diy solar

diy solar

An Enphase Ensemble Installation

Yes, an isolation transformer would do the trick, but that is a big cost. I only have a 100 amp main, but that would still be a 24 KVA transformer. The name plate picture is perfect at 25 KVA. 265 pounds, wow. But even more scary, 115C rise!! At 40C ambient, that is 155C winding temp at rated power. How many watts are being turned into heat?

If they do start billing on the two legs independently, I will just work harder at balancing my loads. Maybe switch a few things to run on 240 volt. All My PC's can do that. The clothes washer and dryer are 120 volt, but since they each only pull 350 watts, they are on the same circuit. I could split them and have on on each leg to balance it better.

Here is a snap shot of what is happening right now.
XW-status2.PNG
The load side has -2,995 watts. That is the Enphase solar - all my load in the backup panel.
The AC1 "grid" is getting 82 watts. And 2,690 watts is charging the battery.
The load side shows the imbalance of loads in the backup panel. -13.5 amps on L1 and -8.2 amps on L2 for a 5.3 amps difference. That difference carries back to the AC1 input as well. 2.1 amps vs -3.3 amps. 5.4 amps difference.
 
GXMnow said:
...Maybe switch a few things to run on 240 volt...
Click to expand...
Be sure to get one with a 3 pin plug. When I first saw the graph to the
right I thought, that's impossible, it's a 240V drier. When I looked, I saw
it was a 4 prong plug (L1, L2, neutral, & ground). Looks like they use it as
120V and 240V.

That imbalance is 8 to 10 amps. L1 Dropping at 10:55 is one thing. But L1
and L2 reversing at 10:57:09 was surprising. But that imbalance... it's all from
one 240V appliance (whirlpool).		1675283990697.png
 
Last edited:
Our old-school electric dryers operate motor and timer off 120V, heating element off 240V. Optionally wired to 120V plug, 1/4 the power.
Used to be 3-prong, with neutral and ground sharing a pin. Now 4-prong is proper, but they still get used both ways depending on house wiring.
After a couple tries, I successfully rewire mine for 4-wire and got it to work on a 2-pole 30A GFIC breaker (laundry is on the porch, exposed to blowing rain.)

Motor is probably 5 or 7 A continuous, heating element maybe 20A intermittent. (half that on "low", I think. Two elements.)
I'll consider condensing for the new place I'm setting up, but will also put in external vent so I can get a $100 used conventional dryer instead of $1000+ new one.

An earlier model 240V Miele dryer and washer set had washer plugged into dryer. I presume when it detected washer drawing heater current, dryer switched off its heater, so they could be on a single circuit.
 
GXMnow said:
Yes, an isolation transformer would do the trick, but that is a big cost. I only have a 100 amp main, but that would still be a 24 KVA transformer. The name plate picture is perfect at 25 KVA. 265 pounds, wow. But even more scary, 115C rise!! At 40C ambient, that is 155C winding temp at rated power. How many watts are being turned into heat?
Click to expand...

This transformer is potted.
Transformers with low temperature rise are few and far between. Most run the windings hot enough to fry an egg (at full load). Exterior probably much lower, I would assume not over 65 C. Transformers are also driven part way into saturation; I've captured scope traces of current waveforms which peak several times higher than a sine wave should be, under no-load.

This one I finally got after lots of shopping, think it cost me $450 + tax. Most are East Coast, and shipping costs as much as purchase.
I had another purchase to pick up in San Diego area and was able to get this transformer as well. Very difficult to lift, we walked it onto a pallet and blocks, then the seller did most of the grunt work getting it into my pickup, with 2x4 ramps. Unloading with an engine hoist was a trivial one-man job.

This one is exactly the size I wanted, going to use it for derived neutral. Most of the time might run up to 60A, so 1/3 the temperature rise.
 
Updated the OP with the most recent data... Bumper yields so far in 2023:
1677693047948.png
 
1 Year Token!
This is interesting!
Was just in playing with the Enphase Token software I wrote and happened to notice the expiry seemed large compared to the generation.

XML: 
{
   "generation_time":1677693155,
   "token":"...",
   "expires_at":1709229155
}

The generation timestamp was today, but the expiration decoded to: Thursday, February 29, 2024 12:52:35 PM, so it's now a 1 year token rather than a 6 month token. That's better, but I still wish they'd get rid of it - a 1 year token does me little good if the business goes belly up.



And just because it hasn't been in a recent post... here's the thread index again...

Thread Index
SystemOperationMisc
Note: As the thread is a blog of experiences, you may have to hunt a bit above and below the link for the full story.
 
Almost hit 250% Production vs. Consumption!

1679226634706.png​
It was a combination of cool weather where the AC didn't come on and an exceptionally sunny day.

We haven't done as well as we usually do as the winter has been warm this year, we've run the A/C a lot (a couple of hours a day in the late afternoon) and only ran the heat once in January as you can see from the temperature graphs for January and February below.

1679227624286.png

1679227794603.png

The panels are flat, so not able to take full advantage of the equinox.

Solar's been pretty good the last week:
1679228107143.png
 
Last edited:
Hedges said:
"rainfall piezo"
Click to expand...
More commonly referred to as a haptic rain sensor.
The amazon page says Accuracy : TBA; Resolution : 0.1mm, rainfall here is so variable you'd need them side by side to tell.

I know it rained that day as I was out in it. A weather underground station to the north 1/2 mile reported nothing and a station a mile to the south reported 30". The 30" is obviously wrong for this time of the year, but that's the nice thing about a haptic sensor is it's not going to get wildly out of calibration as gunk collects on the bearings. One of the reasons I picked the WittBoy was no moving parts... haven't had it a year yet, but no problems so far - typically they break for me before the end of year two as the salt's wicked on them. There are similar more expensive models people have had for years that they swear by.

Hedges said:
Wonder how wind (raining sideways) affects it.
Click to expand...
It does know the windspeed, so it could do a correlation, but doubt they do.
 

Bluetooth​

Captured some Bluetooth packets and found Enpower, Encharge, Envoy, and some mystery packets.

The Encharge packets have 9 bytes of data followed by the device name and device serial number.

For example, convert the highlighted packets to ASCII you get "Encharg".		1679340077261.png

Unfortunately, the first 9 bytes match the other batteries and the Envoy even with the batteries at different states of charge and activities (charging vs. discharging).

Packets are 5 to 20s apart for the same device, so that's fairly frequent for something that doesn't seem to be changing much.

1679340237215.png 1679340799271.png
 
Last edited:
Switched my system to self-consumption to fully cycle the battery, figured it was time for them to be balanced. About 30 minutes before solar noon I check and there was no power going into the batteries. Two are at 100% and two are at 70%. Called tech support and they said the two at 70% were actually full based on voltage, so they recalibrated them so the GUI would be correct.

I asked if the system ever self-calibrates, and he said it occurred when on full backup after balancing completes (actually he said "after a while", I'm just assuming the balancing).

I don't think this makes a difference to the amount of energy actually available, I suspect it's just the GUI reporting as the batteries have been in savings mode for the last few months. I'll deplete them tonight to verify and let you know if I don't get the expected number of watt-hours out.

Update: Expected to see 10.75 kWh discharged, measured 10.3. I'll charge them up to full and let them balance, then repeat the test.
Update2: Repeated the test from full to 20%, discharge 10.3 kWh, but there should have been 10.752. So, missing 452 Wh or a 3.3% loss? The warranty is at least 70% at 10 years; so guess I'm ahead of the curve still.
 
Last edited:
Just checked and seems like I'm still at the same versions as last November. Wonder if they're hung up again?

IQ Gateway (Envoy): D7.3.121
IQ System Controller (Enpower): 1.5.4813_rel/22.95
IQ Battery (Encharge): 2.0.5441_rel/22.10
Microinverters:
  • 540-00142-r01-v04.28.03
  • 520-00095-r01-v04.28.07
    Profile Set (IEEE 1547:2015-WHB)
There are release notes for a 7.4.29, but it looks like it's only for Italy. 7.4.28 is Germany & Belgium.

Spoke to a gentleman in Austria today, he said his MyEnphaseView (web app) didn't have a slider for the "Battery Shutdown Level". I didn't know the web app was different based on country.

Also updated the OP with March's production, lower than prior years but not off by much.
 
There's an official document now for generating a token!

Even better, they list some additional REST API commands! I've updated post #17 to keep all the known API commands in one place. Of the five new commands, only 1 seems to have new data (although as these are "official" public API commands it makes sense to shift to them as the others are undocumented and more susceptible to change).

The new command is https://envoy.local/ivp/livedata/status and among other things has generator data in it. Possibly they'll expand it with bi-directional data when that becomes available. Doesn't seem to track generator hours, something critical for maintenance. Posted a suggestion for them to add it to the UI.

Unfortunately, it doesn't look like ivp/livedata/status updates very often, examining the data timestamp says it is from ~21 hours ago.
 
Last edited:
Had an interesting experience where the Ensemble system got into a state where it couldn't work. Basically, the Envoy told the batteries to switch to grid, there was no grid and so the Envoy lost power. At that point, even though there was solar and battery available, without 240V for the Envoy to switch them back on, there was no power available. Wrote the whole tragic tale up, hopefully amusingly, in the humor thread here:
Diagnosing a Failure to diagnose a failure
 
Last edited:
GXMnow said:
Yes, an isolation transformer would do the trick, but that is a big cost. I only have a 100 amp main, but that would still be a 24 KVA transformer. The name plate picture is perfect at 25 KVA. 265 pounds, wow. But even more scary, 115C rise!! At 40C ambient, that is 155C winding temp at rated power. How many watts are being turned into heat?
Click to expand...
Worst case, if all the 120V loads are on one phase it would only need to be 12kVA. Half of the loads get power directly from that phase.

The Solaredge auto transfer they supply with their system is 5kVA.

kw
 
I moved two eg4 3000 inverters together, so i could have a 240v circuit. i heard a lot of chatter about balancing loads. at first it appeared that indeed one leg was serving 80% of the load. then i saw a little picture of the inverters with two yellow wires connecting them. i only had one connected. i added the second wire and whalla. no more imbalance.
 
kevin93 said:
Worst case, if all the 120V loads are on one phase it would only need to be 12kVA. Half of the loads get power directly from that phase.

The Solaredge auto transfer they supply with their system is 5kVA.

kw
Click to expand...
That would be true for an autotransformer, not for an isolation transformer.
 
You must log in or register to reply here.
Back
Top