Current Cost Support Forum

Hi all,

I've got a question about the "monthly history" XML output by the EnviR. Since the EnviR has no idea of dates, it can't possibly know about months, their lengths, or what day of the month the EnviR was "born" on.

So what approximation to a "month" does the EnviR use, exactly?

(I ask because I'm developing a plugin for the Micasaverde Vera that collects history data, and I'm trying to figure out what to put on the horizontal axis.)

The full XML specification should this one up for you. Details here:

I have just read that document and before had never considered before that when you setup any envi model all you enter is the time with no date or year information.

Thus it would seem it assumes for historic data every month is 31 days and I guess the start of the month is when you first power it up?

The CC EcoManager product does let you set the full time, date and year so it would be interesting to know if it knows about real months...

Automan.

@currentcost,

Thanks for the reply. I asked here precisely because the XML spec you linked to doesn't define how the EnviR chooses when months begin. Nor does the new PDF version of the spec from January.

I'm assuming that it's something like @Automan's guess, but I would only be guessing too, and since I am coding up a tool to be used by others I'd like to have a definitive answer.

google power meter needs to give their software over for free when they shut down, problem solved

Looking back when I checked the output of the EcoManager device I note it does indeed output the complete date so perhaps this device is more accurate?


<msg><src>EMC2-v1.79</src><date>07/11/2010</date><time>09
:56:00</time><d07m11><kwh>6</kwh></d00m11><d08m11><kwh>0</kwh></d01m11><d09m11><
kwh>0</kwh></d02m11><d10m11><kwh>0</kwh></d03m11><d11m11><kwh>0</kwh></d04m11><d
12m11><kwh>0</kwh></d05m11><d13m11><kwh>0</kwh></d06m11><d14m11><kwh>0</kwh></d0
7m11><d15m11><kwh>0</kwh></d08m11><d16m11><kwh>0</kwh></d09m11><d17m11><kwh>0</k
wh></d10m11><d18m11><kwh>0</kwh></d11m11><d19m11><kwh>0</kwh></d12m11><d20m11><k
wh>0</kwh></d13m11><d21m11><kwh>0</kwh></d14m11><d22m11><kwh>0</kwh></d15m11><d2
3m11><kwh>0</kwh></d16m11><d24m11><kwh>0</kwh></d17m11><d25m11><kwh>0</kwh></d18
m11><d26m11><kwh>0</kwh></d19m11><d27m11><kwh>0</kwh></d20m11><d28m11><kwh>0</kw
h></d21m11><d29m11><kwh>0</kwh></d22m11><d30m11><kwh>0</kwh></d23m11><d01m12><kw
h>0</kwh></d24m12><d02m12><kwh>0</kwh></d25m12><d03m12><kwh>0</kwh></d26m12><d04
m12><kwh>0</kwh></d27m12><d05m12><kwh>0</kwh></d28m12><d06m12><kwh>0</kwh></d29m
12><d07m12><kwh>0</kwh></d30m12><msg><src>EMC2-v1.79</src><date>07/11/2010</date
><time>09:56:00</time><m11y10><kwh>31</kwh></m11y10><m11y10><kwh>31</kwh></m11y1
0><m11y10><kwh>31</kwh></m11y10><m11y10><kwh>31</kwh></m11y10><m11y10><kwh>31</k
wh></m11y10><m11y10><kwh>31</kwh></m11y10><m11y10><kwh>31</kwh></m11y10><m11y10>
<kwh>31</kwh></m11y10><m11y10><kwh>31</kwh></m11y10><m11y10><kwh>31</kwh></m11y1
0><m11y10><kwh>31</kwh></m11y10><m11y10><kwh>31</kwh></m11y10>


Automan.

Anyone writing software to parse the XML messages from the Envi needs to understand how the device stores and rolls its historic data "buckets", since it stores the data as past points in time relative to "now" (whatever it thinks "now" is).

This can introduce a number of inaccuracies and inconsistencies within the data itself (eg. if the device is switched off for any length of time, or the clock is reset).

Below is info I had from Current Cost when asking for detail on the historic data storage :

-----

The daily sectors are

11pm-7am night

7am-3pm day

3pm-11pm evening

The last 31 days Energy Usage is output as 2-hour aggregates of KWHr (3720 values covering 9 Appliances plus Whole House), the data being aggregated starting at odd hours, eg 23:00 to 01:00, 01:00 to 03:00 etc. This accommodates industry requests to view evening and night data segments distinguished at 23:00.

The last 90 days Energy Usage is output as daily aggregates of KWHr (900 values covering 9 Appliances plus Whole House).

For simplicity, this data overlaps the 2-hr history store in the present version.

The last 84 months Energy Usage is output as monthly aggregates of KWHr (840 values covering 9 Appliances plus Whole House).

For simplicity, this data overlaps the 2-hr and Daily history stores in the present version

So the day periods do not start at 00.00, but at 23.00

The monthly data is simplified into periods of 30 days regardless of the month.

I hope this helps you and please don't hesitate to contact me further.

Kind Regards
James Warner
Current Cost

I've reverse-engineered my own answer.

The first "month" of history on a CC-128 appears when there's been 31 days of history. The second "month" appears when there's been 61 days of history.

The very first day of history, when the "days-since-birth" equals 65535, doesn't seem to contribute to any "month". The second through 31st days of usage, added up, equal the first month (with some rounding); the 32nd through 61st days of usage, added up, equal the second month.

Extrapolating this, I'd guess that on the CC-128, all months are 30 days.