Because your voltage reference is only one leg of the split phase, the two will always be opposite and have different signs. I’m pretty confident this is the standby loss of your inverter and will reverse when the dawn breaks.
I’m thinking I started out with the CT on the black wire. The solar was installed about 6 months ago and LOTS has happened since then clouding my memory. I want to say I’ve tried this before. If the numbers go further negative when the sun is shining, I’ll have to change the CT around so the arrow is pointing opposite of power flow? I don’t want to reverse it in configuration. Would that mean the CT is connected to the opposite leg as the voltage reference? And if I do have to reverse the orientation, that would make the nighttime wattage be positive again. If I’m remembering correctly, I think this is why I moved the CT to the red wire. The arrow on the CT was in the correct direction and the readings were correct without using the reverse function. I originally installed the CT on the black wire and did it when the sun was shining. I was looking at the iotawatt in real time when I clipped it on and it showed negative so I turned it around with the arrow pointing the “wrong” direction for daytime power flow. Then on that first night, I saw “output” when there shouldn’t be. That’s when I allowed negative values. My memory may not be entirely accurate. The CT is currently on the black wire with the arrow pointing from the array towards the meter and it’s currently showing negative 18 watts. I have 26 panels and 26 micro inverters. Since the output is doubled, it looks like the inverters are using 9 watts an hour. I can go on my solar app at night and see data from the inverters so I would agree they’re using power at night. But why didn’t the mains see it when I opened the solar breaker? The solar CT went to zero when the breaker was open but the mains never changed. Not one watt. I went to lengths to make sure the house load was near perfectly stable.
Let’s wait and see what the dawn brings. There is another possibility that we can explore.
It’s running further negative.
OK, that’s not it. For now you can either put it back on the other wire or, easier, just check reverse in the CT setup to go back to positive solar power.
At this point, I’d like to take a look at what is going on with the neutral wire. Is it possible that you could put a 50A CT on it?
I can pull the CT from the black and put it on the neutral unless you require a CT on the black while a CT is on the neutral.
Just trying to eliminate a remote possibility. let’s leave that for now.
Can you go back a few days with Graph+ to a time during the night when the +18 was being reported? Select out (just highlight on the graph) a period of an hour or less so the intervals are 5 seconds, then for the solar plot as Watts, VA, PF and VAR all on the same plot. Click CSV at the bottom and post the plot along with the first dozen or so lines of the CSV.
I think this is probably a low power reactive load issue and may want to take a cycle sample tonight.
I’m not a power engineer and so won’t pretend to completely understand what this. I’ll explain my understanding of what’s happening, which may not be completely correct.
The IoTaWatt is coming up with a huge VA of 682 VA. That means the product of the RMS Voltage and RMS Current is 682. With the x2 measurement that would mean the VA is 341. We know the RMS voltage is 120 Volts, so the RMS current must be about 2.84 Amps.
That may seem incredible, but with a power of 8.79 Watts, the power factor is .026. That, along with the reactor power measurement almost equal to the VA, tells me that the inverter is constantly using a lot of current and giving it back, every cycle. There may be a few Watts consumed if the action is causing heating of the inductor. 8 Watts isn’t much, think nightlight. Also, 8 Watts is only about 2.5% of the VA, so it could be more of a wash but not measured more accurately because of the very low values.
I’d like to see one of those cycles. When the system is inactive and stable at 18 Watts, could you enter this command in your browser:
Where n is the input number of your inverter CT that is measuring 18W.
You will get a response of about 640 sample pairs. Please copy and paste to a notepad file and upload.
samples 642
3,-97
21,-97
40,-97
55,-97
71,-97
87,-97
102,-96
118,-96
134,-95
149,-94
163,-93
178,-92
193,-91
208,-90
222,-89
234,-87
248,-87
262,-86
277,-86
291,-86
307,-87
322,-87
337,-88
353,-89
368,-89
382,-88
393,-84
406,-83
420,-81
433,-78
447,-77
460,-77
473,-76
485,-76
497,-76
510,-78
523,-78
537,-79
550,-79
563,-80
576,-80
590,-80
603,-79
616,-78
629,-77
642,-77
656,-76
669,-76
682,-76
694,-76
707,-77
720,-77
733,-78
746,-80
760,-81
774,-81
788,-82
801,-82
815,-82
829,-81
842,-80
855,-80
869,-79
883,-79
896,-78
909,-78
922,-78
935,-78
949,-78
962,-78
974,-78
988,-79
1001,-78
1014,-78
1027,-77
1041,-77
1054,-76
1067,-75
1079,-74
1091,-74
1103,-73
1116,-72
1129,-72
1142,-72
1154,-73
1166,-73
1179,-72
1191,-72
1203,-72
1216,-71
1228,-70
1239,-69
1252,-69
1263,-68
1274,-67
1285,-66
1296,-64
1307,-63
1317,-61
1327,-60
1337,-59
1348,-58
1359,-57
1370,-57
1380,-55
1389,-54
1398,-52
1407,-49
1416,-46
1425,-40
1434,-36
1442,-35
1449,-31
1454,-31
1459,-31
1464,-30
1468,-31
1472,-31
1477,-30
1481,-30
1486,-30
1491,-28
1495,-27
1499,-25
1503,-22
1506,-20
1509,-17
1512,-14
1515,-12
1518,-12
1520,-12
1522,-12
1525,-12
1527,-12
1529,-12
1531,-11
1533,-10
1536,-9
1537,-6
1539,-4
1540,-3
1541,0
1543,0
1544,0
1544,0
1544,0
1545,0
1546,2
1547,3
1547,3
1547,2
1547,3
1547,3
1548,3
1548,3
1547,4
1547,3
1547,4
1548,4
1547,3
1547,3
1547,4
1547,3
1547,3
1548,4
1548,3
1548,3
1548,3
1548,2
1548,3
1548,3
1548,4
1548,4
1548,3
1548,5
1547,5
1547,6
1547,6
1546,8
1545,8
1544,10
1544,12
1542,13
1539,15
1536,17
1533,19
1531,20
1529,21
1527,21
1525,21
1522,20
1519,20
1514,19
1511,18
1510,22
1507,20
1506,23
1503,23
1498,25
1493,24
1489,30
1485,30
1481,33
1478,33
1473,35
1465,32
1459,32
1454,36
1449,37
1446,40
1441,38
1433,40
1425,42
1417,45
1411,49
1404,49
1397,52
1388,53
1378,54
1369,55
1360,57
1351,59
1342,59
1334,61
1325,63
1314,64
1303,63
1292,65
1281,67
1270,67
1260,68
1249,70
1238,70
1227,71
1216,73
1204,73
1192,75
1179,75
1166,76
1154,77
1142,79
1130,79
1117,78
1104,80
1091,81
1077,81
1064,82
1051,84
1037,84
1023,85
1009,85
995,85
981,86
967,86
952,86
938,86
925,86
911,86
897,86
882,85
868,86
855,85
841,85
827,85
813,85
799,85
785,84
771,84
757,83
744,83
731,83
717,82
703,82
689,81
676,81
662,81
649,81
635,81
622,82
609,82
595,82
582,82
568,83
554,83
540,83
527,83
513,83
499,83
485,84
472,84
458,84
444,84
430,84
416,84
402,85
388,85
374,85
359,87
345,88
331,88
316,89
300,89
286,90
271,90
256,91
241,92
226,92
211,93
196,93
180,93
165,93
150,94
134,94
118,95
102,95
86,95
71,95
56,94
41,95
25,95
9,94
-7,94
-25,94
-43,94
-59,94
-74,94
-89,94
-105,93
-120,93
-135,92
-151,91
-165,89
-180,89
-195,88
-209,87
-222,86
-235,84
-248,84
-262,83
-277,83
-292,83
-306,84
-321,85
-336,85
-352,86
-367,87
-379,84
-390,81
-403,79
-417,77
-431,75
-445,74
-458,73
-471,74
-483,74
-495,74
-508,76
-521,77
-534,78
-547,78
-561,78
-574,79
-588,79
-601,77
-614,77
-627,76
-640,76
-653,75
-666,75
-679,75
-692,75
-705,76
-719,76
-732,77
-745,79
-759,80
-772,80
-786,80
-800,81
-814,80
-827,80
-841,79
-855,78
-868,78
-882,78
-895,76
-908,77
-922,77
-935,77
-949,77
-962,77
-975,77
-988,77
-1001,77
-1015,76
-1028,76
-1041,76
-1055,75
-1067,74
-1080,73
-1093,73
-1106,72
-1119,72
-1131,72
-1143,71
-1156,72
-1168,71
-1180,69
-1193,69
-1205,69
-1217,68
-1229,68
-1241,67
-1253,66
-1264,66
-1275,65
-1286,63
-1296,61
-1306,59
-1316,58
-1326,56
-1337,55
-1348,55
-1358,54
-1368,53
-1378,52
-1387,50
-1396,49
-1405,47
-1414,43
-1423,38
-1432,34
-1439,33
-1446,28
-1452,29
-1457,29
-1462,30
-1466,31
-1470,30
-1475,29
-1480,26
-1484,27
-1488,27
-1493,25
-1497,22
-1501,21
-1505,21
-1508,17
-1510,16
-1513,12
-1516,9
-1518,9
-1521,9
-1523,9
-1525,9
-1527,9
-1530,9
-1532,8
-1533,7
-1536,6
-1538,4
-1539,2
-1541,0
-1542,0
-1543,0
-1543,-2
-1544,-2
-1545,-2
-1545,-3
-1546,-3
-1546,-3
-1547,-4
-1547,-4
-1548,-4
-1548,-4
-1547,-5
-1547,-5
-1548,-5
-1548,-6
-1548,-6
-1548,-6
-1547,-5
-1547,-5
-1547,-6
-1548,-6
-1548,-6
-1547,-5
-1548,-5
-1548,-5
-1547,-5
-1547,-5
-1547,-6
-1547,-6
-1547,-6
-1547,-7
-1546,-7
-1546,-7
-1546,-8
-1545,-10
-1545,-10
-1544,-12
-1543,-14
-1541,-14
-1539,-15
-1537,-17
-1535,-18
-1532,-19
-1530,-21
-1528,-22
-1526,-20
-1525,-23
-1522,-23
-1519,-22
-1516,-22
-1511,-23
-1507,-20
-1506,-23
-1505,-26
-1501,-25
-1498,-25
-1493,-28
-1487,-29
-1483,-30
-1479,-33
-1475,-35
-1472,-34
-1468,-36
-1462,-37
-1455,-38
-1449,-37
-1443,-38
-1438,-40
-1432,-42
-1426,-45
-1420,-46
-1413,-48
-1405,-51
-1397,-53
-1389,-55
-1380,-56
-1371,-57
-1363,-59
-1354,-60
-1345,-61
-1335,-61
-1325,-62
-1315,-63
-1305,-65
-1295,-66
-1285,-67
-1274,-67
-1262,-69
-1251,-71
-1241,-73
-1230,-75
-1218,-76
-1206,-78
-1194,-79
-1182,-79
-1169,-80
-1156,-80
-1143,-80
-1130,-80
-1118,-81
-1105,-81
-1092,-82
-1079,-84
-1066,-84
-1052,-86
-1038,-87
-1024,-88
-1010,-88
-996,-89
-981,-90
-967,-90
-953,-89
-938,-89
-924,-89
-910,-89
-896,-89
-882,-87
-868,-87
-853,-87
-840,-87
-826,-88
-812,-88
-798,-88
-784,-88
-769,-87
-755,-87
-742,-86
-728,-86
-714,-85
-700,-84
-687,-83
-673,-83
-660,-83
-647,-82
-634,-82
-620,-82
-607,-83
-593,-83
-580,-83
-567,-84
-553,-84
-540,-84
-526,-84
-513,-84
-499,-83
-486,-84
-472,-84
-459,-83
-445,-84
-432,-84
-418,-84
-404,-85
-390,-85
-376,-86
-362,-87
-348,-88
-333,-88
-319,-90
-305,-90
-291,-90
-276,-91
-261,-92
-246,-92
-231,-92
-216,-92
-201,-92
-186,-92
-171,-93
-156,-93
-141,-94
-125,-95
-109,-95
-94,-95
-79,-95
-65,-95
-49,-95
-33,-96
-17,-96
-2,-96
Here is what the signal looks like.
The blue trace is the voltage signal, and the red trace is the current signal. Don’t know if I can explain what is going on here but I’ll try.
Ordinarily when the voltage is positive the current is also positive and when the voltage goes negative the current goes negative. That is strictly true for a purely resistive load like an incandescent light or a toaster. When there is a capacitive or inductive load, those components store a portion of the energy and return it at some other time. Motors and transformers will typically do this to some extent.
The total energy is called VA - the product of RMS voltage and RMS current. In that purely resistive load, VA = Watts.
In this graph, when the voltage is positive, the current is about half positive and half negative. Same when the voltage is negative. So nearly all of the energy consumed is returned to net zero energy used. The voltage and current are roughly 90 degrees out of phase.
VA is 862, but Watts is only 18. That means practically all of the energy is returned as what is called VAR or Volts Amps Reactive. The Watts used is called Real Power and is what you are billed for.
There may be a small amount of net energy consumed as heat. 18 Watts is about what a couple of LED lights release, but most of that 18 Watts is probably measurement error. Your experiment to turn off the breaker and watch the mains demonstrated that.
There is no fix for this right now. An electrical engineer could probably add a capacitor somewhere to cancel it out, but I don’t know how to do that.
You are not the only user that experiences this, but it i usually only a few of Watts. I’m considering adding a threshold power setting for each input below which the power recorded would be zero. There is currently a hard coded 2 Watt threshold.
Thank you. You’re light years ahead of me on understanding electricity. A threshold power setting would definitely be one way to solve it. Like you said, it’s more of nuisance than anything. I looked for some kind of timer switch that worked with the CT plugs thinking maybe I could set it up where a switch opened at night and closed in the morning. Totally different subject but I’d love to have a time function on the iotawatt where I could track how long an input had been above a certain watt or amp setpoint. I’d love to track runtime hours on some equipment. Maybe a time function could also be used for dealing with errant solar inputs as well… Anyway, thanks again for your help.
I’m considering adding a threshold power setting for each input below which the power recorded > would be zero. There is currently a hard coded 2 Watt threshold.
That would be great!
Did anything come of this?
I can easily filter the phantom 4W my solar panels “produce” at night in Home Assistant but it would be nice to do it at the source.
Edit: the subtract 4 and max zero will work for me.
For any others that may read this later, just to clarify, the so calle “phantom” power that inverters draw at night is really standby power used by the inverter when it is not producing power. There are some conventions in IoTaWatt installations that make this difficult to deal with, but it is really a legitimate thing.
BACKGROUND: With split-phase power, IoTaWatt users measure only one of the phases. The other phase is assumed to be a mirror image - or 180 degrees out of phase. When a load on the non-monitored phase is measured, the math produces a negative number. For convenience, IoTaWatt by default will use the absolute value, thus making those measurements positive. There is a setup option to “allow negative values” for any input. Most folks with solar check this box for the mains CTs to be able to identify the negative exported power.
Most other inputs should never legitimately go negative, but the solar inverter is an exception. They provide power (+) when the sun shines and consume some standby power (-) when it doesn’t.
If I were to put on my bean-counter hat to treat the load-center as an account, I’d setup an IoTaWatt with all of the CTs oriented, either physically or using the “reverse” option, so that inputs supplying power are positive (credit) and inputs consuming power are negative (debit). I would select “allow negative values” for all inputs. Now a negative main is an export (the grid is consuming power) and a positive main is an import (you are consuming power from the grid). Similarly, when the Solar Inverter input is positive you are consuming power from the solar panels, and when it is negative you are providing the inverter with standby power.
Of course, all of the loads will be negative as you are “exporting” power to them. Now everything is in balance. In theory, if all of the breakers in a panel are measured, the sum would be zero. More probably, the sum of all of the inputs would be a positive number indicating a surplus of power which is the absolute value of the unmeasured loads.
The panel would “balance” and there would be no need for subtraction or sign manipulation.
As a practical matter, nobody does this including me, but if you want to consider the standby power more appropriately, there are some approaches besides zeroing it out (which incidentally reduces your generated power as well).
Check “allow negative power”. Now when you use Solar in a script, if you are measuring production use (Solar max 0). If you have something that adds up your loads, such as a “misc” or “unmeasured” calculation, you can subtract any inverter standby load by adding (Solar min 0).
1 Like
That is indeed a neat solution. 
I forgot to consider that Solar, with allow negative values, is not monotonically increasing. To be completely accurate it would need to be integrated:
solar Wh = Solar
Then rather than (Solar max 0) you would use solar.pos and for (Solar min 0) you would use solar.neg.
1 Like
I think the way I am using it it will be ok.
I’m using the output (Solar max 0) as my solar production which will be monotonically increasing as it can’t go negative.
My Grid integrator will pick up the inverter self load at night.
Edit: is there a reason we can’t use outputs in outputs?
Is it to prevent circular references?
I needed to change these to (Solar max 0) as well. It would have been easier if I could have used the solar_prod output instead I had to replace (Solar) with (Solar max 0).
Update: it was not OK.
At the end of the day the solar energy would decrease.
Using the integrator is the solution.
1 Like