diy solar

diy solar

An Enphase Ensemble Installation

Tried the API out in AP mode (e.g., envoy acting as a wireless access point), but it wants a token too. The beta Enlighten did connect in AP mode and showed the CT values & SoC, but I'm not a fan of putting it in AP mode (it requires manual intervention from a ladder and the internet is most likely to go out in the worse weather (e.g., hurricane). But, better than nothing. Still no joy on a "6 month" token.1644147962358.png
 
Got an email from Enphase they'd closed my microinverter not updating ticket. Woot! My microinverters are all now updated! My Envoy was updated too, possibly for the "in premise networking", the new names are now in Enlighten, and from the date it looks like the Enpower was updated (but it still has the same version number).

IQ Gateway (Envoy): D7.0.71 (a48a57) Software Build Date 07 Jan, 2022 1:38 AM
IQ System Controller (Enpower): 1.5.3637_rel/21.19 Updated Feb 6, 2022 3:39 AM
IQ Battery (Encharge): 1.2.4384_rel/21.19 Updated Feb 3, 2022 11:27 PM
Microinverters:
  • 540-00142-r01-v04.28.03 -
  • 520-00095-r01-v04.28.07 - Tue Sep 14, 2021 09:45 AM EDT
    Profile Set (IEEE 1547:2015-WHB)
 
Last edited:
Since I had a power glitch 2 days ago, I figured I would take a peek in the Enphase logs and see if they saw anything odd. But as I was scrolling back in time to the power glitch, I saw something odd that happened earlier today.

13 of my 16 iQ7's failed to report between 1:19 pm and 1:56 pm. It was a nice day, almost got up to 80F outside, good sun, and my system made a nice 20.8 KWH's, and there is no dip in the production curve, around this time, so I think it was a loss of power line communication.

Most of them failed "set" and then "cleared" just once, but a few panels did it 2 and 3 times. Before the first failure to report, I had the 16 all come out of sleep in the morning, and then after this, I had a few going offline as the shadows hit the lower panels, followed a bit later by the rest of the panels shutting off at sunset.

Also in the morning, one panel through this
Device AGF Profile Propagation Failure: Set
It was then cleared an hour and a half later. What does that mean?

Thanks to my shadows hitting about half my panels, and the wind moving the palm tree, I have a LOT of messages about inverters going DC power too low, and recovering, so it chewed up the 500 log entries. It does not go back far enough to see my power glitch.

My micros are still on software version 520-00082-r01-v04.27.04 Updated 09/03/2021 8:37 AM 5 months ago.
and my profile is still Profile Set (CA Rule21 201902 VV VW FW)

My Envoy software build is D5.0.49 (77afa8) from April 2020.

Since I am just a basic grid tie only setup to Enphase, I guess I don't need the latest firmware??
 
13 of my 16 iQ7's failed to report between 1:19 pm and 1:56 pm. It was a nice day, almost got up to 80F outside, good sun, and my system made a nice 20.8 KWH's, and there is no dip in the production curve, around this time, so I think it was a loss of power line communication.
I've never seen a powerline failure. Sure that's just not the Envoy not reporting for an hour? It should catch up if it was.

Most of them failed "set" and then "cleared" just once
Since it was most of them, it sounds like the grid was outside the profile spec. Does your Schneider give decent grid readings? I would dearly love something akin to a frequency trace at times of trouble. I'm never sure if it's something from inside the house, the grid, or if one of the microinverters gets noisy when it's overloaded causing others to trip. I'll start a thread on that, possibly someone knows of an inexpensive tool.

Also in the morning, one panel through this Device AGF Profile Propagation Failure: Set ... What does that mean?
I see that every so often too, not really sure what it means. AGF probably stands for “aggregated generating facility”, so it sort of sounds like the profile change didn't succeed on all the micros. I checked my logs but didn't see any AGF messages, was hoping some event from the same device/time gave more of a clue.

My Envoy software build is D5.0.49 (77afa8) from April 2020.Since I am just a basic grid tie only setup to Enphase, I guess I don't need the latest firmware??
Sounds like it's hung up. I think most of the fixes have been around the Ensemble tech, but D5.x seems a long way from D7.x. As I recollect, you got the important fix to reset the microinverters when they got hung.

Probably wouldn't hurt to have the latest. You might want to wait until they get past that whole API token-issue for no-internet access before calling them. I'm guessing the installation gets a fault, so probably trying to DIY it through the toolkit won't work. I just got past that issue with those 17 microinverters, took four calls (mainly because the "update" takes a few days and they'll want to try it a few times before kicking it to engineering).
 
Last edited:
There's a lawsuit over Enphase's new remote token to get access to the local API access: https://enphaselawsuit.com/
They're cognizant of the HA work-around and the page even lists code to get the token, the suit is entirely about the real issues, e.g.:

With a natural disaster, oftentimes the internet is out for days or weeks as the infrastructure is without power and lines are down. This mean you would be blind to the operating of your system during the entire outage rendering it nearly useless for emergency preparedness.
 
Was able to get a 6 month token using this new tech brief: https://store-d9.enphase.com/downlo...=answer_links&utm_campaign=internal_reference

Login into Enlighten with Enlighten credentials
Open the Enlighten access token URL: https://enlighten.enphaseenergy.com/entrez-auth-token?serial_num=<envoy serial number>
It came back with this: "expires_at":1660526973, which converts to "Monday, August 15, 2022 1:29:33 AM"
Also decoded the token, got, exp":1660526973, so Woot! Guess I can code that up tomorrow!
 
Last edited:
Fairly straightforward... use post #361 for the rest of the code...

Java:
// This request logs you into enlighten
        String urlParams = "user[email]="+ enlightenLoginId + "&user[password]="+password+"&commit='Sign In'";
        HttpRequest request = HttpRequest.newBuilder()
                .uri(URI.create("https://enlighten.enphaseenergy.com/login/login"))
                .headers("Content-Type", "application/x-www-form-urlencoded")
                .POST(BodyPublishers.ofString(urlParams))
                .build();

// this request gets you the 6 month token in the response body, it's "token":
        request = HttpRequest.newBuilder()
                .uri(URI.create("https://enlighten.enphaseenergy.com/entrez-auth-token?serial_num="+envoySerialNumber))
                .GET()
                .build();

// this request authenticates the Envoy session using the 6 month token:
        request = HttpRequest.newBuilder()
                .uri(URI.create("https://"+envoyIP+"/auth/check_jwt"))
                .headers("Authorization", "Bearer "+token)
                .build();
After that, you can perform the normal get requests via https. Nothing is left now but writing up the bulk of the code (e.g., replacing prior 6 month tokens, testing if the internet is up, deciding to use the 6-month token or not).

I saw some folks on the Enphase forum saying some api commands didn't work, but they worked for me so I suspect they didn't switch to https or lost the session id/cookies.
 
Last edited:

WiFi & Zigbee Interference​

Looking at my ZigBee signal strength on the IQ Battery and wondering if I can do anything to improve it. The Envoy, Enpower, and IQ batteries are all fairly close, yet I get 5 bars on the Enpower (right next to the IQ Combiner with the dongles) and this for two IQ batteries: 1645634597277.png.

Could switching WiFi channels help?

Both protocols operate in the same frequency band. WiFi starts with channel 1 with a
mid-frequency of 2412 MHz, and there’s another channel every 5 MHz, so channel 2 is
2417 MHz and channel 14 is 2484. The width is 20 Mhz. WiFi has 3 non-overlapping
bands and Zigbee has 16 channels that are 2 MHz wide and 5 MHz apart.

So, if your neighbor is on channel 1, you might set your WiFi channel to something
greater than channel 5 so the signals don’t interfere. The same thing with Zigbee, you
want to make sure the frequencies don’t overlap.
1645634658984.png

So, from putzing around with my CC2531 I have two strong Zigbee channels at 15 and 25 and one of them must be the IQ Batteries (possibly both as I have two dongles). Can’t set the Zigbee channels, but I can set the WiFi. From this ref, channel 15 is 2425 MHz (between wifi channel 3 & 4) and channel 25 is 2475 MHz (not used by WiFi in North America, ref).

Checking, my WiFi is on channel 9, which is 4+5, so altering it probably won't have any effect. My neighbors are unfortunately channels 1, 2, 5 & 6, so not much I can do about that. They’re in the -80 dB range. So doesn’t look like this will do anything for me but it was worth checking out.
 
Last edited:
Switching channels, moving antenna to change distance (standing waves), rotating or tilting antenna (radiation pattern) are possibilities.
Different models could have different sensitivity.

Shielding. Put up foil or metal screen (small mesh relative to wavelength, so window screen) between your system and his. A bit will bleed around edges, but should greatly reduce signal strength.

Wave guide. Reflective metal tube, > 1/2 wavelength in width, spanning from one antenna to the next. Absorber material at the end, or just open to the air and not aimed at neighbor.

Probably very low data rate. Too bad they aren't apparently CDMA, rather something like TDMA, so they collide.

Antennas with coaxial connectors? Run through attenuators (e.g. -60 dB, i.e. 1000:1 resistor divider) and join together. "Cable TV" rather than "Broadcast"
No accessible antenna? Run cables between units with E-field (open end whip 1/4 wavelength) or B-field (1/2 wavelength loop to ground) at each device. Whether "E" or "B" works better for close coupling depends on which is used inside. The infamous first iPhone had an E-field antenna under user's hand, and water absorbs electric field.
 
The 2.4 GHz spectrum is a complete mess. So many devices are trying to use the same space, it is a wonder it works at all. I have hard network cable running just about everywhere I can. I need to add another switch so I can eliminate the WiFi hope to my garage. Beyond that, the only WiFi devices are phones, my work laptop, two Fire sticks, and my son's PC. Wow crap, I do have a lot of data on WiFi still.

Any device that can do 5GHz is working up there. It's beter on 5gig, but it has a shorter range. And you can't exactly tell your neighbors to switch all their stuff to 5gig to free up ZigBee space.
 
Switching channels, moving antenna to change distance (standing waves), rotating or tilting antenna (radiation pattern) are possibilities.
Did try a parabolic reflector in #366 (didn't do anything, but there wasn't much room in the IQ Combiner) and reorientating (got me to two bars). Since then haven't had issues where the batteries bottom-out, but would still prefer a strong signal. I haven't tried shielding from the neighbors, but a part of that is I'm pretty sure the Enpower is what's blocking the signal and I'm not sure how the signal is reflecting in.

On the Enphase diagrams the sequence is always Enpower, Envoy (In the IQ Combiner), batteries... but here the physical sequence is Envoy, Enpower, Batteries. Here's what it looks like, you can see the battery isn't far from the Envoy:

1598017157586-png.20502
And yes, that's stucco over concrete, so the wire mesh behind it is doing god-knows-what.

Antennas with coaxial connectors?
That would be nice. But no, the dongles supplied are the USB dongles without connectors. My current thought is to get a USB extender and snake it from the combiner box into the Enpower enclosure. I have one ... should probably try it just to see what happens.
 
Is the porch area open on two walls, or just one?
Each pair of facing walls (also floor/ceiling) can support standing waves.
Chicken wire should be somewhat porous to 2.5 GHz but also something of a ground plane.

Is your neighbor behind the stucco and chicken wire? Or the open area? If open, add a copper window screen. Put it at an angle so not parallel to stucco, e.g. tilted out like an awning, maybe only 1' out at bottom of opening.
If behind stucco, try foil or window screen between him and the units. Inside for first test, could be relocated between wall and units if successful.

Maybe just window screen over the units, forming a wave guide with the chicken wire. Aim open ends away from offending sources. Chamfered corners would be good, but looks like Encharge is against the other wall.

You might be able to close-couple an antenna to each device and run cabling between them.
Free space has 377 ohms impedance, while close up coupling could behave different. Using a resistor divider to drop 377 ohms to 50 ohms could help couple into cable. maybe further attenuation needed, maybe not. A shield around outside should reduce pickup of neighbor's transmissions, and absorptive inside reduces standing waves (shield rooms for EMI test are anechoic with carbon loaded foam cones. Your need is narrow band high frequency.)

Tests of telemetry systems prior to flight that I observed had a coupler plate placed next to antenna and connected to test equipment.
 
Is the porch area open on two walls, or just one?
Open area, everything is hung on the western wall. Although you can see the concrete beam supporting the deck above, so there is some interference from that undoubtedly.

Is your neighbor behind the stucco and chicken wire?
He has vinyl siding and is line-of-sight. His signal is ~-80 db though so I'm dubious.

... forming a wave guide with the chicken wire....close-couple an antenna to each device and run cabling between them.
I like these ideas!

... current thought is to get a USB extender and snake it from the combiner box into the Enpower enclosure...

Putting the dongles in the Enpower doesn't work, turns out it has an RF shield (you can see the metal attached to the inside cover in the photo below). Some signal get's out, but it's just a little better than in the IQ Combiner.

But, using a USB extender and draping the cord over the Enpower, then letting the dongles dangle as shown in this photo (can't really see it, but the dongles are plugged into the Envoy's USB port near the top of the IQ Combiner) does seem to work...
1645641175812.png

... signal went up two bars: 1645638987380.png.

Possibly put a waterproof connector on the IQ combiner box and string the dongles to a waterproof box mounted above the window? Below the IQ 10 might be better, but sure how much signal that chunk of concrete is absorbing. Easy to test....
 
Last edited:
Picked up this project box to put the dongles in (see #395) and thought before drilling the holes I should check to see if I have any signal losses as
I'd read that the anti-UV agent can impact the signal strength.

So, hook it up... and the signal information is not getting updated. At all. Got a call and half an hour I look at the signal and it's all good... What?

Turns out the software was getting updated... curiously... it was to the same version? Here are the previous and current version numbers:
IQ Gateway (Envoy): D7.0.71 (a48a57) Software Build Date 07 Jan, 2022 1:38 AM
Microinverters: 540-00142-r01-v04.28.03 & 520-00095-r01-v04.28.07 - Tue Sep 14, 2021 09:45 AM EDT
IQ Battery (Encharge): 1.2.4384_rel/21.19 Updated Feb 3, 2022 11:27 PM
IQ System Controller (Enpower): 1.5.3637_rel/21.19 Updated Feb 6, 2022 3:39 AM
Profile Set (IEEE 1547:2015-WHB)

IQ Gateway (Envoy): D7.0.71 (a48a57) Software Build Date 07 Jan, 2022 1:38 AM
Microinverters: 520-00095-r01-v04.28.07 Updated February 5th, 2022 9:26 AM
IQ Battery (Encharge): 1.2.4384_rel/21.19 Updated March 1 2022 9:58 AM
IQ System Controller (Enpower): 1.5.3637_rel/21.19 Updated March 1, 2022 9:58 AM
Profile Set: IEEE 1547:2015-WHB
 
Last edited:
It looks like they swapped the names??
IQ Battery went from a 1.5.3673 to a 1.2.4384
IQ system controller went from 1.2.4384 to 1.5.3673
Microinverters went up so they are all now 4.28.07

I thought Enpower was the transfer switch, and Encharge was the batteries. Have they been changing things around.
 
Lol!.. the names did changed. He just had them mistyped!
I was, ahhhh, just testing to see who was paying attention. ; -) Good catch @GXMnow and nice observation @el-guapin!
But yeah, in #384 the new labels I copied from were already reversed (e.g., IQ Battery is the new name for Encharge) and didn't catch it either time. Then compounded the error by swapping the version numbers in the previous. Guess it's not just calculators I can't be trusted with. ; -)
I edited/corrected the two prior posts so as to less-confuse newcomers.

Let's try that again...
IQ Gateway (Envoy): D7.0.71 (a48a57) Software Build Date 07 Jan, 2022 1:38 AM
Microinverters: 520-00095-r01-v04.28.07 & 540-00142-r01-v04.28.03 Updated February 5th, 2022 9:26 AM
IQ Battery (Encharge): 1.2.4384_rel/21.19 Updated March 1 2022 10:35 AM
IQ System Controller (Enpower): 1.5.3637_rel/21.19 Updated March 1, 2022 10:35 AM
Profile Set: IEEE 1547:2015-WHB

This is curious... the timestamp changed again?

Previously it was 9:58 AM. In the snapshot to the right, it's 10:35 AM.

I was putting everything back around then as I can't do the actual installation until the
5mm glands arrive (needed for the IQ Combiner and new box).

Either the timestamp reset to the current time when it was reconnected or after being
disconnected for a while it feels the need to reinstall. Curious.

On a side note, the WHB in the profile might be Whole House Backup &
unpdated the OP with February's numbers.
1646220635081.png
 
Last edited:
Back
Top