Skip to content

Conversation

@scottleibrand
Copy link
Contributor

We often see non-meal-related rises that benefit from SMB/UAM correction. For people who have been running SMB long enough to be comfortable running it overnight, this provides an option to enable such corrections 24x7 (except when overridden by a high temp target). This is not recommended for new users, so it will not be exposed in displayedDefaults.

@ghost
Copy link

ghost commented Sep 26, 2017

that'll be a great idea, for permanent SMB usage I often triggered with temp target

@tim2000s
Copy link
Contributor

Presume this is "enableSMB_always":"true" in preferences.json?

@scottleibrand
Copy link
Contributor Author

Yeah. You can also confirm that by looking at the diff: https://github.com/openaps/oref0/pull/677/files

@jaylagorio
Copy link
Contributor

This one looks good to go

@jaylagorio
Copy link
Contributor

I like having SMB turn off for a big jump, but can the reason for being disabled on that run (for any reason) maybe go to the OpenAPS pill's hover text?

@scottleibrand
Copy link
Contributor Author

Good idea. I just added delta > 10% to the reason field. The minGuardBG < threshold is already displayed there, so I think that's already clear.

@jaylagorio
Copy link
Contributor

And what do you think about having the percentage be slightly higher? Maybe 15% is a good starting point? Thinking about my SMB-enabled meals lately, it's not uncommon for the first jump from being around ~100ish to be 11 or 12, then slowing down when it gets in the 150s, and capping out around 170s. 10% would slow the overall response time in at least my case. This is a hard one since it's all subjective and a judgement call.

@scottleibrand
Copy link
Contributor Author

Yeah, I think 10% might be a tad too low. Do you see "real" 10-15% deviations for more than 10m after meals in your recent data? If so, I'll probably raise the threshold from 10% to 15%.

@jaylagorio
Copy link
Contributor

I don't think so, but if so not for more than 15 minutes and so far not often.

@scottleibrand
Copy link
Contributor Author

Ok. Let's keep an eye out for it happening, and it if does we'll raise the threshold. I've already seen a few cases where 10% has helped avoid SMBing on false jumps from a noisy new sensor and from calibration, so there's benefit to having it as low as we can without messing up meal SMBs.

@scottleibrand
Copy link
Contributor Author

Gonna go ahead and merge this. We can tweak further as needed with another PR to dev.

@scottleibrand scottleibrand merged commit 3bd4605 into dev Sep 29, 2017
@scottleibrand scottleibrand deleted the enableSMB_always branch September 29, 2017 03:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants