diy solar

diy solar

An Enphase Ensemble Installation

Watching the videos and found some interesting things:
  • the encharge is designed to have water (rain) run through a part of it as the IQ8s are NEMA 6 (underwater) and it doesn't matter how wet they are. The other electronics (and presumably battery) are in a water tight NEMA 3r enclosure.
  • You can daisy-chain them together, but restricted to a max of 8 gauge (max 40 amps).
  • The envoy is the "brains" that tell the batteries if they should be charging or discharging.
  • The IQ8s can operate in Grid-Tied (current source for TOU) or Grid-Forming (voltage source to supplement load) mode. Sounds like the Envoy switches it back and forth depending on the situation.
  • Communication to the batteries is via the Envoy.
  • A 10 kW Encharge is physically three 3.3 kW Encharge units. The downside of this is that the batteries in the 10 kW are not shared...each is discrete. So, if you want to add a DIY 10 kW battery to one of them you'll be limited to 1280W inverter it has. To get the 3840W from the inverter for that 10 kWh you'd have to augment all 3 sub units.
  • Losing the internet connection can affect the warranty (sounds like you have to be offline for over a year).
  • The Enpower reports grid and Encharge frequency to the Envoy, the Envoy has the brains to tell the batteries to adjust.
  • The IQ8s in grid forming mode are responsible for stabilizing voltage and frequency to what Envoy specified.
  • A cellular modem is required (although I don't believe this was quoted in my system...so old data?).
  • IQ8s in Grid forming mode are fast (20 ns) in terms of supplying current/voltage, much faster than power curtailment which is 2-3s and controlled by the Envoy.
  • They have smart breakers in their lab: Alexa: You only have 15 minutes power left, should I turn off your air conditioner?
 
Last edited:
Great recap Svetz. I haven’t made it through all the videos yet but so far i was disappointed to hear I’ll need the cell modem. The whole system seems like a sharp departure from the simplicity of past enphase products. However, i think they’ve put in the hard work and after it’s installed you won’t need to fiddle, massage or otherwise worry about the functionality...apart from manually shedding load in grid down situations to stretch capacity. Also, it hasn’t been clear to me until now, that both panels and batteries can contribute to total kW production simultaneously during an outage. Do i understand that correctly? Also, as far as communication PLC/Zigbee it was clear as mud...but i took it as everything can work independent of the other parts.

Thanks,
WL
 
......apart from manually shedding load in grid down situations to stretch capacity....
I'm working on that! See Features for the Perfect Monitoring Software for a Residential System, Smart Solar/Battery, and RF and the Smart Solar Home.

... it hasn’t been clear to me until now, that both panels and batteries can contribute to total kW production simultaneously during an outage. Do i understand that correctly?
Yes you do! Way slick eh?

... Also, as far as communication PLC/Zigbee it was clear as mud...but i took it as everything can work independent of the other parts.
As far as I know the envoy is powerline to the IQs and WiFi to the battery, the WAN, and the smart switch. I'm still not convinced the cellular modem is required. Why would it be if the Envoy can hold a year's worth of data waiting for the WiFi to come back up.

You can put a zigbee stick in the Envoy, looks like it's for their revenue grade meter. Not sure if it works with anything else.
 
The warranty/internet connection seems odd to me. So is enphase telling me I can operate it in a way to harm it or will enphase adjust charge/discharge rates to conform as needed to keep warranty terms intact.

I still like this system the best, but I want 350vdc+ for my battery bank still keeps putting me back to SMA storage or when ever the one from ossiaco.com is released but the more I talk to them the more I think these guys dont want to support an offgrid users yet.
 
Tesla does the same thing (requires an internet link, although supposedly to update firmware to get the 10 yr warranty). It raised a stink in Florida a few years ago as long internet outages due to hurricane damage would have voided their warranties (I believe Tesla extended the times for affected locations). The Tesla Powerwall might be a less expensive way to go with Enphase.
 
Taking Ensemble sizing webinar as i type. Saw this interesting question and answer. It makes a strong argument for waiting for IQ8 on the roof.

"Q: Can you install more than 1.5x PV / Batery with the system automaticaly curtailing the PV amount above the 1.5x factor? If not, then with the PV IQ8 models will this change?

A: You should maintain the ratio for PV to Battery. Outside this range, there can be circumstances wher the microgrid can't absorb the full production and cause instability in the microgrid.
With rooftop IQ8, there will not be any ratio restriction."
 
As the Air conditioner kicks in more frequently the daily power consumption should start to outstrip the solar generation. Put together a little chart from the data I've been collecting from ( Features for the Perfect Monitoring Software for a Residential System ) and it looks like the trend lines crossed a few days ago. Unfortunately the data's not to precise as the "day" starts from when I turn the PC on which varies (for example, I've not yet had a day where solar production was over 50 kWh. In retrospect I should have measured at a time my PC was always on or run it on a PC that is always on.

Capture.PNG
Update: Realized I could just divide by the time to get the daily average to smooth out the time differences... graph updated. The power usage looks awfully high, not sure if that's just the heat wave + stay-at-home or if something is starting to break.
 
Last edited:
Last couple of days I've been getting, and ignoring, the soiled panel alert. I figured it was a case of the tolerances being too tight. I'm sure they are, but I went up to the roof to check and sure enough, the local trees had deposited a fine layer of pollen and the last couple of inches were particularly laden (probably from a very light rain we had). To hot to rinse them off now, but I'll do it tomorrow morning.
I had no idea that washing the panels would cause it to rain. Had I just waited a day I wouldn't have lugged the hose up to the roof, which tripped the latch mechanism on the ladder causing it to collapse (while on the roof), nor scared the bejusus out of my wife by tapping on the upstairs window with the hose end so I could get off the roof, nor have her spin around so fast that she caused her full coffee cup to go flying. Good thing the dog house gets wifi.
 
Last edited:
Heard from my Ensemble battery installer...sounds like the earliest I can expect it is June and that really depends on what CA does with reopening.

Looks like all that pollen on the panels was costing me about 10% energy. Guess I shouldn't have ignored the soiled panels warnings. Anyway, the system produced 52 kWh yesterday and there were some clouds around noon so hopefully will do better today. Nice to finally break 50 on a 7.8 kW system.
 
Thanks for sharing your experience svetz, I'm getting an Ensemble IQ7 system (no batteries) on my house in the very near future and I'm excited to start making some electricity soon.
 
Ran across this image on the AltE store and it reminded me I hadn't mentioned it:

encharge_pairingchart_image2.png

As you can see, there's a maximum number of microinverters based on the number of batteries (Encharge) installed. This is one of the reasons I suspect the "brains" for controlling the inverters on the roof is inside the battery rather than the Envoy.

But, it's probably more about the problems incurred with AC coupling. We know with AC Coupling there's a time element to frequency shifting inverters such that they can safely handle sags (an appliance turns on) and spikes (e.g., the inverters are producing enough to power the Air Conditioning which suddenly turns off). That is, just as inverters have to handle start-up loads, AC Coupling must handle shut-down loads.

Sags are easy as the batteries can make up a power difference until the inverters ramp up. But there's only so much you can do with a power spike as the inverters ramp down. Also from the chart you can see the Encharge units handle more of the lower wattage microinverters than they do the higher (e.g., the IQ6 is 195W, the IQ7A is 366W). It looks like the maximum a 3 kW unit can handle is about 1800W for probably a tenth of a second. I bet there's some fabulous engineering in there, wish they'd publish a white paper on it like Outback did.
 
It's pretty amazing that even on the dark stormy days the LGs have been cranking out some power.

The worst day so far has been the 14th, 4.67 kWh produced that day, less than a tenth of a good day. Yes, it rained all day...but look at the numbers from the nearby solar radiation meter:

1590505089953.png

So from 1 to 2 about 175W possible were available. On a good day we usually cross that by 8:00 A.M. or 6:30 PM.

Given the panels are almost 20% efficient, you might expect to see as much as 35W/m^2.

The program recorded the data from the panel and calculated it's efficiency compared to the meter and adjusted for the panel's actual size, here's what it reported:

Panel Eff %Watts MeterWatts PossibleWatts ActualAmbientPressure inHgWind mphHumidity
13:02:2695.6180.462597430.053.4 E99
14:02:2671.2104.536257530.039.8 NE99
15:02:26042.91507530.014.9 ENE98
16:02:2656.159.320117530.003.8 E99

The columns are a bit out of order as the Panel Eff is calculated from watts actual / watts possible (numbers reported are rounded); it's first as that was what I was most interest in. These aren't exact of course as the solar sensor is about a mile distance, but interesting data none the less.

What really impressive about the LG panels is that they produced at all in these circumstances, they were generating power when it was as low as 60 Watts/m^2 ambient. That's just about their cutoff point, you can see when the ambient light fell to 42W/m^2 no power was produced.
 
Last edited:
In the data above, it's interesting to see how the panel efficiency starts to degrade with decreasing ambient light.
1590669215369.png
 
Those of you with properly tilted panels are now well past clipping season, but with my inefficient flat panel setup suspect I'm just starting to enter it as the solstice approaches. So, for the next 36 days I expect to see a few watts lost around solar noon (assuming it's not raining).

It was a down-side of going with the LGs, the IQ7s are 295W, the higher voltage 7Xs the Panasonics would have used are 320W and I probably wouldn't have seen any.

So digging into the logs what do I find? The first time my panels hit 295W was two months ago... solar noon on 4/20. So does that mean I have 4 months of clipping? On 4/22 I have the highest number so far at 301W, that's higher than the 7X's rating of 295W! From 4/22:

1 (east) : 292.0
2 (east) : 300.0.0
3 (east) : 285.0
4 (east) : 298.0
5 (east) : 294.0
6 (east) : 298.0
7 (east) : 295.0
8 (east) : 289.0
9 (east) : 291.0
10 (east) : 297.0
11 (east) : 301.0
12(east) : 299.0
13 (east) : 283.0
14 (east) : 297.0
15 (east) : 291.0
16 (east) : 296.0
17 (east) : 291.0
1 (west) : 296.0
2 (west) : 295.0
3 (west) : 297.0
4 (west) : 297.0
5 (west) : 296.0
6 (west) : 295.0










Since then, the period has been getting longer where I see 295+ W from the panels, right now it's peaking about 1:20 to 2:20 - so an hour. How much power am I losing per day? It's hard to say, but at a guess not much as I'd expect some of the inverters to reporting above 295 and the numbers are just around that now at solar noon.

The numbers in the table above were at an ambient of 830.7W/m2, I've seen as much as 1024W/m2 where the panels were only outputting an average 270 W, so it's either weather related (e.g., heat/wind) or I need a more local/accurate meter. Here's the weather data:

Code:
          Timestamp   Watts,  UV, Temperature, Dewpoint, Humidity, Pressure,  Wind,  Wind Gust, Direction
04/22/2020 13:41:42   830.70,   9,          86,     66.4,       52,    30.01,   2.9,        3.4,       ENE
06/01/2020 11:39:14  1024.60,  11,          89,     81.3,       79,    30.06,   2.0,        2.2,        SE

So, even though there was more light yesterday, on 4/22 the panels produced on average 30W more per panel. Considering the panels are flat, they have a better tilt angle now. It looks like the humidity might have a roll, but I can't see why, the solar meter should be reading the available energy level.
 
Weather forecast for the next 15 days has cloudy with possible thunderstorms, so looks like I don't have to worry about losing any power to clipping... ?

Still don't know why the panel readings aren't higher now that the panels are at a better tilt angle than 4/22. I cleaned the surface on half of them (there was bit of residue), but there's no detectable power difference.
 
Is Encharge defying the laws of LiFePO4 chemistry?

From the quick install guide:
Following local standards, choose a well-ventilated location where the ambient temperature and humidity are within -15° C to 55° C (5° F to 131° F) and 5% to 100% RH, non-condensing, preferably out of direct sunlight. The optimum ambient temperature range for installation location is 0º C to 30º C (32º F to 86º F).

So, from what I understand, charging LiFePO4 below 0º C can cause irreparable damage. Couldn't find anything regarding their having a low-temp cutoff or if they heat their batteries. That doesn't mean it's not there.

Mine will be outdoors, but this isn't a problem I'll encounter unless everyone switching to solar triggers an ice-age. Something my numerically higher latitude friends might need to consider though.
 
How Accurate is SAM?

It's based on average weather so I wouldn't expect it to be exact, but it should get closer the longer the time period is. Here's a quick comparison between SAM's prediction and actual harvested amounts:

SAM kWh/moActual kWh/mo
Feb890919
Mar12091306
Apr13491315
May14041317
Total-to-date48524857

Pretty scary that it's within 5 Wh over 4 months. Probably just means it's an average year so far. Supposedly we're going to have a bumper crop of hurricanes, so it'll probably be less accurate over the next 6 months.
 
Learned that Covid will again delay the battery installation... so July?
Covid is rampant in Florida just now, last three days were: 1700 new cases, then 1900, then 2300 new cases yesterday. No talk of a another shut-down AFAIK, but definitely not good. Stay safe out there!
 
From another thread (but posted here so I can find the data later):
...Anyone seeing any effects from the dust? Sky is real milky here in S Fla....
Yep... I probably live South of Brewman and I was wondering how it was affecting production. It looks like the peak of the curve was shaved off, still I'm better off then the previous rainy days. Overall probably probably only a couple % (2 to 3 kWh?).

Despite (or because of) the Saharan dust, panels were outputting until after 8pm last night (if you call 6W output ;-). More likely it's just because we passed the summer solstice and the days are still longish.

Code:
7:25:07, 6171648, 102W, 32.90W meter, Ambient= 86F, Pressure= 29.99 inHg, Wind Speed= 1.3 mph SE, Humidity= 84%
8:00:03, 6171718,  70W,  6.90W meter, Ambient= 85F, Pressure= 30.01 inHg, Wind Speed= 1.3 mph NE, Humidity= 85%
8:24:02, 6171724,   6W,  0.00W meter, Ambient= 85F, Pressure= 30.01 inHg, Wind Speed= 0.2 mph N,  Humidity= 85%
9:00:01, 6171724,   0W,  0.00W meter, Ambient= 84F, Pressure= 30.03 inHg, Wind Speed= 1.3 mph E,  Humidity= 85%

The Watt meter capped out at 907 W/m^2 just before noon, all in all not a bad day with an insolation of 6.5 (panels are at 0 tilt).
1593426768686.png
Next year, assuming we get a dust free year (dust from the Sahara happens every summer AFAIK, it just doesn't always hit us). You can easily see it from the GOES Satellite during when it's daytime over the atlantic.

The dust has actually been overhead a few days. Rain & clouds have been more of an issue this month:
1593428673686.png
 
Last edited:
Back
Top