Archived Forum

Noise? That is the most likely explanation. Without knowing what your maximum full-scale power is, 5.7 W is probably way below the value at which you can expect reliable and accurate measurements.

I know it was way below for accurate or reliable measurements but how the hell did it get so low ?

Had it running now for 3 day and over night it never went under 100W.
All network devices are always on (moder, router, switch, UPS...) it can't go so low.

What is your residual load? Is it possible you had a half-hour outage and your UPS carried it through?

Residual load? Have 0 ideas what residual load is :)

Just had a look at what my arduino in reporting.

Irms = 0.46A

Vrms = 230,16V

RealPower = 10.3W

This does not look ok :) The value is directly displayed on my LCD (arduino) and the same value is send to emoncms.

This is my code:

 emon1.calcVI(20,2000); // Calculate all. No.of half wavelengths (crossings), time-out
 
  porabaT = emon1.realPower;
  porabaP = emon1.apparentPower;
  faktorP = emon1.powerFactor;
  Vrms = emon1.Vrms;
  Irms = emon1.Irms;

printToLCD(porabaT, Vrms, Irms);

And in setup

emon1.voltage(2, 219.4,  1.7);
emon1.current(1, 60.6); // 100A CT

It appears to me that your monitor correctly recorded zero power for two periods, one of thirty minutes and one of about 8 minutes. I'm reasonably confident that the monitor itself worked throughout the period because what you graph shows is consistent with what I've described. You are going to have to do a bit of detective work before anyone can help you further.

Detective work on what ?
 

I have no idea what to look at or what to test :)

Let it run a while and see if a pattern occurs. If it does then try to find out what's unique about those occurrences.

Paul

It's there on every night when the usage dropped below 80-100W. When no other devices were running only my network stuff. And goes away even if you switch lights on in any room ie. usages goes up and over 100w.

Which sketch are you using? can you provide a link

The one from the provided example in the library for arduino for current and voltage.

Sketch from here http://openenergymonitor.org/emon/buildingblocks/how-to-build-an-arduino...

Which sketch are you using? can you provide a link

http://openenergymonitor.org/emon/buildingblocks/how-to-build-an-arduino...

This one at the bottom.

Edit - deleted duplicate posts - BT

That's a well established software version and checking 'git issues' nobody else is reporting similar problems with it.

I suppose a logical approach would be, if it's occurring every night, is to establish if it's actually recording what is occurring, or is producing false readings.

I presume that as a electrical engineer you will have the knowledge and tools to identify which it is - as Robert Wall has said - detective work!

Paul 

Current and voltage are almost spot on when messured with my multimeter.

So Vrms and Irms are OK.

Current and voltage are almost spot on when messured with my multimeter.

'almost spot on' infers that you are comparing 2 values. Are you comparing the multimeter reading and emoncms - if so, then there's no issue with emoncms - it's recording accurately.

If you are comparing your multimeter reading with something else, please tell us what, and why the readings are 'almost spot on'?

Paul

Almost spot on is few % off (231V emoncms / 230V multimeter). 

Will try to log the values from my multimeter and add voltage and current to feed to compare both.

My last post in this thread.....

Paul

Where in the world are you? Knowing that might give us a clue.

As I wrote above, the very low and wildly fluctuating values below about 8 W look consistent with the sort of noise that is generally reported from Arduino builds. So I'm reasonably confident that the Arduino and the recording chain right up to that graph is accurately recording the power from the point where, as separate voltage and current, it meets the inputs to your Arduino.

It's not likely that your ac adapter jumped out of the wall and back again, or your CT dropped off the cable and magically reattached itself, a few minutes later, so given that you listed a UPS as part of your normal residual load (and you as an electrical engineer have never heard that term? ! ! !), based on what little you've told us of your set-up I think the likeliest explanation is the supply really did go off where the graph shows the wildly fluctuating very low - and a few negative - values.

If you have an analogue electric clock of the sort that has a synchronous motor inside, by running that off the suspect supply and checking whether it loses time and by how much, you'd be able to see how that correlates with your emoncms records.

When I have the lowest power recorded (negative) the voltages is at 230V on the LCD and the current at 0.46A. So the only wrong value is the calculated power.

Irms = 0.46A, Vrms = 230,16V, RealPower = 10.3W

That's not necessarily wrong, but it does imply a pretty poor power factor.   At the time of the low power readings, what do porabaP and faktorP contain?   You say the sketch you're running is from an example in the building blocks, but I can't find any sketch there with variables with those names.

Why not attach your sketch, as an actual .ino file that we can download and look at?

Is most of your expected 100W background load plugged into the UPS?  Do you get similar numbers/graphs if you kill the power to your UPS?  Does your UPS log when it's decided to switch to battery power?  Is it possible your grid power quality is poor enough that the UPS has decided to kick in because frequency or voltage is out spec?

Are you saying that when it's in low-power mode, you only need to turn a light on somewhere to fix it?

Will look at what porabaP and faktorP values are at when I'm home. And they are just local variables for Apperent power (porabaP) and Power factor (faktorP) nothing else!

And I will post the *.ino file! when I'm back home.

No even when I disconnect my UPS I still get the same when the power (realpower) dropps under aprox. 90W.

Even mesured current draw and voltage on my electricity meter! Voltage is the same even on the decimal point and It did not dropp below 229V and above 242V over night. Current went down to 0.5A and when it goes under 0.5A it's all wrong I get from 100W to 10W to -1W and so on...

Yes, when I switch something on that will draw enough power to get the realpower over 90W all is fine again!
 

Edit - attached sketch directly to post vice third party site - BT

 Current went down to 0.5A and when it goes under 0.5A it's all wrong I get from 100W to 10W to -1W and so on...

I'm not entirely sure I'm following the exact sequence of events, but what you've described above sounds like an uncorrected phase error in your CT.  Check the report in the building block section but you'll see as current goes very low, phase error can get very high and potentially even push you into the next quadrant, which would explain the negative power.

I see!

But can I fix this?
 

It's very difficult to deal with varying phase error... even fairly high-end energy ICs (at least the ones I've played with) only support a constant phase error adjustment.

Did you calibrate the PHASECAL setting in your sketch as per the calibration instructions?  What current did you calibrate it at?

You could throw money at the problem, in the form of more expensive CTs, but they are often of the 333mV form, which are not well suited to the OEM design.

Can't find anything on PHASECAL setting in calibration instructions or I'm looking at the wrong page!
 

My bad, I think I misnamed it.  It's steps 4 and 5 here:

http://openenergymonitor.org/emon/buildingblocks/calibration

Aha OK. I left it at 1.7

So I need to calibrate it.

Adjusted the value as per calibration instructions.
Real power and Apparent power are now very close! 
Power factor is at 0.99 or 1.00 depending on power draw.

Will test it over night to see if anything changed.