Each night, my solar shows that it is generating about 5 watts. I’m almost sure there is no sun out at 12 midnight (I am not in the arctic circle) 
My belief is that the CT sensors for the solar are showing a small amount of current that is not really there, so, my suggestion is to offset that value (measured late at night) so that the value is zero if this is, in fact, the case.
With all the other “errors” I really could care less that it shows 2 watts at night for mine - that does not add up to a hill of beans
[quote=“Howard, post:2, topic:347”]
With all the other “errors” I really could care less that it shows 2 watts at night for mine - that does not add up to a hill of beans [/quote]
It’s not that bad in terms of data, but it is a distraction on the bubble screen, and may just be a zero set calibration issue that could be easily remedied.
technically I think that’s one purpose of calibration - to zero out no solar.
I’m afraid if you did “zero” it manually it would still fluctuate +/- 0. I think that flow is not entirely imaginary - it’s either imperfect zeroing out during calibration or actual watts. Sure, your panels may not be producing, but depending on how your wires are laid, electrons moving in one circuit could be generating enough of a magnetic field to push electrons in other lines. Not +/-5w, so I suspect the answer is an imperfect calibration.
Mine reads between (-)1 to (-)4w after dark
Actually I am surprised my solar reads 2 watts at night as I recall the inverters actually draw a small amount of power all the time and I would expect if not showing zero it would be a small negative number at night.
[quote=“Howard, post:5, topic:347, full:true”]
Actually I am surprised my solar reads 2 watts at night as I recall the inverters actually draw a small amount of power all the time and I would expect if not showing zero it would be a small negative number at night.[/quote]
I agree that is probably the case. I am absolutely sure they are not generating power from stars (except for the sun) and the moon/planets (even though I have seen both my Moon shadow and Venus shadow )
Hi there, thanks for reporting this. The situation you’re describing might actually be a bug that we’re already looking into. The solar inverters are drawing energy over the solar at night and the Sense app is reporting a positive value, which should be negative. I’ve already submitted this to our team to look at, so hopefully this will be fixed soon!
Not all inverters draw energy at night. In my case, for example, I use microinverters that are 100% powered from the panels themselves. At night, they shut down and draw no power whatsoever.
That said, I do have negative readings at night (-6W and -36W), but that is not from the inverters, but rather various always on equipment connected to the sub-panel that my Sense CTs are monitoring.
I think the solution is to allow the user to individually trim out this night consumption by allowing a correction to be applied to each leg, to get a 0w / 0w night reading.
According to a post in a solar forum, they said many microinverters don’t draw from the AC side, but by the DC side. Hence, there should be no draw once they drop below a threshold.
However maybe there is a partial load on the AC output, e.g. a transformer winding, but I doubt it.
Right, my microinverters only draw from the DC side as per my post above. I verified with a clamp on current probe on each of my 6 strings. My reading at night is 0.00A on the 20A scale (lowest scale my digital snap-on hand held probe does). So I guess it could have been 0.004A or less, which would translate to 0.96 watt or less. Close enough to 0 in my book.
Different inverters do different things. I have a Fronius 7500 unit that when the sun goes down it goes dark. The good side is it draws no power, the bad side is you can’t check the metering or other statistics when it is not online. I also have a Sunnyboy 2500 and 1100 and they draw a very small amount of power when off line thus you can read their meters at night.
[quote=“peter, post:10, topic:347”]
I verified with a clamp on current probe on each of my 6 strings[/quote]
How big or small were your 6 strings? My 4 strings are at most 12 panels of 345w each.
[quote=“Howard, post:11, topic:347”]
the bad side is you can’t check the metering or other statistics when it is not online.[/quote]
Do you mean the historical values are not available when the panels are not powered up? Wouldn’t that be stored on the cloud someplace?
What metering or statistics would be there if not powered up?
My inverters are not hooked up to the internet or anything like that. I have had my system for about 12 years and that was not a “thing” then. I did have a Sunnyboy control unit that did connect via a serial cable and log data to spreadsheets and the software on my PC had the ability to upload it to a Sunnyboy web site. I replaced two of my old Sunnyboy 2500 inverters with the Fronius unit. It does not have built internet capability but you can buy a highly overpriced module to allow that. I decided it was a much better value to buy Sense to accomplish the same for cents on the dollar. Further, even if I went that route with the Fronius I still could not capture the Sunnyboy data. Sense does everything I need to track my solar and usage.
I have 4 strings with 14 * 260w panels and 2 strings with 12 * 260w panels.
They did not make them that big when I installed mine. The largest were 185 watt panels which is what I have. I have two strings of 8 each on the 2500 watt inverter, 1 string of 6 on the 1100 inverter, and four strings of 8 on the 7500 watt inverter.
My microinverters communicate with a device known as a Solar Gateway. The sole purpose of the Solar Gateway is to communicate with each of the microinverters (over low power Blu-tooth) and collect that stats which are stored on the gateway, and then connect to my LAN via Ethernet. My switch powers the Solar Gateways via PoE.
So at night the microinverters are completely dead and the Solar Gateways report no connection to them. However, I can still talk to the gateways and pull historical data. The data is compressed on the gateway and it can store about 50 years worth of stats (doesn’t take much room). I can also pull the data from the gateways using its REST API and keep it all in a local database on one of my servers and create pretty graphs. Would be nice if Sense would expose a REST API so that I could suck data in from it as well. 
I can ssh into each of my gateways and run “suntop” to see how each microinverter is doing real-time like so:
I have some cloud cover this morning and there is some shade the larger roof mounted array, but the ground mount is in the clear so I’m getting better numbers from it at the moment.
Yes, I have 2 failed microinverters at the moment. The replacements are actually being delivered today (along with a couple of spares).
[quote=“peter, post:17, topic:347”]
I have 2 failed microinverters at the moment. The replacements are actually being delivered today (along with a couple of spares)[/quote]
OT, but what type of microinverters are you using and how many failures over how many years?
I’m using the system from these guys:
System has only been in operation since December 2016 and I have had a total of 5 microinverter failures (out of 80) in December and early January. I got a very early batch of inverters and the infant mortality rate was very high at 6.25%. I don’t anticipate any more failures (fingers crossed). Vendor has made some changes to the manufacturing process that should improve the reliability going forward.
I too, am in the Infant mortality period, as my solar turned on on 12/20/2016. My individual monitoring had to wait for some electrical work be done, and I am waiting for some programming to enable that feature (any day now!).
Your failure rate seems high, but your microinverter company may not have as many out in the field (like Enphase) and thus may be experiencing growing pains. Best of luck once you clear out the bad ones!
