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

8th
Jan 2019
Martin Haeberli
@mhaeberli
Jan 08 02:00
Screen Shot 2019-01-07 at 17.56.46.png
trying to understand why SOME 2500 mah batteries have the connector polarity reversed, as here; the AdaFruit batteries seem to be correct.
Scott Leibrand
@scottleibrand
Jan 08 03:43
Standards are hard?
Jens Heuschkel
@juehv
Jan 08 13:23
@mhaeberli the connector polarity isn't defined in a standard. So actually there is no reversed :smile:
renegadeandy
@renegadeandy
Jan 08 13:32
@scottleibrand no other errors, if you look up you will see my full log from that loop.
viq
@viq
Jan 08 14:02
@sethgagnon I found your comment https://github.com/openaps/oref0/issues/965#issuecomment-428233852 - did you manage to get it working? I'm struggling with the same right now
renegadeandy
@renegadeandy
Jan 08 14:23
Hi everyone. I am finding my openaps is doing a great job at removing my highs, but I am finding that there is a snaking pattern in my bloods which frequently bring me to 3.8 or so.
What can I change on the algorithm to help remove these lows? I am assuming changing my existing basal rates won't really make any difference?
Eric
@ecc1
Jan 08 14:40
@mhaeberli if you get batteries from anywhere except adafruit and sparkfun, you have to check the polarity yourself. Don't even believe that red is always positive!
sethgagnon
@sethgagnon
Jan 08 14:56
@viq no I did not get it working
Raymond Richmond
@PedanticAvenger
Jan 08 15:42
@mhaeberli I'll underscore @ecc1 here. I had a battery that the wire colors were swapped w.r.t. polarity and I ended up cooking off an explorer board. The only rule is check, check, then check again!
Scott Leibrand
@scottleibrand
Jan 08 17:20
@renegadeandy are you running autotune? if so, what has it detected so far?
renegadeandy
@renegadeandy
Jan 08 18:01
@scottleibrand I believe i am
renegadeandy
@renegadeandy
Jan 08 18:07
What file gives a summary of its detections?
renegadeandy
@renegadeandy
Jan 08 18:45
I am also finding that when the edison has been connected to the bluetooth hotspot, and I move back into a wifi zone, it actually doesn't drop the bluetooth connection.
Scott Leibrand
@scottleibrand
Jan 08 19:08
You should have a cat-autotune alias on your rig.
renegadeandy
@renegadeandy
Jan 08 20:28
Ah
here it is

Parameter | Pump | Autotune

ISF [mg/dL/U] | 39.600 | 47.732
Carb Ratio[g/U]| 9.000 | 9.257
Basals [U/hr] | - |
00:00 | 0.950 | 0.975
00:30 | |
01:00 | | 1.001
01:30 | |
02:00 | | 0.958
02:30 | |
03:00 | | 1.040
03:30 | |
04:00 | | 1.000
04:30 | |
05:00 | 1.200 | 1.261
05:30 | |
06:00 | | 1.065
06:30 | 1.9 |
07:00 | | 1.570
07:30 | |
08:00 | | 1.559
08:30 | |
09:00 | | 1.643
09:30 | |
10:00 | | 1.760
10:30 | |
11:00 | | 1.706
11:30 | |
12:00 | 1.800 | 1.623
12:30 | |
13:00 | | 1.584
13:30 | |
14:00 | 0.900 | 1.019
14:30 | |
15:00 | | 0.973
15:30 | |
16:00 | 0.950 | 0.959
16:30 | |
17:00 | | 0.924
17:30 | |
18:00 | | 0.986
18:30 | |
19:00 | 1.300 | 1.270
19:30 | |
20:00 | | 1.340
20:30 | |
21:00 | 1.000 | 1.122
21:30 | |
22:00 | 0.900 | 0.984
22:30 | |
23:00 | | 0.981
23:30 | |
image.png
so it looks generally, like autotune is giving more insulin across everything, which may be fine, but I am still feeling a number of dips across the new 24 hour pattern
Dana Lewis
@danamlewis
Jan 08 20:49
Your pump basal at 6:30-13:00 is pretty high - I would consider changing the 1.9 and maybe the next 1.8 down, and rerun autotune, and see if it continues to tune your daytime basals down further from there
renegadeandy
@renegadeandy
Jan 08 21:02
so will autotune always take into account my current basals?
i believe it runs everyday at 04:00
Scott Leibrand
@scottleibrand
Jan 08 21:07
it only uses the pump's basal schedule to as a guideline for how far it can tune. if the autosens_min is 0.7, then it can go as low as 70% of what's in your pump.
renegadeandy
@renegadeandy
Jan 08 21:10
Gotcha, excellent
@danamlewis ill make those suggested changes now to my basals
ill move them 1.7 and 1.6 respectively (0.2 lower in each case)
renegadeandy
@renegadeandy
Jan 08 21:16
So I assume with autotune turned on, as per phase b (which I believe I am setup to run with), my pump will be delivering insulin using the figures on the right hand side of the autotune configuration
Does autotune keep a history of its changes
(if not, I may add this to the openaps project through a pull request as I would find it useful)
Another question would be - would lowering my sensitivity factor on the pump i.e how many mmol 1 single unit of glucose lowers my blood again perhaps curve off some of these lows?
i.e does the openaps use this sensitivity factor ( i know it asked me during setup hence why i ask)
Scott Leibrand
@scottleibrand
Jan 08 21:45
@renegadeandy each night's autotune run is logged in a timestamped file
autotune also adjusts ISF: it has raised yours from 39.6 to 47.7 mg/dL/U
as with the basals, you should probably raise the pump closer to what autotune recommends if you want to allow it to keep raising it
Zach Gohr
@zgohr
Jan 08 22:04
Interestingly, autotune does not work for us. It lowers basals consistently. We ran it until we had our basals basically 1/2 what they should have been until we’re constantly running high (thinking the whole time we’re doing something wrong.) Then we disabled it, set basals to where we knew they should be, and everything got better.
Scott Leibrand
@scottleibrand
Jan 08 22:08
did you ever try categorize-uam-as-basal?
Abigail Cember
@acember
Jan 08 22:09
@renegadeandy I was having some problems with lows, too, but I think two things I did fixed it: a) made the "peak insulin time" longer than I had had it previously and b) turned on super-micro-bolus (SMB), so that that is now primarily used for correcting highs rather than increased temp basals. Relatedly, @ZachGohr -- that's really strange. I haven't really been watching what my autotune is doing, but my pump seems primarily to use SMB to manage to my bloodsugar overnight, which seems to be working very well.