p1 water incorrect since 2026.1

[janvanbesien]

Version: 2026.1
Platform: raspberrypi
Plugin/Hardware: P1 smart meter USb
Description:

Since the upgrade to 2026.1, the water meter readings from P1 seem incorrect. The usage for the current day is correct, but past days are wrong. Here is an annotated example from a holiday period with very little usage.

Code: Select all

"DateTime","Water"
"2026-03-02 00:00:00", // values was indeed 0 on these days
"2026-03-03 00:00:00",
"2026-03-04 00:00:00",
"2026-03-05 00:00:00",
"2026-03-06 00:00:00",
"2026-03-07 00:00:00",
"2026-03-08 00:00:00",
"2026-03-09 00:00:00",91 // probably correct
"2026-03-10 00:00:00",3 // probably wrong
"2026-03-11 00:00:00",3 // certainly wrong, same value as previous day
"2026-03-12 00:00:00",3 // certainly wrong, same value as previous day
"2026-03-13 00:00:00",3 // certainly wrong, same value as previous day
"2026-03-14 00:00:00",21 // probably wrong (I think it was 12)
"2026-03-15 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-16 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-17 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-18 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-19 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-20 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-21 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-22 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-23 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-24 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-25 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-26 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-27 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-28 00:00:00",21 // certainly wrong, same value as previous day
"2026-03-29 00:00:00",21 // certainly wrong, it was 343 on the day itself
"2026-03-30 00:00:00",43 // correct (for now)

Values in the database seem correct?

Code: Select all

sqlite> select * from meter_calendar where devicerowid=7 order by Date desc limit 20;
7|343|151659|5.6502|2026-03-29
7|0|151316|0.0|2026-03-28
7|0|151316|0.0|2026-03-27
7|0|151316|0.0|2026-03-26
7|0|151316|0.0|2026-03-25
7|0|151316|0.0|2026-03-24
7|0|151316|0.0|2026-03-23
7|0|151316|0.0|2026-03-22
7|0|151316|0.0|2026-03-21
7|0|151316|0.0|2026-03-20
7|0|151316|0.0|2026-03-19
7|0|151316|0.0|2026-03-18
7|0|151316|0.0|2026-03-17
7|0|151316|0.0|2026-03-16
7|0|151316|0.0|2026-03-15
7|0|151316|0.0|2026-03-14
7|12|151316|0.1977|2026-03-13
7|0|151304|0.0|2026-03-12
7|0|151304|0.0|2026-03-11
7|0|151304|0.0|2026-03-10

[gizmocuz]

Are you on the latest beta version? If not, try the latest beta first.
There have been no changes for this since ages, then again I do not have a P1 meter with Water.

Could you post some P1 diagrams?

But I am quite sure it should be OK, maybe post a few P1 datagrams if possible, or at least 2, beginning of the day and next day, with water usage

[janvanbesien]

Here are the diagrams.



The "today" and "this month" values seem correct.

The first graph is also correct.

The second graph (usage last month) is wrong. The values should not be the same on multiple days. The first day of every such period (the 9th, 10th, 14th, ...) might be correct but I have doubts.

The third graph (usage last year) shows a similar pattern. Periods where the usages was zero in the past, now have a consistent non-zero value. In feb and march 2026 my usage was indeed very low, but not like shown (should be consistently zero, not consistently non-zero).

[gizmocuz]

Ahhh I know what is happening... I will have a look at it. It detects a peak and thinks based on the previous day(s) that the meter was not running and therefore is averaging it

Regarding the P1 datagrams, I actually meant the serial datagrams from the P1 meter... but thanks to your screenshots I know where to look

[janvanbesien]

Meanwhile I updated to the latest beta as you suggested, and that seems to fix it.

[gizmocuz]

Perfect, yes always update to the latest beta first