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

15th
Feb 2018
andyhandy
@andyhandy
Feb 15 2018 04:26
Hi All - how do I go about changing the default values in "eating soon" and "activity" modes. Thanks.
Dana Lewis
@danamlewis
Feb 15 2018 04:29
In Nightscout care portal drop down, do you mean? Or something else?
Dave Acklam
@dcacklam
Feb 15 2018 04:31

@scottleibrand
I get that...
In my specific case (Running UAM with no meal announcement, and a very-low non-temp basal value) it has the side effect of delivering substantially slower than I'd personally prefer....

What I am suggesting, is that when UAM is active, the preferences.json value for minutes-of-basal (which defaults to 30 anyway) should be used even-if IOB seems greater than present carbs-in-effect....

If that is obviously unsafe, then don't do it... But it seems that keying off UAMduration would provide the safety against non-meal-related false-highs (still using 30min override if UAMduration == 0), while also being more responsive to increases triggered by actually eating food...
Dave Acklam
@dcacklam
Feb 15 2018 04:39

On another note: Bluetooth issue...

The pi zero's bluetooth does NOT stay in scanning mode perpetually & thus on Pi, the following code block results in bluetoothd being CONSTANTLY reset (such that BT-PAN network connectivity is impossible):

if ! ( hciconfig -a | grep -q "PSCAN" ) ; then
killall bluetoothd
/usr/local/bin/bluetoothd --experimental &
fi

With this removed, bluetooth is more stable - although it still flaps through 'up, down, back-up' constantly...

Rick Warren
@riwarren
Feb 15 2018 04:44
I haven't used my openaps rig in several months. I have it up and running now....what would be the most direct way to update to the latest version?
Dana Lewis
@danamlewis
Feb 15 2018 05:05
@riwarren http://openaps.readthedocs.io/en/latest/docs/Customize-Iterate/update-your-rig.html I would also read the release notes to compare changes between the last version you were running and current master (0.6.0)
Scott Leibrand
@scottleibrand
Feb 15 2018 05:15
@dcacklam ok, sounds like you're using oref0 in a way that we sorta designed it for, but then backed off of for safety and never fully optimized for. one thing you might test for is the response to unannounced rescue carbs. if you're drifting low, and do a small quantity of fast carbs to correct, does it end up doing SMBs that more than cover the rescue carbs, resulting in a subsequent low? it did when we originally put that limit in place, but perhaps some of the more recent UAM changes (particularly using slopeFromMinDeviation to limit UAM Duration) adequately mitigate that now.
if you find that it's always safe to use the maxSMBBasalMinutes, and that UAM never overcorrects, then we can probably add some sort of preference to enable that behavior
re: bluetooth, do you know what's different about raspbian that prevents it from keeping bluetooth in PSCAN?
Dave Acklam
@dcacklam
Feb 15 2018 05:18
I'll keep testing the UAM situation...
As-for PSCAN, all I can find from googling is that it is a 'known issue'.
It goes into 'noscan' shortly after boot.
andyhandy
@andyhandy
Feb 15 2018 05:19
@danamlewis yes in the care portal drop down but also interested to know whether you can
Dana Lewis
@danamlewis
Feb 15 2018 05:20
I believe you can.
Dave Acklam
@dcacklam
Feb 15 2018 05:22
The PSCAN thing is both with the Pi's native BT (on both Pi3 and ZeroW) and with a usb dongle)
andyhandy
@andyhandy
Feb 15 2018 05:22
Whether you can change lowest values openaps (autosens) will use when lowering targets when high
To make it less aggressive
Scott Leibrand
@scottleibrand
Feb 15 2018 05:23
@dcacklam k. we also have a few folks testing Pi based rigs in https://gitter.im/openaps/hardware-dev so I gave them a heads up to see if we can figure out the best way to make it work with both platforms
@andyhandy can you be more specific on what you mean by "change lowest values openaps (autosens) will use when lowering targets when high"?
Dave Acklam
@dcacklam
Feb 15 2018 05:27
@scottleibrand
Perhaps a platform.ini file, that can be .-ed into various scripts for use in an if-block (if edison, then this-way, elif pi, then that-way, fi)...
Scott Leibrand
@scottleibrand
Feb 15 2018 05:28
we have better ways of detecting platform: see oref0-setup in either dev and/or go-explorer-hat branches
Dave Acklam
@dcacklam
Feb 15 2018 05:28
Got it...
Scott Leibrand
@scottleibrand
Feb 15 2018 05:28
so I'm fine with a platform conditional if that's what would be best
but sounds like even with the PSCAN stuff commented out it's still flapping, so want to try to figure out how to make it work reliably
Dave Acklam
@dcacklam
Feb 15 2018 05:28
That is what I'm thinking, but obviously the rest of the pi users should weigh in.
Scott Leibrand
@scottleibrand
Feb 15 2018 05:28
I haven't really tested BT-PAN with the Pi rigs yet
andyhandy
@andyhandy
Feb 15 2018 05:29
@danamlewis @scottleibrand sorry I meant adv_target_adjustments which I understand from the docs will set the targets as low as the eating soon mode
Scott Leibrand
@scottleibrand
Feb 15 2018 05:29
ah, k. FWIW we don't recommend using that any longer.
Dave Acklam
@dcacklam
Feb 15 2018 05:29
It's flapping, but it is flapping less.
With the PSCAN test in there, you are killing bluetoothd on Pi every single time oref0-bluetoothup runs.
Scott Leibrand
@scottleibrand
Feb 15 2018 05:29
yeah, removing that sounds like a good start
Rick Warren
@riwarren
Feb 15 2018 05:30
date: invalid date ���@���
???
Scott Leibrand
@scottleibrand
Feb 15 2018 05:31
@andyhandy now that oref0 can predict how low BGs will go with a zero temp (ZTpredBGs) it can more safely bring down highs that way, so adv_target_adjustments is now off by default
@riwarren that's usually a symptom of something else. what other errors do you have in your logs?
Rick Warren
@riwarren
Feb 15 2018 05:31
rm: cannot remove ���enact/smb-enacted.json���: No such file or directory
Warning: clock input Invalid Date is unzoned; please pass clock-zoned.json instead
and meal.json
I just rebooted after updating though. so it could be erroneous
Scott Leibrand
@scottleibrand
Feb 15 2018 05:33
the lack of an smb-enacted.json just means it hasn't looped yet since the upgrade
the Warning: clock input Invalid Date is unzoned; please pass clock-zoned.json instead is normal and can be ignored if you use MDT CGM (Enlite)
Rick Warren
@riwarren
Feb 15 2018 05:34
k I'll give it a bit before I start looking for problems :) yes I use MDT unfortunately. as if there is anything else to use in Canada sigh
andyhandy
@andyhandy
Feb 15 2018 05:34
@scottleibrand got it thanks. What about changing the defaults in NS in the careportal drop down? Thanks
Scott Leibrand
@scottleibrand
Feb 15 2018 05:34
ok, give it 20m, and then paste an entire pump-loop.log run from Starting to Completed if it's still not working and you can't figure it out.
@andyhandy changing which default from what to what?
Rick Warren
@riwarren
Feb 15 2018 05:34
oh its working.
andyhandy
@andyhandy
Feb 15 2018 05:37
@scottleibrand changing default (pre-populated) temp target range values for "eating soon" and "activity" modes
Scott Leibrand
@scottleibrand
Feb 15 2018 05:38
k. are you wanting to change those in your own fork of NS, or thinking the defaults should be changed for everyone?
andyhandy
@andyhandy
Feb 15 2018 05:39
Only for me
andyhandy
@andyhandy
Feb 15 2018 05:40
I find they are a bit too aggressive for my son
Terrific thanks
johnaras98
@johnaras98
Feb 15 2018 06:52
0
4
PeetuHongisto
@PeetuHongisto
Feb 15 2018 07:03
raise AlreadyInUseException("%s already in use" % self.device)
mmeowlink.vendors.serial_interface.AlreadyInUseException: /dev/ttyMFD1 already in use
Setting up edison + breakout board + ti-usb stick and I get this
PeetuHongisto
@PeetuHongisto
Feb 15 2018 08:25
Never mind! I broke my breakout board :D I guess I just have to wait for the explorer board to get here
Eric
@ecc1
Feb 15 2018 14:47
@dcacklam @scottleibrand I've been using BLE on both Edison and Pi0W for months in my non-openaps loop, and never kill or restart bluetoothd. What was the original need to do that?
joannestevens
@joannestevens
Feb 15 2018 17:35
Anyone ever had any problems with carbs not degrading? Ate at lunchtime and they are really going down!!
Not
Michael Spradling
@M1Sports20
Feb 15 2018 17:42
I have a question about autotune. I have been running it in a closed loop form. If I have No IOB from carbs and no carbs on board. My autotune_recommendation setting should match my temp basis for that hour correct?
I guess my question is. I should see temp basals mathcing the autotune_recommendations. Or do I have to manually configure my pump every hour. I didn't think this was needed
Scott Leibrand
@scottleibrand
Feb 15 2018 17:47
@joannestevens did you have CGM data, or was there a gap?
dan lightner
@danimaniac
Feb 15 2018 18:11
Help me understand how the offline webpage (which I love so far BTW) relates to xDrip+ and/or xDripAPS. It seems at first glance the offline webpage is "xDrip+ plus the OpenAPS pill info" for offline situations. My rig is NEVER directly online but is always bt pan to phone with xDrip+ - xDripAPS on rig uploads info when phone connects to wifi. I must keep xDrip+ running in this scenario, correct?
Scott Leibrand
@scottleibrand
Feb 15 2018 18:12
xDrip+ is getting the BG from your CGM, right? If so, yes, you need it for that. the offline webpage is just for visualizing the data OpenAPS has available
rhysjones
@rhysjones
Feb 15 2018 18:13
When running autotune I get the following message:
Limiting adjusted ISF of 8.28 to 31.67 (which is pump ISF of 38 / 1.2 )
Does that mean somethings off somewhere
It thinking my ISF being that far out seems just wrong
dan lightner
@danimaniac
Feb 15 2018 18:18
@scottleibrand Yes, xDrip+ is getting BG from CGM. Thank you.
Scott Leibrand
@scottleibrand
Feb 15 2018 18:34
@rhysjones it's normal to get a single day's data producing outliers like that. the Limiting thing prevents those from affecting ISF too much.
rhysjones
@rhysjones
Feb 15 2018 18:35
ok thanks
One more thing, I seem to be having spiking after meals, the basal kicks in but massively overcorrects and I end up going low. Am I correct in thinking I need to raise DIA?
Currently 6 hours
Scott Leibrand
@scottleibrand
Feb 15 2018 18:36
what does autotune say about your CR?
rhysjones
@rhysjones
Feb 15 2018 18:36
it's recommending I drop it from about 1:7 - 1:6
on 1 week of data
Scott Leibrand
@scottleibrand
Feb 15 2018 18:37
hmm... if you eventually go low, that usually means your CR is too low (too aggressive), or possibly your mealtime basals are too high. what does it recommend for basals?
rhysjones
@rhysjones
Feb 15 2018 18:37
screen shot incoming....
Screen Shot 2018-02-15 at 18.37.46 1.png
mid afternoon is the main issue hypo wise, autotune is saying basal is good
Scott Leibrand
@scottleibrand
Feb 15 2018 18:39
that's across-the-board recommending more insulin, for basals, ISF, and CR. do you tend to run high and need to correct a lot? are you looping with autosens? where does your sensitivityRatio usually run?
rhysjones
@rhysjones
Feb 15 2018 18:39
it used to run at about 32
but the first couple of autotunes drove it the other way
maybe that was the error
Scott Leibrand
@scottleibrand
Feb 15 2018 18:40
32? you mean ISF?
rhysjones
@rhysjones
Feb 15 2018 18:40
yes
Scott Leibrand
@scottleibrand
Feb 15 2018 18:40
sensitivityRatio is a ratio, with neutral = 1.0
so you're saying autosens was adjusting ISF down from 38 to 32, until autotune kicked in and made adjustments in the same direction?
rhysjones
@rhysjones
Feb 15 2018 18:41
it was running @ 1.2 for the past few days, but again I kept going low
Scott Leibrand
@scottleibrand
Feb 15 2018 18:42
I understand you're having lows. are you also having lots of highs requiring high temping or correction boluses?
rhysjones
@rhysjones
Feb 15 2018 18:42
So I switched back to Ma
I think so
"rebound lows"
staying around 180 in US units a lot for 3-4 hours after meals then overshooting when it finally lands
Scott Leibrand
@scottleibrand
Feb 15 2018 18:43
so the lows are usually after meals?
rhysjones
@rhysjones
Feb 15 2018 18:43
yeah 4-5 hours after maybe
Scott Leibrand
@scottleibrand
Feb 15 2018 18:43
in that case, yes, raising DIA from 6h to 7h sounds like it'd help
rhysjones
@rhysjones
Feb 15 2018 18:44
ok will try that and drop the autotune recommendations on
Dana Lewis
@danamlewis
Feb 15 2018 18:45
@M1Sports20 you don't have to change anything manually. But it depends on your insulin needs for the hour based on other factors that drives temp basal settings. So they may or may not match your underlying basal rates from Autotune, if you need something else at that time for other reasons.
Michael Spradling
@M1Sports20
Feb 15 2018 18:45
ok, makes sense. I was more trying to verify it was actually working.
It hasn't adjusted my basals much at this point, so thats good
Dana Lewis
@danamlewis
Feb 15 2018 18:47
If you're looping, it's working. And you can check the Autotune report to see what rates it is using for the calculations
Michael Spradling
@M1Sports20
Feb 15 2018 18:48
ok thanks. will look
rhysjones
@rhysjones
Feb 15 2018 18:49
Is Rapid-acting Oref the best insulin profile to use on Novorapid/log?
Scott Leibrand
@scottleibrand
Feb 15 2018 19:00
are you using AndroidAPS?
rhysjones
@rhysjones
Feb 15 2018 19:00
yep
Scott Leibrand
@scottleibrand
Feb 15 2018 19:01
you could confirm that in https://gitter.im/MilosKozak/AndroidAPS, but I think that one will work well
rhysjones
@rhysjones
Feb 15 2018 19:01
Will do, just a job to know whether it's best to try that at a longer dia, or one of the others with more action at the end with a shorter dia
benjaminbake
@benjaminbake
Feb 15 2018 19:25
New to looping. My rig isn't communicating with my pump. Any troubleshooting ideas?
Getting the 0, -99
kallnap
@kallnap
Feb 15 2018 19:41
@benjaminbake do you have the correct frequency setting for US, WW pump?
benjaminbake
@benjaminbake
Feb 15 2018 19:42
I selected us in the terminal setup
root@baker:~# grep mmtune /var/log/openaps/pump-loop.log
mmtune: "916.636", 0, -99 waiting for 78 second silence before continuing
Dana Lewis
@danamlewis
Feb 15 2018 19:54
What kind of rig?
benjaminbake
@benjaminbake
Feb 15 2018 19:56
Edison with explorer/712
Diadon81
@Diadon81
Feb 15 2018 20:09
Dear someone can explain me, why there is still 16 COB after 3,5 hours meal?
image.png
image.png
maybe I don't understand something?
is it possible to UAM to add carbs to COB value?
Scott Leibrand
@scottleibrand
Feb 15 2018 20:23
No, COB is determined by how much carb absorption is actually observed, as measured by how much BG rose (or didn’t fall) relative to what it would’ve done from insulin activity alone.
Diadon81
@Diadon81
Feb 15 2018 20:35
So it means that if my BG rose or didn't fall according to predicted IOB rig thinks that I still have a carbs?
Scott Leibrand
@scottleibrand
Feb 15 2018 21:16
if your BG dropped and required a lot of zero temping, or if you under-bolused and it didn't rise, then it will think you still have COB, yes
there is a min_5m_carbimpact that ensures a minimum level of carb absorption even when BG doesn't indicate any
you should also check whether your CR is accurate: see if autotune recommends raising it
joannestevens
@joannestevens
Feb 15 2018 22:01
@scottleibrand yes I had data, just updated to 10.2, temp basals still not showing, carbs have now degraded, will try again
benjaminbake
@benjaminbake
Feb 15 2018 22:26
Dana
jsw2480
@jsw2480
Feb 15 2018 23:08
Hi all, first time on Gitter and also flashing the Edison with jubilinux. I've gotten as far as downloading everything, and running the code. However when I flash, I end up in Rescue Mode: Starting Rescue Shell...
[ OK ] Started Rescue Shell.
[ OK ] Reached target Rescue Mode.
Welcome to rescue mode! Type "systemctl default" or ^D to enter default mode.
Type "journalctl -xb" to view system logs. Type "systemctl reboot" to reboot.
System Maintenance Mode
root@edison:~#
Scott Leibrand
@scottleibrand
Feb 15 2018 23:09
What happens when you Type "systemctl default" or ^D to enter default mode?
jsw2480
@jsw2480
Feb 15 2018 23:09
the Edison reboots but goes right back to Rescue mode. But... I'll try it again...
Scott Leibrand
@scottleibrand
Feb 15 2018 23:10
ok. new question: which version of jubilinux did you flash?
jsw2480
@jsw2480
Feb 15 2018 23:11
0.2.0
Scott Leibrand
@scottleibrand
Feb 15 2018 23:17
and when it told you that it was going to take ~10m to flash, how long did that step actually take?
jsw2480
@jsw2480
Feb 15 2018 23:19
pretty short before I saw the prompt "root@edison:~#" in the bottom terminal window... I did not (this time) press ctrl-d. In prior attempts I did that repeatedly, not knowing if the script was executing, or had stopped.
Scott Leibrand
@scottleibrand
Feb 15 2018 23:20
if that step didn't take at least 5 minutes, that's your problem
it means the OS didn't fully flash to the Edison, which is why it can't boot it
check the troubleshooting section of the docs for how to increase your virtual memory / swap and try again
jsw2480
@jsw2480
Feb 15 2018 23:21
ok will do. thx!
kallnap
@kallnap
Feb 15 2018 23:21
if I remember correctly looking for decent USB cables is also important. the Edison seems pretty picky
Scott Leibrand
@scottleibrand
Feb 15 2018 23:22
yes, that could interrupt it as well
garykidd
@garykidd
Feb 15 2018 23:25
Whats latezt on pi zero rigs?
Scott Leibrand
@scottleibrand
Feb 15 2018 23:25
work is ongoing to get the go-explorer-hat branch working with them, in https://gitter.im/openaps/hardware-dev
garykidd
@garykidd
Feb 15 2018 23:26
Thx.
John Benjamin
@Bender1061
Feb 15 2018 23:26
so I see that Hamshield has the preorder setup for the explorer hats. are we telling people to preorder yet?
Scott Leibrand
@scottleibrand
Feb 15 2018 23:27
yes, if you already have a working Edison rig, and are interested in a Pi rig as well, you can go ahead and order an Explorer HAT.
for newbies we still recommend getting the Edison + Explorer Board first
John Benjamin
@Bender1061
Feb 15 2018 23:27
ok cool
garykidd
@garykidd
Feb 15 2018 23:30
Scott have you/dana ever rellaced g5 xmitter batteries? Just curious.
Scott Leibrand
@scottleibrand
Feb 15 2018 23:31
Dana's still using a G4, so no.
kallnap
@kallnap
Feb 15 2018 23:32
@garykidd as far as I understood you can not change batteries on G5 and still use the official Dexcom receiver with it
Scott Leibrand
@scottleibrand
Feb 15 2018 23:33
that's correct. you have to use xDrip+ or xdrip-js or some other DIY receiver