We had a roof top system installed in 2013, with an Aurora 3.6 kW inverter. In order to get paid for Solar Renewable Energy Credits (SRECs), they installed a Locus Energy "Revenue Grade" power metering box because they said the numbers from the inverter weren't accurate enough. The Locus box uses an Ethernet connection to report the power generated to a web site, and the web is the only readout we have. Originally, we were paying several hundred dollars to Locus for 5 years of on-line monitoring. They also automatically reported our numbers to the SREC folks so we would get paid. We renewed this arrangement after the first 5 years. At the 10 year mark, we were no longer eligible for SRECs, and can only get paid (at a much lower rate) for what we generate. It would cost over half of what we'd get paid to continue the arrangement with Locus, so we canceled that. We now get a reading ourselves and report that directly to the energy credit folks.
As it turns out, we can still access the Locus web site and get readings that way. It's not clear if that will continue if they notice that we aren't paying them anymore. I've been carefully tracking both the Locus numbers and the inverter numbers to come up with a "correction factor" so we can use the inverter numbers in case they pull the plug on the web site.
With this additional scrutiny, I've noticed something that makes no sense. We can get readings off the web throughout the day of hourly power generated. If I check the total power generated for the day in the evening, and then check it the next morning, our power generation takes a large hit overnight! If the inverter is running, but generating less power than it draws, I do occasionally see small negative power readings, but the reported overnight drop is multiple kWh. Most of the night, the inverter is in a very low power standby mode (like 6 Watts).
It gets weirder... They don't just scale all the hourly power numbers generated down by some factor. The power generated at various times during the day gets chopped by differing amounts. Here's a plot of the power generated for March 4th, as reported in the evening:

Here's a report I ran 2 days later after they had fudged the numbers:

Most of the 4.49 kWh that went missing is getting taken out of the 11:00 reading, but not all of it. The 1:00 reading also took a big hit, but not the 12:00 reading.
Over the course of December, I collected the evening readings, and then the readings the next day for the previous day. If I total up the power it claims I had generated every evening, I get 92.6 kWh. The total Locus claims for the month (which appears to be based on the next day readings) is only 51.48 kWh! That's the number the REC folks want us to report. Ordinarily, I would claim that the meter is clearly busted, except that all of the monkey business appears to be happening in their web system overnight. if I was into conspiracy theories, I'd think that Locus was getting paid off by the SREC folks to under-report our power generation.
What is equally screwy is that the total power the web system reports each month is within ~ 5% of what the inverter says it generated. Based on that, I would think the Locus box isn't too far off. I've got three different ways of estimating the power:
1) Total power generated calculated from the evening readings from the Locus box. This WAY bigger than the other numbers.
2) Total power generated according the Locus web site a day later (it doesn't change after the first night). This is the smallest number each month.
3) Total power generated according to the inverter. This is only slightly more than what the Locus web site reports at the end of each month.
None of them match, and the biggest discrepancy is between the Locus web readings from one day to the next.
If I had discovered this back when we were getting ~ $700 a year for SRECs, I would have quickly had a chat with Locus about why their web system was randomly "stealing" hundreds of kWh in the dark of night. As it is now, we're getting paid peanuts, and I don't want to raise a ruckus and have them just pull the plug on the web site.
HOWEVER, if anyone else out there is using a Locus system and their web site to collect real $$, I STRONGLY recommend double checking that your system doesn't pull any slight-of-hand with your power numbers overnight. If they are, and you complain, I'd LOVE to hear their explanation.
Thanks for wading through all of this!
As it turns out, we can still access the Locus web site and get readings that way. It's not clear if that will continue if they notice that we aren't paying them anymore. I've been carefully tracking both the Locus numbers and the inverter numbers to come up with a "correction factor" so we can use the inverter numbers in case they pull the plug on the web site.
With this additional scrutiny, I've noticed something that makes no sense. We can get readings off the web throughout the day of hourly power generated. If I check the total power generated for the day in the evening, and then check it the next morning, our power generation takes a large hit overnight! If the inverter is running, but generating less power than it draws, I do occasionally see small negative power readings, but the reported overnight drop is multiple kWh. Most of the night, the inverter is in a very low power standby mode (like 6 Watts).
It gets weirder... They don't just scale all the hourly power numbers generated down by some factor. The power generated at various times during the day gets chopped by differing amounts. Here's a plot of the power generated for March 4th, as reported in the evening:

Here's a report I ran 2 days later after they had fudged the numbers:

Most of the 4.49 kWh that went missing is getting taken out of the 11:00 reading, but not all of it. The 1:00 reading also took a big hit, but not the 12:00 reading.
Over the course of December, I collected the evening readings, and then the readings the next day for the previous day. If I total up the power it claims I had generated every evening, I get 92.6 kWh. The total Locus claims for the month (which appears to be based on the next day readings) is only 51.48 kWh! That's the number the REC folks want us to report. Ordinarily, I would claim that the meter is clearly busted, except that all of the monkey business appears to be happening in their web system overnight. if I was into conspiracy theories, I'd think that Locus was getting paid off by the SREC folks to under-report our power generation.
What is equally screwy is that the total power the web system reports each month is within ~ 5% of what the inverter says it generated. Based on that, I would think the Locus box isn't too far off. I've got three different ways of estimating the power:
1) Total power generated calculated from the evening readings from the Locus box. This WAY bigger than the other numbers.
2) Total power generated according the Locus web site a day later (it doesn't change after the first night). This is the smallest number each month.
3) Total power generated according to the inverter. This is only slightly more than what the Locus web site reports at the end of each month.
None of them match, and the biggest discrepancy is between the Locus web readings from one day to the next.
If I had discovered this back when we were getting ~ $700 a year for SRECs, I would have quickly had a chat with Locus about why their web system was randomly "stealing" hundreds of kWh in the dark of night. As it is now, we're getting paid peanuts, and I don't want to raise a ruckus and have them just pull the plug on the web site.
HOWEVER, if anyone else out there is using a Locus system and their web site to collect real $$, I STRONGLY recommend double checking that your system doesn't pull any slight-of-hand with your power numbers overnight. If they are, and you complain, I'd LOVE to hear their explanation.
Thanks for wading through all of this!