2017 05:04
according to the openaps pill iob was -0.75 at midnight, -0.07 at around 8:40 pm
practically zero at 3 am
Scott Leibrand
@scottleibrand
Dec 14 2017 05:07
what was sensitivityRatio then?
Paul Dickens
@thebookins
Dec 14 2017 05:07
OK, I'll enable low_temptarget_lowers_sensitivity. Is there a write-up on what is does exactly?
sensitivityRatio was 0.91 at 3 am
Scott Leibrand
@scottleibrand
Dec 14 2017 05:09
if you set a target of 90, it sets sensitivityRatio to 1.2. with 99 it does 1.01 or so (so basically just cancels any autosens sensitivityRatio)
so based on a 91% basal, you had basically zero IOB. if you recalculated IOB based on a 100% basal, IOB would have been negative
(but only by approximately 10-20% of hourly basal, probably)
Paul Dickens
@thebookins
Dec 14 2017 05:11
so that explains the - s in the autosens log?
Scott Leibrand
@scottleibrand
Dec 14 2017 05:12
probably. it wasn't very negative, but just slightly (and consistently)
Paul Dickens
@thebookins
Dec 14 2017 05:12
is the sensitivityRatio capped at 1.2? So If i set a target of 80, I still get 1.2? Or is that governed by preferences.json?
Scott Leibrand
@scottleibrand
Dec 14 2017 05:15
it's your autosens_max
Paul Dickens
@thebookins
Dec 14 2017 05:16
thanks for your help - good to know I can disable a low sensitivityRatio using NS now. I guess we are stuck with all the xs effectively preventing autotuning during the day.
Scott Leibrand
@scottleibrand
Dec 14 2017 05:16
if it's higher than 1.2, an 80 target could give you up to 1.5x sensitivityRatio
Paul Dickens
@thebookins
Dec 14 2017 05:16
Unless Oli doesn't eat :smile:
OK, so that's a pretty strong effect then.
Is this all in determine-basal.js?
Scott Leibrand
@scottleibrand
Dec 14 2017 05:17
when autosens detects sensitivity, you basically need to wait for the next night for it to adjust that. but if you set high temp targets, autosens will use those as a sensitivity signal. so corrective action to any daytime sensitivity will cause detected resistance to decay pretty fast
yeah, search for the preference names to find the relevant code
Paul Dickens
@thebookins
Dec 14 2017 05:18
wil do, thx
Scott Leibrand
@scottleibrand
Dec 14 2017 05:18
the "use high temp targets as a sensitivity signal" logic is in the autosens lib
and there are PRs that describe the behavior too, but maybe not much better than the code comments
Paul Dickens
@thebookins
Dec 14 2017 05:21
OK, got it
Just one other thing: the sensitivityRatio doesn't scale carb ratio too does it?
Reason being, oref0 zero-temped for a long time after breakfast bolus today
I'm assuming because the 0.85 was causing the predictions to trend down
I just re-read your earlier comment. So is the converse true: "if you set low temp targets, autosens will use those as a resistance signal"?
Or does it only work in one direction?
Dana Lewis
@danamlewis
Dec 14 2017 05:29
Depends on which preferences you've enabled
Can go both ways if you enable everything
Scott Leibrand
@scottleibrand
Dec 14 2017 05:30
No, the autosens effect is only one way.
Autosens-set sensitivityRatio does not affect CR. Manually set ones via temp targets do.
Paul Dickens
@thebookins
Dec 14 2017 05:32
Righto, good to know.
DaisyHK
@laurathiessen
Dec 14 2017 06:54
@scottleibrand I am no having an issue with Bt auto thethering when I lave the house. It use to work fine until I had to switch to master branch and re-eintstlall oref. Rig worked fine at home on wifi but wont connect to my phone when i leave the house now. I double checked using the docs instructions, it was able to turn on the BT sharing on my phone, but when I went to double check and turned off wifi it would not automatically connect. I have the app BT auto tether enabled and the phone and it is successfully paired with the rig. Phone is a oneplus 5T
image.png
Samuchco
@Samuchco
Dec 14 2017 07:19
NS is telling me that
OpenAPS is not looping, but it is. Based on papertrail and the temp basal pill, and the temp basal on pump
DaisyHK
@laurathiessen
Dec 14 2017 07:21
@Samuchco if you look on facebook there are quit a few posts on that issue... cant remember how to resolve it though....
Samuchco
@Samuchco
Dec 14 2017 07:22
I can see recent SMB s on NS. Yet pump pill shows nothing and OpenAPS pill shows not connected for two hours
@laurathiessen thanks. Never looked for issues on FB. How do you search?
DaisyHK
@laurathiessen
Dec 14 2017 07:24
@Samuchco i am trying right now to find some posts for you... on facebook go to the goup on the left side there is a searc bar to type in words to find in posts
DaisyHK
@laurathiessen
Dec 14 2017 07:33
@Samuchco I am sorry but I cant find your issue specifically on facebook and I gotta head off to bed. Either use the search took or wait for someone else on here to answer you could also look at the docs with rig--Ns issues to see if you can reolve it http://openaps.readthedocs.io/en/latest/docs/Troubleshooting/Rig-NS-communications-troubleshooting.html
Samuchco
@Samuchco
Dec 14 2017 07:35
@laurathiessen I found it. Thank you. You were very helpful
Fixed
Colin Lennon
@colinlennon
Dec 14 2017 12:31
@Samuchco I have the same issue, could you tell me how you fixed it please? I deleted all documents from the devicestatus collection via Admin Tools, but that hasn't fixed it. Thanks
Samuchco
@Samuchco
Dec 14 2017 16:43
@colinlennon That is what I did. Needed to wait for NS to update. Go to http://openaps.readthedocs.io/en/latest/docs/Troubleshooting/Rig-NS-communications-troubleshooting.html and scroll to mLab maintenance and follow the directions
Colin Lennon
@colinlennon
Dec 14 2017 17:17
Thanks @Samuchco
Brian Cadre
@DrCadre_twitter
Dec 14 2017 18:35
To the community: We looped successfully for a few hrs last night (1st time!). Worked well. However, when I tried to add another WiFi network, rig would no longer connect to any network. I removed the added network from the file (leaving the original), rebooted, but it still won’t connect. Maybe I corrupted the file? Is there a way to selectively wipe then re-add the original WiFi file or do I need to re-run the entire install? Help please...
egirard
@egirard
Dec 14 2017 18:41
cannot access settings/profile.json
(oops - meant to search on that;)
Cas Eliëns
@cascer1
Dec 14 2017 18:48
Hey guys, I found a weird issue in my Nightscout graph. I just did a meal bolus with the bolus wizard on my pump but didn't follow the suggested dose (suggestion was 11.4 units, I took 7.4). However, the bolus circle in the graph still shows 11.4 units, and the hover text is also weird: https://ducohosting.com/screenshots/deskc2264b8a7aa147e9c80208cf08a0658d.png
The IOB seems to still use the acutal 7.4 bolus information, so I'm hoping it's just some issue sending the data from my rig to nightscout
alimhassam
@alimhassam
Dec 14 2017 19:24
Small question, just noticed I can set different BG targets in my pump at different times of day. What happens if I use that does it use the different targets?
Dana Lewis
@danamlewis
Dec 14 2017 19:24
yes
alimhassam
@alimhassam
Dec 14 2017 19:25
Cool! Thanks! I'll try it.
Brian Rabinovitz
@bdr1177
Dec 14 2017 20:33
We're getting the Could not get subg_rfspy state or version error almost every day now. I can always fix it by flashing the radio chip with the ccprog, but I'm not sure why this is happening every day now. I've tried the other troubleshooting steps listed in the docs (here: http://openaps.readthedocs.io/en/master/docs/Resources/troubleshooting.html?highlight=ccprog#could-not-get-subg-rfspy-state-or-version-have-you-got-the-right-port-device-and-radio-type), but haven't had any luck. Does anyone have any other ideas?
Dave Acklam
@dcacklam
Dec 14 2017 21:21
@bdr1177 What pump model are you using?
And what does mmtune output?
@cascer1
You are not the only one who sees wierdness with the wizard...
Mine takes Wizard boluses & records them as 'normal bolus, no wizard' - stripping any carbs input off & only reporting the insulin.
In other news... Something came to mind when I was at the doc's office today...
They tried to pull historical data off my pump, and were having problems doing that while it was suspended with a zero-temp-basal active...
I have also seen OpenAPS have 'issues' pulling Old-pumphistory during and after a suspend-while-zeroed state - even now that my pump correctly reports suspended/not-suspended due to command-code changes for the status report...
Brian Rabinovitz
@bdr1177
Dec 14 2017 22:14
@dcacklam pump is 522. Been looping for about 6 months with this particular pump. This error has become much more frequent in the past few weeks. Mmtune is generally good, but once I get the error I need to reflash radio before it will run mmtune again.
Dave Acklam
@dcacklam
Dec 14 2017 22:32
@bdr1177 Well, that doesn't sound like what I was experiencing (my firmware-dumps got better when I fixed mmtune for my model of pump - but I was getting all-99s before that fix)...