Please help me decide on which way to go. (enPhase vs SolarEdge vs LG)

[foo1bar]

From many previous posts we all know it is not that accurate - what the manufacturer writes on the side is not really meaningful. So you proceed to use it for comparisons?

Many people will and do use the report from their inverter as a measure of how much power they're producing.
You'll note that in this thread there is question of whether SE is inflating their reported numbers - that people are still speculating about it as they don't have real proof.
I was just adding an additional anecdote.

There are many meters available for residential use that are revenue grade - cheaper than TED for sure.

I couldn't find anything that would track four 240V circuits for less than ~$420.
I even asked on this website if anyone had suggestions.
Of course a week after I bought it, I see TED5000s with 3 inputs for $150 that could probably be upgraded to 4 for an extra $80.

[russ]

I couldn't find anything that would track four 240V circuits for less than ~$420.

Not there so I don't know - here I can go to the builders supply and buy them cheaply.

[SunEagle]

I bought a TED5000 and set it up a few days ago.
The TED5000 reports via pvoutput.org that for today (5/12) I produced 57.47kwh
The Solaredge monitoring website says I produced 58.54kwh
(1.8% higher reported by SE)

The TED5k only is measuring 1 leg (120V, not 240V), so there could be some error introduced there.
And I see yesterday 56.644kwh vs. 58.696kwh (3.6% difference) so it's not a constant % difference.
Yesterday and today had some clouds so some error could be due to sampling rate/size.

And it's quite possible that my TED is off - they claim +/-2% and SE claims +/- 5%.

But it seems likely that SE is calibrated to report slightly higher.

More than likely the two meters use different sample rates.

Also based on what they measure (volt & amps) the kWh calculating equation for SE may have less decimal places which tends to round up the numbers more then what the TED software uses.

[sensij]

More than likely the two meters use different sample rates.

Also based on what they measure (volt & amps) the kWh calculating equation for SE may have less decimal places which tends to round up the numbers more then what the TED software uses.

The SolarEdge data is not limited by sample rate or rounding error. I have data that proves this.

Measuring inverter output on one leg is probably fine. I am monitoring both legs on mine, and the difference is insignificant.

[SunEagle]

The SolarEdge data is not limited by sample rate or rounding error. I have data that proves this.

Measuring inverter output on one leg is probably fine. I am monitoring both legs on mine, and the difference is insignificant.

I guess the TED meter is either less accurate or the SE data might be inflated.

[sensij]

I guess the TED meter is either less accurate or the SE data might be inflated.

Yes, SE data is inflated; their energy counters, in particular. See this post earlier in the thread for an example, but there is substantially more data that supports the conclusion. It still complies with the 5% standard they need to meet, and even though more accurate data is already available in the inverter if you know how to get it, they choose to publish the higher number.

[J.P.M.]

Yes, SE data is inflated; their energy counters, in particular. See this post earlier in the thread for an example, but there is substantially more data that supports the conclusion. It still complies with the 5% standard they need to meet, and even though more accurate data is already available in the inverter if you know how to get it, they choose to publish the higher number.

FWIW, I think I agree w/Sensij on this one even though I'm unfamiliar with the SE reporting system.

Solar mfg./suppliers have nothing to lose by knowingly over reporting output to the extent of an over tolerance or more to the extent they can get away with it, just like vendors have nothing to lose and a lot to gain by underestimating array performance.

Users in states/areas or situations where utility grade meters are required - for SREC requirements for example, probably have a better handle on accurate output.

If output accuracy tol. is +/- 5 %, I'd put my money on closer to the -5% than +5 % every time and hope to be proven wrong when confronted with verifiable #'s.

I've not got a revenue grade meter (yet), but an educated guess puts my Sunpower reporting about 2-4 % higher than actual output. The array output is also about ~~ 2.8 - 3.0% higher than I'd otherwise calc., at this time probably due to panel over tolerance, which seems common for many reputable panel names.

Bottom line: monitor #'s coming from mfg./vendor monitoring equipment is from/by someone with skin in the game and therefore a stake in putting up as large an output # as possible.

Higher #'s look nice and feel good. I try to remember that not everyone who tells me what I want to hear is telling me the truth.

Follow the money.

[Ian S]

Just for kicks, I'm going to see how my Sunpower 7000 kwH display compares with the revenue grade meter for my solar production on the other side of the wall.I suspect it will take a couple of days of observation.

[J.P.M.]

Just for kicks, I'm going to see how my Sunpower 7000 kwH display compares with the revenue grade meter for my solar production on the other side of the wall.I suspect it will take a couple of days of observation.

I'm a bit surprised you haven't done that already. If I'd known/figured out you had a revenue grade meter, I'd have asked. I'll be interested to find out what you observe. Honest. I'd suggest the whole day output using the 5 min. increments after the dropouts show up. I'm sure you know the drill.

Regards.

[Ian S]

I'm a bit surprised you haven't done that already. If I'd known/figured out you had a revenue grade meter, I'd have asked. I'll be interested to find out what you observe. Honest. I'd suggest the whole day output using the 5 min. increments after the dropouts show up. I'm sure you know the drill.

Regards.

Both the revenue grade meter and the inverter display give total kWh as integers with no decimals. The revenue grade meter was changed out to an electronic one during year 1 so what I've done is take readings on both meter and inverter at the same time then I will wait for a couple of days to repeat then subtract to get the delta kWh for each. Over two days or so, the delta should be 85 to 90 kWh so should get an idea of how close the readings are. Now what did I do with the sticky note that I wrote the first set of reading on?!!!

re: dropouts, haven't noticed any lately.

[Ian S]

We have the revenue grade meter as the utility gets brownie points for the solar production - part of the deal in exchange for their $1/watt subsidy.

[sensij]

We have the revenue grade meter as the utility gets brownie points for the solar production - part of the deal in exchange for their $1/watt subsidy.

Thanks for looking. I only have access to SolarEdge so that is all I can post about, but PVOutput shows data from many systems and it is interesting to compare. I hope more people like you are out there, with revenue grade meters already installed, and will let us know how far off their inverter is.

[Nekota]

That test was rigged to generate a particular outcome, as many are. There should be very little difference in the output between SE and Enphase.

In what ways do you think this test was rigged? PVEL is an independent test lab with many members. It is interesting to note that SolarEdge is not one of it's members. I am not saying the test wasn't rigged, I'm curious in what ways you say it is rigged.

[Nekota]

So you have never seen an inverter efficiency curve

here you go.

[ATTACH=CONFIG]6695[/ATTACH]

And one for Enphase.

http://www.gosolarcalifornia.ca.gov/equipment/inverter_tests/summaries/Enphase%20M215-60-2LL-S2x%20(240V).pdf

[sensij]

In what ways do you think this test was rigged? PVEL is an independent test lab with many members. It is interesting to note that SolarEdge is not one of it's members. I am not saying the test wasn't rigged, I'm curious in what ways you say it is rigged.

I would start by questioning the array configuration. If it is truly a side by side test, and 12 panels are used in the enphase array, why not use 12 in the Solaredge array?