Hey @pswired. Have you already filed a support ticket for this? I believe you’ve given us permission to view your monitor previously, so I can pass along your information on top of the Support request.
Questions for you:
When did you first notice the Always On bubble/usage was gone?
Is all of your “always on” usage captured via smart plugs?
I’ll have to look in a little greater detail, but it looks the fix might have coincided with an uptick in my Always On, possibly roughly equal to the sum of my smartplug Always Ons. Old Always On was about 330W prior to Aug 31st. New Always On is around 420 (but bouncing around more than I would expect). Sum of amrtplugs is around 160W
After talking with @JonahAtSense, this is a bug unrelated to the recent fix for smart plugs not nesting under always on. If you are experiencing this issue, please report it to support@sense.com so we can track it and develop a fix.
We made some changes on our end, @pswired. Is your Always On visible again?
Any chance that sense can create a toggle switch for this functionality? I’ve grown accustomed to and now prefer having my smart plugs not showing up as nested.
Hey @senseinaz. You’re the first I’ve seen to mention this, but I think there might be more folks out there as well who feel similarly. Can you create a Product Wishlist post with this feature request?
I tried one more experiment with Always On but it was sort of inconclusive. I tried to emulate Sense’s Always On calculation by doing a 1% bin on the 15 min data coming from my utility over the previous 48 hours, then comparing. Three main vectors for differences:
15 min sample vs 1/2 second samples - one would expect the 1% bin of 1/2 second samples to be smaller than the 1% bin for 15 min samples over the same period of time. As expected, most of the PGE Always On values lie above the unity line.
I had to add hourly Sense solar data to my 15 min PGE meter data since the meter data would otherwise be net meter info, unusable for Always On. Usually the solar production is fairly stable over an hour and in most cases daytime usage with solar does not contribute to pulling down Always On, so this shouldn’t add too much error. But hard to quantify.
Missing data - I have a few hours of missing data in June, plus a few more hours throughout where Sense solar or Sense Always On data is missing. Missing data makes the calculations and comparisons in those regions suspect. I could spend a fair amount of time trying to fill or fence off those regions, but it would be ugly.
The good news is that the Sept (month 9) is where it should be, with the PGE Always On emulation slightly above and somewhat linear with the Sense Always On data.
ps: I do have one technical question raised from this experiment - how is the 1% binning done ? There are two ways:
Discrete - does Sense look at the 3,456th smallest sample out of the 345,600 samples from the past 48 hours. I chose this approach (except for me, it was the 2nd smallest value out of the 192 samples in a 48hour period).
Statistical - or does Sense fit the 48 hour sample window data to a distribution and produce a 1% value. Thats the way the R quantile function works. They give very different results.
As of 10 pm late this evening the 26th my smart plugs disappeared from my always-on.
Always-on shows 45 watts. Should be 136 to 143 area.
The bubbles that are from any of my smart plugs no longer show the the always on wattage, just the current wattage.
I noticed around this same time before they disappeared my phone could talk to the monitor, but I couldn’t get anything from sense.
Meaning all my stats was gone, it’d say something like can’t connect to sense. But I could watch my power meter in action… and also see my on now bubbles operate accurately.
But no stats on equipment or stats on always on, id get that notification saying retry, can’t connect to sense.
This was all between 10pm and mid night