the minute trace on my dash (which comes from the pulse output of a PRO75D meter at 1600imp/kwh) is lumpy, i don't see this problem with my generation pulse/meter output. averages out ok, but makes FluksoViz display rather ugly.
https://dl.dropbox.com/u/7642167/peaks.jpg
any ideas please?
Thats because the pulse output is just that. A higher current flow will see it smooth out more.
Bazzle
hmm, well i see the same behaviour when it's pulling large currents too. perhaps i don't see it in the generation because it's only 1000 imp/kwh.
I can recommend a current clamp for a more 'live' reading, check the Flukso shop.
Internally the FLM works with electricity 'pulses' of 1Wh. If your meter's pulses are not a fraction of 1Wh (i.e. 0.5, 0.25, 0.2, 0.125, ...), then you're bound to get these 'bumps'. In your case, you've got a meter constant of 0.625Wh/pulse.
The real-time power calculation does not take the fractional parts into consideration. This means you'll get a higher instantaneous reading between pulses 2.500 and 3.125.
Please note that this phenomenon only shows up in the real-time readings. It does not impact overall accuracy.
ok, that explains it, thanks.
So am I best not to upgrade the firmware from 216 to avoid this bug since as consumption meter is a 1600imp/kwh one?
If its not broke :)
> Please note that this phenomenon only shows up in the real-time readings. It does not impact overall accuracy.
True, but it also makes for very wobbly consumption graphs on PVOutput because the snapshot value is used, and that can vary by as much as +-30% or so with my 400imp/kwh meter. It really would be nice to have this fixed in the next firmware update. The fluctuations also make it hard to identify how much power a particular appliance actually draws. When I turn something on, the graph jumps wildly up and down, and I can't just say "aha, this thing draws 50W".
Michi.
Here is an example of what is logged for the past 30-odd hours. Note the sinusoidal pattern between midnight and 7:00am. That should be a horizontal steady up and down as the fridge turns on and off. Instead, I get these growing and shrinking waves due to the beat effect with the way the Flukso samples the data. It's not even vaguely resembling real consumption.
http://www.triodia.com/staff/michi/images/Graph.jpg
Cheers,
Michi.
Please sign up for the r241 alpha release at this forum post.