Monitoring PV output

[DamoM]

I have had a PV system installed since December and have been struggling to get any decent pruduction out of it which has been attributed to the angle of the panels (15 degrees) not being ideal for this time of year but reaping dividends in the summer. This remians to be seen I guess.

In the meantime I am trying to monitor my output without having to access an outbuilding and so used my internet-connected wireless Current Cost electricity monitor (with the hall sensor around the live wire from the inverters to the meter). This gives reasonably accurate readings of production during daylight hours but still shows that the inverters are drawing about 105w during dark hours.

I am talking to mastervolt about this (who say that it "could be the unload of the condensator from the inveter").

However, in order to eliminate the possiblity of error in reading and/or draw from the inverters I was looking to install a dusk til dawn light sensor in between the inverters and the meter so that it only connected the circuit when it was dawn and when electricity was being generated.

Has anyone else ever doen this?

[dimogga]

My OH has made something to measure how much we're generating ... but was looking at a light meter sort of thing too.

I can tell if we're generating from
(a) the shadow of our roof across the road (and late afternoon on a neighbours roof without even moving from my seat)
(b) from the owl meter showing bizarre readings like 1.500 kw when nothing is on.

[JohnS]

Do you have a shading problem?

It looks as though there are a lot of trees around your site and also the rows look to be close together.  In simple terms if there is any shadow on your panels, there will be a dramatic fall in output and hence your low production.

John

[CeeBee]

...This gives reasonably accurate readings of production during daylight hours but still shows that the inverters are drawing about 105w during dark hours.

I am talking to mastervolt about this (who say that it "could be the unload of the condensator from the inveter").

It's not unusual (it seems) for inverters to appear to be drawing this kind of power when in standby (mine - Fronius IG20 does), but it isn't real power - it will be mainly reactive power (current and voltage waves out of phase with each other) so no real power is going anywhere - the possible result of capacitance or inductance in the inverter. Most simple clamp-on energy monitors only measure the current and not the voltage - they just assume that they are in phase. So you're seeing the effect of current flowing to and fro, but the CurrentCost is displaying this as power when it isn't. I was able to check the mainly out of phase current & voltage on mine with a clamp-on monitor which does monitor the voltage signal as well (and which I happened to have worked on development of). The real electricity meter will be measuring both current and voltage so as to get an accurate reading.

Sounds like MasterVolt's explanation has been through a poor translator from some other language - perhaps they were trying to say the same as me?

[DamoM]

My OH has made something to measure how much we're generating ... but was looking at a light meter sort of thing too.

I can tell if we're generating from
(a) the shadow of our roof across the road (and late afternoon on a neighbours roof without even moving from my seat)
(b) from the owl meter showing bizarre readings like 1.500 kw when nothing is on.

Who/what is OH?

Love the shadow thing but what if it's cloudy??! But you don't know how much, even when sunny. Is your OWL meter specifically for monitoring PV?

[DamoM]

Do you have a shading problem?

It looks as though there are a lot of trees around your site and also the rows look to be close together.  In simple terms if there is any shadow on your panels, there will be a dramatic fall in output and hence your low production.

John

There is some shading but not as much as that picture suggests. The panels face south-east and that big tree (my neighbours!) blocks the light for a very brief period but to be honest the readings don't really reflect this. I can see that them being close together will impact in winter as they will shadown each other and also have a poor angle to the sun but preumably the summer readings will be the opposite?

[dimogga]

My OH has made something to measure how much we're generating ... but was looking at a light meter sort of thing too.

I can tell if we're generating from
(a) the shadow of our roof across the road (and late afternoon on a neighbours roof without even moving from my seat)
(b) from the owl meter showing bizarre readings like 1.500 kw when nothing is on.

Who/what is OH?

Love the shadow thing but what if it's cloudy??! But you don't know how much, even when sunny. Is your OWL meter specifically for monitoring PV?

My other half. He is into electronics so has built A gizmo thingy. His wireless bit arrived for it yesterday but he is too ill to be soldering today.

The owl just monitors how much is going through the big wire by the meter. It doesn't know what direction. Bit of maths and knowing normal use and remembering what is switched on meansi have vague idea of what we making.

Not very accurate but better than just looking at the sun

[M]

...This gives reasonably accurate readings of production during daylight hours but still shows that the inverters are drawing about 105w during dark hours.

I am talking to mastervolt about this (who say that it "could be the unload of the condensator from the inveter").

It's not unusual (it seems) for inverters to appear to be drawing this kind of power when in standby (mine - Fronius IG20 does), but it isn't real power - it will be mainly reactive power (current and voltage waves out of phase with each other) so no real power is going anywhere - the possible result of capacitance or inductance in the inverter. Most simple clamp-on energy monitors only measure the current and not the voltage - they just assume that they are in phase. So you're seeing the effect of current flowing to and fro, but the CurrentCost is displaying this as power when it isn't. I was able to check the mainly out of phase current & voltage on mine with a clamp-on monitor which does monitor the voltage signal as well (and which I happened to have worked on development of). The real electricity meter will be measuring both current and voltage so as to get an accurate reading.

Sounds like MasterVolt's explanation has been through a poor translator from some other language - perhaps they were trying to say the same as me?

CeeBee, sorry to bother you mate, but how common is this? It's just that that reading jumped right out and slapped me in the face, as I was recently reading a 'discussion' on the MSE site about an OWL monitor connected to a Fronius inverter reading 113watts all night and possibly during very low readings (less than an amp). Sounded extremely similar, but the other discussion couldn't decide if the 113watts was there all the time, even at high 'revs'?

Mart.

[Pat_]

Just for the record, my Fronius (when it's off at night) gives the same reading as these (+/-) on a Wattson. I have assumed it is spurious but will take the trouble soon to make some more detailed and accurate measurement.

[asorton]

If you upload to PvOutput which is easy as current cost already export your data to Pachube and in Pvoutput you just link in the Pachube Feed.

Pvoutput also has a rule you cn set to remove this overnight figure.

Better still change the clamp on your Current Cost meter for a wireless Optismart sensor thats attaches to your generation meter, it records the pulses so you get spot on generation figures, search some of AlanIOW posts he has this setup and has posted several times about it.

[CeeBee]

CeeBee, sorry to bother you mate, but how common is this? It's just that that reading jumped right out and slapped me in the face, as I was recently reading a 'discussion' on the MSE site about an OWL monitor connected to a Fronius inverter reading 113watts all night and possibly during very low readings (less than an amp). Sounded extremely similar, but the other discussion couldn't decide if the 113watts was there all the time, even at high 'revs'?

Hi there Mart

For somone who reckons to only tell stuff that I actually know, rather than what I just think, or just have read somewhere, then perhaps I over-stated it - but it is what I believe. I've only taken measurements on my own Fronius IG20, but I seem to recall several people here on this forum saying this same thing - readings up to around 100W even during the night using simple clamp-on monitors.

As I said, thanks to the fact that I was working on an energy monitor which senses the voltage signal as well as current, I could use special firmware to get to see the voltage and current sine-waves. During standby, I was seeing a current signal which, if one just assumed it was in phase with voltage, would have corresponded to about 100W. But in fact, the current was out of phase with voltage, so not real power at all. Looking at the reading on it just now, it reads 20W - OK - not zero, but this is going beyond what it was designed to do, and it only takes a small phase-error (around 10 degrees) to cause a 100 VAr reactive signal to look like it's 20W of real power. A monitor not sensing voltage would be reading around 100W. I also observed that when the solar started generating a small amount, then the current waveform wasn't very sinusoidal, and current/voltage still weren't entirely in phase - so indeed quite likely that the ~100VAr reactive power was still there. At higher solar powers, a nice sinusoidal current dominates (and in phase with voltage, not surprisingly, otherwise you'd never generate anything which would register on the proper meter!). Who knows whether the small out-of-phase component is still there at high powers - doesn't really matter, and would need fancier equipment to detect. And even if it is still there, it wouldn't cause a simplistic monitor to read 100W high - the in-phase and out-of-phase currents would add up in a 'square root of sum of squares' sort of way, so 100 VAr reactive would make very little difference to the readings if there was also 1kW of proper power there.

I also monitor generation/import/export by counting LED flashes on the proper meters (RFXCom equipment). I know that baseload of the house is below 50W (accounted for by modems, router, etc.), so no way there's actually 100W going into the inverter during the night.

[don0301]

thanks for more information CeeBee

yes, I'm the guy M mentioned!

I have a Fronius IG Plus 35V, with an Owl CM160 monitor that displays a constant 113W at night (well, actually reads 0.49A, Owl is set at 230V), or when the inverter is not "producing" (I have the monitor set to read power generation only, sensor inside the CU).

I did some research, and have proved the inverter doesn't "draw" power at night by switching all fuses off on the CU except the inverter one. reading the outside house meter then leaving over night (13) hours, and re-reading. the meter had not moved even 0.01kWh. the Fronius handbook says it draws 0.23W in "night mode".

I'm very interested in your reply, cos since installation just over 2 months ago I have been trying to understand the error in my Owl kWh reading and Gen. meter reading over time.

I have my Owl set to 230V as I read somewhere this "compensates" for the reactive errors in the household, no idea if thats right.

hope you can give me more insight into this so i can understand it fully

your explanation makes perfect sense to me, and is how I have been thinking the error is explained. I can't see how that 0.49A error is there constantly no matter the Current, or sine wave being produced

(which, as others have argued on mse, is the case).

also, i believe the Owl as stated by the manufacturer is less accurate at low current readings, and more accurate at higher ones (below 1A then not specified,1 to 3A then less than 10%,3 to 71A then less than 5%)

after 53 (30 Nov to 21 Jan) days from installation, my Gen meter was reading as near as damn it 300kWh, my monitor reading 420kWh set at 230V.

this was my "fag packet" maths:

assuming 13 hours "nighttime" each day, thats 0.113kW*13*53=78kWh

assuming ? hours no output/very low output during each day, no idea but guess 4? so thats 0.113kW*4*53=24kWh

420-(78+24)=318

318/300*100=6% error

i thought that put my "guess" of 6% error as within the stated error of the monitor by the manufacturer

does this make any sense? 

regards

Don

[Pat_]

I don't know if it helps, but assuming an indicated current of 100W and a voltage of 240V ( which is closer to reality than official nominal) that would represent a load of 576 Ohms if it were entirely resistive.  If it is in fact a capacitative load it would by calculation be 5.5uF. Without knowing anything about the filtering at the inverter output, it is not beyond the realms of possibilty that the circuit exhibits such a capacitance. It is a much better explanation than the other I have heard relating to 'digital noise'  because it is present when the inverter has shut down for the night, and therefore not producing noise.

On a British domestic import meter we pay for watts, and not volt-amps, so this perceived current is innocuous. We're still working on our reporting circuitry so I don't have any solution yet.

[don0301]

I don't know if it helps, but assuming an indicated current of 100W and a voltage of 240V ( which is closer to reality than official nominal) that would represent a load of 576 Ohms if it were entirely resistive.  If it is in fact a capacitative load it would by calculation be 5.5uF. Without knowing anything about the filtering at the inverter output, it is not beyond the realms of possibilty that the circuit exhibits such a capacitance. It is a much better explanation than the other I have heard relating to 'digital noise'  because it is present when the inverter has shut down for the night, and therefore not producing noise.

On a British domestic import meter we pay for watts, and not volt-amps, so this perceived current is innocuous. We're still working on our reporting circuitry so I don't have any solution yet.

isn't current measured in Amps?  

the inverter doesnt shut down, it goes into "night mode" according to my manual it draws 0.27W at night or when not "producing".

the error, as you say, is innocuous regarding payment. but over time there is a large error in kWh reading on the monitor, to the eqpt, which is whats being discussed. and for me, trying to be fully understood.  

if it helps, the manual states the inverter is 0.85 - 1 inductive/capacitive (depending on country setup or device specific settings!)