An Enphase Ensemble Installation

[svetz]

Batteries freaked out again last night, two started discharging, one started charging, and the other behaved normally. Called Enphase and they did a soft reboot. The system was already recovering on its own, I mainly just called them to let them know D7.3.121 still had issues. Here's the data, the columns are timestamp, battery ID [battery SoC, Battery LED state]

Timestamp                  ID1 Soc           ID2 Soc          ID3  Soc          ID4  Soc
12/09/2022 18:00:01, SoCs= 94 [60%  LED=15], 30 [60%  LED=15], 88 [59%  LED=15], 89 [60%  LED=15]
<no night time data collection, pc off>

12/10/2022 06:00:01, SoCs= 94 [34%  LED=16], 30 [32%  LED=16], 88 [87%  LED=14], 89 [56%  LED=12]
12/10/2022 06:43:10  SoCs= 94 [33%  LED=16], 30 [32%  LED=16], 88 [87%  LED=14], 89 [58%  LED=12]
12/10/2022 07:00:01, SoCs= 94 [33%  LED=16], 30 [32%  LED=16], 88 [86%  LED=14], 89 [59%  LED=12]
12/10/2022 07:43:10, SoCs= 94 [34%  LED=12], 30 [33%  LED=12], 88 [87%  LED=12], 89 [60%  LED=12]
12/10/2022 08:00:01, SoCs= 94 [35%  LED=12], 30 [33%  LED=12], 88 [88%  LED=12], 89 [61%  LED=12]
12/10/2022 08:43:10, SoCs= 94 [38%  LED=12], 30 [37%  LED=12], 88 [91%  LED=12], 89 [64%  LED=12]
12/10/2022 09:00:01, SoCs= 94 [40%  LED=12], 30 [38%  LED=12], 88 [92%  LED=12], 89 [65%  LED=12]
12/10/2022 09:43:10, SoCs= 94 [43%  LED=16], 30 [42%  LED=16], 88 [91%  LED=13], 89 [64%  LED=13]
<reboot>
12/10/2022 10:21:52,  SoCs= 94 [46%  LED=16], 30 [44%  LED=16], 88 [89%  LED=13], 89 [61%  LED=13]

There doesn't seem to be any actual discharge, so probably one SoC was off and one battery send current into another?? <sigh>

 

[GXMnow]

That looks like very odd behavior. I have essentially 3 separate battery banks, but in my case, they all feed DC to/from a single inverter/charger, so it will never be able to act like that. The way the Enphase batteries work, each 3.3 KWH of battery has 4 inverter/charger units. I like the idea of the redundancy, and having the separate inverters on all my solar panels is great. But in the battery case, it looks like they ae not coordinating the charge and discharge across all of the battery inverters. I have to think about it though.... Are they doing it for a reason?

This is all speculation on my part, just trying to think why they would let it do that. One thing I have learned from reading the forums and watching a lot of You Tube videos is that tracking the state of charge on LFP cells is a pain in the butt. That super flat voltage discharge curve that works so great for loads makes determining the state of charge almost impossible. You can only find full charge by charging into the top knee. And you need to discharge into the bottom knee to find full discharge. With my big single inverter/charger, that would mean that to calibrate the State of Charge, I have to run my whole system down to the lower knee and then charge it back up to the top knee, or the other way, to find the true state and capacity of the battery bank. In theory, it could be done with PV solar providing the charge power, and the house loads used to discharge it, so the energy is not just wasted, but if there was an actual grid outage or poor solar production during a test like that, it could be a huge problem. With all of the separate inverter/charger sections in the Enphase system, and having separate battery banks. it does give them an interesting option. They could be having it test the state of charge and capacity of individual batteries. With enough separate batteries, they can test two at a time. Using the energy from one battery to charge up another. The rest of the batteries are still working the house normally. The one battery reaches the top knee, it now knows 100% on that one. The discharging battery hits the lower knee, it now knows 0% on that one. Then reverse them, have the full one discharge into the dead one, and go until they hit the opposite knees again. Now both batteries know their full capacity and the 100% and 0% states. Depending on the total configuration of battery/inverter units, maybe it tests 1 or 2 batteries every once in a while? After the test, they normalize the state of charge to match the other batteries in the group.

I could be completely wrong, but something like this might explain what they are doing. Enphase is a very smart company. Coding the system to keep the batteries in sync does not seem too difficult. So when they go way out of sync like that, I have to believe they have a reason for doing it.

To be able to safely do something like this to self test the batteries, you would need at least 3 battery banks with separate inverter/charger units. The Enphase battery system is one of the few that actually does have that.

 

[svetz]

It's doing it again, this time the two that went low stayed stable and the one that went high went low. #89 continues to behave normally.
There's no sign of any current moving, so I suspect it must be a truing up of the SoC between the systems. If so, #89 should trip in a few days.

Previously:

Timestamp                  ID1 Soc           ID2 Soc          ID3  Soc          ID4  Soc
12/10/2022 06:00:01, SoCs= 94 [34%  LED=16], 30 [32%  LED=16], 88 [87%  LED=14], 89 [56%  LED=12]...

Currently:

Timestamp                  ID1 Soc            ID2 Soc      ID3 Soc         ID4  Soc
12/12/2022 06:00:01, SoCs= 94 59%  LED=15, 30 59%  LED=15, 88 33%  LED=12, 89 59%  LED=12...

Updates
12/13 : #30 is 49% this morning, the rest are normal.
12/15: #94 is 41% this morning, the rest are normal.

12/16: #30 dropped to 33% at 5:17 pm.
12/18: #94 dropped to 34% at 6pm

 

[Thebravo]

Hi Svetz,

I will be starting my own thread (so as not to clutter up yours), but I'm just a few hours north of you (Port St Lucie) and I just had a similar Enphase system installed (IQ7+ inverters, IQ controller, 13KWH of batteries and 11.5 KW of Qcells panels). I'm just waiting on FPL for the meter swap and the permitting office to do the final inspection then I should be live and producing. I am very much into the deeper workings and gaining full control over the system as it appears you (and the others on this thread) are as well. Thank you for the very detailed thread detailing how to get into the data and deeper workings of the system, Looking forward to trying t some of your finding on my own system. I want to be able to control my system, battery SOC and such so it works the way I want it to. My first step once the inspection is done is rewiring the loads panel as they didn't give me nearly enough circuits on backup

-thebravo

(BTW I was down in Key Largo last year for some snorkeling, it's absolutely beautiful down there!)

 

[svetz]

Hi Svetz,

I will be starting my own thread (so as not to clutter up yours),

Sweet! I'll add you to my watch list! Why IQ7s and not 8s? Your battery & Enpower are a couple gens up from mine, it'll be interesting to see if you see the same issues.

My first step once the inspection is done is rewiring the loads panel as they didn't give me nearly enough circuits on backup

The Enpower is a 200 amp passthrough, I got the whole house backed up (down side is I have to turn off breakers during an outage (until I automate anyway), upside is 100% control of where the power goes.

(BTW I was down in Key Largo last year for some snorkeling, it's absolutely beautiful down there!)

Sure is! Love the water, but could do with a bit less humidity/heat in June through August or a more effective cooling vest.

 

[Thebravo]

Sweet! I'll add you to my watch list! Why IQ7s and not 8s? Your battery & Enpower are a couple gens up from mine, it'll be interesting to see if you see the same issues.


The Enpower is a 200 amp passthrough, I got the whole house backed up (down side is I have to turn off breakers during an outage (until I automate anyway), upside is 100% control of where the power goes.


Sure is! Love the water, but could do with a bit less humidity/heat in June through August or a more effective cooling vest.

At the time, when it was quoted the IQ8's were quoted as notably more expensive than the 7's, that said the actual price difference on the website doesn't seem that great at this point... other than slight efficiency boost and the ability to restart the system if the batteries are fully depleted there didn't seem to be a lot to justify the (what was quoted to be) a much higher cost at the time... We had to replace the roof as part of this as well, so the loan for this project is nearly 100K.

They installed my system with just a 100 amp backup loads breaker in the Empower and gave me an a total of 8 circuits in the critical load panel (2 of which are the grinder for the wastewater) I could easily swap 6 of those breakers to the "cheater" double breakers and get 12 circuits which would actually be most of the non heavy duty load circuits in the house. Alternately I could swap the 100 amp Empower breaker to a 200 amp and hook it to my main panel and get rid of the sub panel, I'd have to see if the system can realistically run those heavier loads (at least during peak sun conditions). I pushed them to give me more circuits up front and they were not willing to do so... and I get it, for the general population folks won't be careful about monitoring their power draw and having the system tripping off on overload a bunch when the customer turns on too many items would be a bad customer experience... But the cost to add more batteries to get "full home" backup was financially not doable for us right now. Even if I don't put the large loads on, I have decent lower power backup options to *most* of my heavy load items, the exception being the hot water heater which would be nice to have. I will have to do some calculations and see if doing it as a "whole home" backup manually toggling off breakers makes more sense it might, but not having actually seen what I can actually generate for power yet I'd like to see what I realistically get for numbers. Mechanical engineer here, so while I can handle this type of electrical work, it is still a little foreign at times

I'm originally from new England so heat and humidity are not my friend at all and I spend most of the non winter months here being sweaty whenever I step outside lol.

 

[svetz]

Interesting sequence of events starting at 1:40 AM, there's no report of being off-grid or erroneous discharge... in Enlighten but the battery reports to the Enphase mothership 0% . Here are the events from the log with dups removed. Looks like a fairly normal switch to the battery, but there are no events as to why it switched to battery. My guess is the local power got crappy but the grid didn't actually go down.

     01:46  052 "SOC below reserved level : Clear", "envoy",
     01:46  988 "EnchargeOnGrid", "Encharge Controller",
     01:46  762 "translation missing: en.env_enl_api.event_type.enpower.generator_agf_profile_prop_done", "Enpower Device",
     01:46  762 "translation missing: en.env_enl_api.event_type.enpower.mmongrid_profile_prop_done", "Enpower Device",
     01:46  762 "EnpowerAgfProfilePropagationDone", "Enpower Device",
     01:45  762 "EnpowerControllerStartClosingEC1Relay", "Enpower Device",
     01:45  762 "EnpowerGridIsGood", "Enpower Device",
     01:46  237 "EnchargePCUStateThrottled", "Encharge Micro",
     01:46  146 "Multi Mode On-Grid", "Encharge Micro",
     01:45  194 "EnchargeGridModeFromEnpowerMMOnGrid", "Encharge Controller",
     01:42  052 "SOC low between 5 to 10 percent: Clear", "envoy",
     01:42  230 "translation missing: en.env_enl_api.event_type.encharge.pcus_not_sensing_ac_voltage", "Encharge Controller",
     01:41  194 "EnchargeGridTied", "Encharge Controller",
     01:41  989 "translation missing: en.env_enl_api.event_type.no_pcu", "Encharge Controller",
     01:41  052 "SOC low between 5 to 10 percent: Set", "envoy",
     01:41  052 "SOC below reserved level : Set", "envoy",
     01:41  237 "EnchargePCUStateIdle", "Encharge Micro",
     01:41  237 "Grid-Tied", "Encharge Micro",
     01:41  237 "AC Frequency Out Of Range", "Encharge Micro",
     01:41  237 "AC Voltage Out Of Range", "Encharge Micro",
     01:41  237 "Grid Instability ", "Encharge Micro",
     01:41  762 "EnpowerZigbeeCommFailure : Clear", "Enpower Device",
     01:40  762 "EnpowerZigbeeCommFailure : Set", "Enpower Device",
     01:41  230 "EnchargeGridModeFromEnpowerGridTied", "Encharge Controller",
     01:40  762 "EnpowerGridOperatingFull", "Enpower Device",
     01:40  762 "EnpowerControllerStartClosingMIDRelay", "Enpower Device",
     01:40  762 "EnpowerControllerNoNcOneRlyGridAct", "Enpower Device",
     01:40  762 "EnpowerControllerNC1RelayClosed", "Enpower Device",
     01:40  762 "EnpowerSwitchedToPrivateKey", "Enpower Device",
     01:40  762 "EnpowerSWReset", "Enpower Device",
     01:40  762 "EnpowerControllerStartOpeningEc2", "Enpower Device",
     01:40  762 "EnpowerControllerStartOpeningEC1Relay", "Enpower Device",
     01:40  762 "EnpowerControllerStartOpeningNFTRelay", "Enpower Device",
     01:40  988 "EnchargeIslanding", "Encharge Controller",
     01:40  494 "Islanding", "Encharge Micro",

 

[svetz]

#94 dropped to 33% again (see #503), so I've set it for full backup (a top balance won't hurt ; -). No corresponding sudden energy discharge, so it's just a SoC reporting issue.

 

[goldserve]

To clarify, is it no longer possible to get installer access to change settings like adding a consumption meter on v7 firmware? I was able to access installer mode on v5 with the installer password generator.

 

[el-guapin]

#94 dropped to 33% again (see #503), so I've set it for full backup (a top balance won't hurt ; -). No corresponding sudden energy discharge, so it's just a SoC reporting issue.

I have seen this same behavior with my system as well. Does not happen often but when it does, there nothing in event logs to indicate there is an issue. The high-RES data shows everything normal. no current drops, no frequency shifts, nothing. On the next reporting cycle, the values return to normal again.

 

[svetz]

To clarify, is it no longer possible to get installer access to change settings like adding a consumption meter on v7 firmware? I was able to access installer mode on v5 with the installer password generator.

Yes and no. You're correct that the password from the generator won't work anymore as a local login...but, you can still login as an installer if authorized (which you can become by taking the university courses).

 

[svetz]

... On the next reporting cycle, the values return to normal again...

That looks like a communication loss. I get those too.

The issue I'm talking about is where the SoC on one or more batteries suddenly goes from 60% to 30% and stays there despite good communication. It causes the other batteries to try and charge to compensate. Here's an image of it, as you can see there's no energy transfer out of the battery, it just decided to change it's SoC.

​

For example, in the table below all the batteries started at 60% and they all have decent signal strength. But then around 6ish two suddenly dropped and now 988 is overcharging to compensate for 989 and 194 misreporting. As the two lower batteries start to charge they'll suddenly realize at some point their true SoC and then 988 will discharge back into the house. There's really only inefficiency losses and additional un-needed cycles on the battery, so it's a minor peeve.

 

[goldserve]

Yes and no. You're correct that the password from the generator won't work anymore as a local login...but, you can still login as an installer if authorized (which you can become by taking the university courses).

Oh cool. So it doesn’t have to be the original installer of the envoy. Just any enlighten account that is an installer?

 

[GXMnow]

How often are you seeing those state of charge drops and overcharge events? Does there seem to be any pattern to it, or is it just random?

My setup, using the Schneider XW-Pro for the batteries, is not very elegant, but now my PLC has full control over the battery charging and discharging. When I go to LiFePO4 batteries, I am curious if I might need to do something like that to better track the true state of charge. Does it need to bring each battery up to a full charge once in a while to reset the SoC counter and also top balance the cells?

Enphase tech support is being a pain in the neck with me on my microinverter communication issue, and I am getting more than a little annoyed with them now. The new version 7 is not user friendly. They had remotely enabled the consumption metering, but I don't have the CT's installed. That changed the monitoring page and having zero consumption all the time was making for an odd chart. It took a week, but I finally got them to put that back. But they are still claiming the new firmware fixed the microinverter communication issue. I am not convinced.

 

[svetz]

How often are you seeing those state of charge drops and overcharge events? Does there seem to be any pattern to it, or is it just random?

Seems to occur at the end of the day (6ish) after a few days of left being at 60% SoC. It doesn't happen if the batteries are cycling, only when held steady.

... they are still claiming the new firmware fixed the microinverter communication issue. I am not convinced.

Have you been getting good signaling throughout the day?

 

[GXMnow]

Have you been getting good signaling throughout the day?

Without being able to see the logs, I don't know if I am getting drop outs. I have not been able to catch one live yet. I don't stare at my PC, and the Enlighten app is very good at hiding if it lost data for a couple hours. As long as it get's data some time near the end of the day, it just fills in the graph and without the log, I would not know if there was a loss. http://envoy.local/home#overview is still showing solid 4 bars on signal, and I can see the individual inverter power with http://envoy.local/api/v1/production/inverters

I still have not been able to sign up for the online classes. For whatever reason, the Enphase site says my Enlighten account is for a country or language other than USA/English. I have not found a way to fix that yet. And since I can't guess what it is set for, it will not allow me to sign in at the Enphase University. Once I get logged in, I will start taking the classes.

 

[goldserve]

Without being able to see the logs, I don't know if I am getting drop outs. I have not been able to catch one live yet. I don't stare at my PC, and the Enlighten app is very good at hiding if it lost data for a couple hours. As long as it get's data some time near the end of the day, it just fills in the graph and without the log, I would not know if there was a loss. http://envoy.local/home#overview is still showing solid 4 bars on signal, and I can see the individual inverter power with http://envoy.local/api/v1/production/inverters

I still have not been able to sign up for the online classes. For whatever reason, the Enphase site says my Enlighten account is for a country or language other than USA/English. I have not found a way to fix that yet. And since I can't guess what it is set for, it will not allow me to sign in at the Enphase University. Once I get logged in, I will start taking the classes.

Sign up with a new email and when they ask, choose self installer. You will get installer access and take the courses. Fastest I was able to finish all the Pv design and install in 2.5hours

 

[GXMnow]

I thought the e-mail had to be associated with a system? It says it must be registered in Enlighten.

I just replied to tech support. They asked me if Enlighten was now displaying properly without the CTs. I have the previous day and year graph overlays back now. At the end of the reply, I asked how to fix my account not being USA / English.

 

[goldserve]

Oh cool. So it doesn’t have to be the original installer of the envoy. Just any enlighten account that is an installer?

Looks like this might be bad info. For a commissioned system, you have to be the original installer to access the equipment. The token page allows you to choose commissioned or uncommissioned system and when I select commissioned, I don’t see the gateway in the list.

 

[el-guapin]

Looks like this might be bad info. For a commissioned system, you have to be the original installer to access the equipment. The token page allows you to choose commissioned or uncommissioned system and when I select commissioned, I don’t see the gateway in the list.

FWIW, I am not the original installer of my system. I wasn’t even an installer!. But after becoming certified through their online university program, I gain full access to all equipment. I can get replacement parts sent to me directly instead of those being sent to the CO that installed my system then have then sent a tech. I can commission and decommission every and all components of my system without affecting warranty.