Adding storage to my Enphase system

[GXMnow]

I will look it up again. I am out of town for a few days. It was listed as "Maximum Grid Sell current".

Found it
0x01B4 Maximum Sell Amps

It does that amount of current back to the input, plus any current on the output, and when my AC coupled solar is producing, it sells less from battery so the total selling back through the XW is this number. It will also go to less, if the current on the output side is high enough that you either hit the 6,800 watt limit, or your battery current limit.

 

[400bird]

Thanks, that's what I was using.
I'm having difficulty with that specific address. It must be something on my end, I'll try it again.

 

[Hedges]

Hex syntax? Maybe you mixed up "O" with "0"?
Probably zero-x-zero-1-B-4
The leading zero says to parse it as a number, and the x says it is hex.

 

[400bird]

Thanks Hedges.
I'm using node red, it takes addresses in decimal not hex.
So, I'm looking at address 436 (from memory, maybe 437).

Plus there is the counting issues where some start addresses at 0 and some start at 1.
I've tried one address in either direction.

The data is 32 bits, which is two addresses long.
I can poll data successfully from the address.

My issue is writing data to the address. I think I need to involve another function to format the data as a list [0,data] as this is how the retrieved data is presented by node red.
I've got more research to do on why it isn't working

 

[GXMnow]

I was looking up some info before I do my firmware update and I came across his video.

This is one of the reasons I ended up with the Schneider XW-Pro. This inverter is an absolute beast. No matter what load I throw at it, it just plain works. If all goes well, I will be up to current firmware tomorrow, and I am working on a few changes to my PLC control as well. In this crazy hot weather, I am gong to have it charge a little harder while the sun is up and the A/C is running. I will still be sure to be charging at less power than my solar array is making, but I will be buying a little more of the cheap grid power to run the A/C while the sun charges the batteries to then run the A/C from 4 pm to 9 pm. With this heat wave, I am just not making enough solar to run the A/C and get enough charge. It is after midnight, and it is still over 81 degrees out.

 

[Hedges]

The data is 32 bits, which is two addresses long.
I can poll data successfully from the address.

If read was successful, I guess you already got "Big Endian/Little Endian" correct.

Since this is a bus not a memory, there may be something more to the protocol to write.
The GUI I used for SMA's RS-485 interface has me enter a number (or pulldown option), also "save" or "write". Although, that may be GUI software related, not a function on the device. I also have to enter a code to be allowed to write. But that also may be GUI (or monitoring gizmo) related, not implemented in the device.

 

[400bird]

If read was successful, I guess you already got "Big Endian/Little Endian" correct.

Since this is a bus not a memory, there may be something more to the protocol to write.
The GUI I used for SMA's RS-485 interface has me enter a number (or pulldown option), also "save" or "write". Although, that may be GUI software related, not a function on the device. I also have to enter a code to be allowed to write. But that also may be GUI (or monitoring gizmo) related, not implemented in the device.

It turns out it was node red trying to write two int16 addresses, not a single unit32 address. I was able to get it straightened out. Thanks for the help. Sometimes it just takes a little kick to figure it out. You're mentioning hexadecimal numbering reminding me it might be an issue with the data formatting.

I was looking up some info before I do my firmware update and I came across his video.

Careful with the 1.11 software update. Myself and a few others have had issues with that build.

This is one of the reasons I ended up with the Schneider XW-Pro. This inverter is an absolute beast. No matter what load I throw at it, it just plain works. If all goes well, I will be up to current firmware tomorrow, and I am working on a few changes to my PLC control as well. In this crazy hot weather, I am gong to have it charge a little harder while the sun is up and the A/C is running. I will still be sure to be charging at less power than my solar array is making, but I will be buying a little more of the cheap grid power to run the A/C while the sun charges the batteries to then run the A/C from 4 pm to 9 pm. With this heat wave, I am just not making enough solar to run the A/C and get enough charge. It is after midnight, and it is still over 81 degrees out.

Watching, if the Sol-Ark fanbois see your post, they'll come running to defend!

In all seriousness, ambient temp was 115 in my backyard. The inverter is in my uninsulated garage, so the temp was likely similar inside, and the inverter ran at 6kw for 3 hours straight, no problem at all. I should have ventured out into the heat to check the fan speed and air temp exiting the top of the inverter, but it sounded like an awful idea yesterday in the disgusting evening heat.

 

[Hedges]

I was looking up some info before I do my firmware update and I came across his video.

This is one of the reasons I ended up with the Schneider XW-Pro. This inverter is an absolute beast.

Husky compressor drew 1.5 kW steady-state. If it has induction motor, 7.5kW surge; if brush type, less. It is fed by 120V, so if Schneider has transformer able to put power from both its legs on a single phase, that is an advantage.

(Midnight's demo started four DeWalt compressors simultaneously, but I found brushes as replacement parts for those.)

People say he has produced paid hit-pieces about SolArk, on behalf of his master. Sorry, NOT this guy. I mixed him up with another. Intentionally unbalancing L1 vs. L2, not updating firmware, not speaking to support.

Although he mentioned heat pumps, background loads were only 1.7kW.

He said SolArk quits if it approaches spec limits, while Schneider was running about 50% over for a while. We hope Schneider does adequately protect itself, and SolArk may deliberately have more conservative limits until they've gained more confidence.

He fully loaded (and overloaded) Schneider with benign loads, electric dryer. I'd want to see central A/C and well pump being started.
It would be fair to assume he experimented first, and if he found a way to overload it, selected a demo to present it in the most favorable light.

But it is good if Schneider handles heavy loads and unbalanced loads better.

Would rather see comparisons from someone without an axe to grind, mapping out the operating space of each product.

 

[Hedges]

In the video he mentioned installing smoke detectors for the battery. Wireless. Inside a Faraday cage?
Although, door seal is plastic, does provide a slot for radio waves to pass through (polarizer).
Hope he tests signal strength/communication.

 

[EnderWiggin]

People say he has produced paid hit-pieces about SolArk, on behalf of his master. Intentionally unbalancing L1 vs. L2, not updating firmware, not speaking to support.

Not sure what "hit pieces" you're talking about. He originally caught my interest because he posted a video specifically saying Schneider was garbage and how he was replacing it because it was failing in a pure backup scenario. He then was contacted by Schneider to try their product again and began this more recent series.

He has one or two videos where he tries to break the Sol-Ark at his home and succeeds rather easily, mostly through unbalanced loading, but I never saw anything there that specifically led me to believe he was creating "non-normal" situations that wouldn't be seen in the real world. Just watched one where he tests AC coupled solar with SolarEdge and it worked flawlessly with a Sol-Ark.

Perhaps you're confusing him with DavidPoz, who made some videos that were directly responded to by Sol-Ark because he loaded his legs intentionally unbalanced, and how he went back to recommending using it with an autotransformer?

Would rather see comparisons from someone without an axe to grind, mapping out the operating space of each product.

Again, I don't see evidence of this from Ben. The closest I've seen in this area would be Abney's video, and people yell on Youtube about how his test was jaded, as well...

 

[Hedges]

Thanks, Ender, I did confuse him. Updated my post.

Each inverter has its own 120V and 240V wattage limits.
Having a maximum allowed imbalance for SolArk seems odd, in the absence of an auto-transformer.

What are Schneider's capabilities, for max wattage on 240V, and max wattage on a single 120V phase?

 

[Hedges]

... it was failing in a pure backup scenario. He then was contacted by Schneider to try their product again and began this more recent series.

What were the conditions that failed initially?

 

[EnderWiggin]

The closest thing to a video or review I've seen online about Schneider and unbalanced loads would be the Abney Solar one, where he takes a massive compressor and places it on a single phase of the XW Pro to kick it up over 9kW:

This video isn't definitive - he glosses over, IMNSHO, many of the setup finer points that some of the critics of the video note (i.e. did he change the Sol-Ark's defaults, etc), but it does show excessive unbalanced loads on the Schneider and how it stays running.

The XW Pro's nominal, steady-state run wattage is 6800 in split-phase mode @25C, 6000W @40C. They don't specifically call out 'how unbalanced' it can get, though the same inverter when configured to run in single-phase mode (or as a member of a three-unit 208Y setup) is 5760W continuous, at 25 or 40C. In both configurations, they can surge to 12kW for up to 60 seconds, and for 30 minutes they can get to 8500W (split) or 7200W (single/208). So Ben's video was in spec, as long as the dryer didn't run its element 100% of the time for a full cycle.

I'd be interested to see someone try and put all 8kW of load on a single phase of a Schneider; my guess is, it would handle it though, based on Abney's video. For a time, at least. Schneider and Victron typically limit their inverters not by any kind of measurement of current/wattage, but how hot their systems get, so your mileage may vary. I think I read somewhere that Sol-Ark actually watches the legs and if it sees more than 2500W difference between the two legs, it can potentially trip to protect itself, but don't quote me on that.

As for the conditions for how it failed for Ben, you'd have to watch the videos I linked above. My understanding is, the Schneider was configured in a purely failover mode of operation - meaning, it was not normally delivering to loads, only when the grid was lost, and the software was horribly glitchy and failed to supply power when required (when it was switched to by some other transfer switch). That, coupled with what was, at the time, horrifically bad customer service/support from Schneider, made him replace the unit with a Sol-Ark.

 

[GXMnow]

I did not mean to dis on the Sol-Ark, just showing how you can beat the crap out of the Schneider and it just keep working.

The XW-Pro is well protected still. But for overload protection, anything under 12,000 watts it only limited by thermal protection. If the mosfets or transformer gets too hot, it will shut down, but it can take a while to get there, so it will do what it can. But if you exceed 12,000 watts, it shuts off pretty quick. I saw my XW peak to 8,000 watts when we were welding and my 3 HP air compressor kicked on. Just like in his video, the lights didn't even flicker. But I was cheating a bit, I was still grid connected, but the XW was providing the power to zero the grid input.

This particular video caught my interest because he is the same guy who did yank out at least 2 Schneider systems and put in Sol-Arks. From what I can gather from his videos, he has the same complaint as us. AC coupled only situations. The Schneider AC coupling needs better software so that it can start a charge cycle without needing an external controller. It really is quite sad that they won't accept this advice. Ever since I got my PLC working, I really can't complain at all. It just flat out works. The Gateway has locked up twice, and my PLC froze twice, but I traced my PLC issue to a divide by zero, caused by a bad read from the cheapo amazon power meters. The gateway issue should be fixed when I update the firmware. I was going to do it this week, but I can't bear to shut down half my power while it is 106F outside.

 

[pvdude]

Saw an article stating that the utility here is raising prices next year.

More loads could be moved on to the system, if I had more PV.
Here is a typical "grid support" snapshot showing the utility supplying the deficit.

I want to add more panels, but there are none of the kind I have installed available (REC365AA)
I saw 400bird's post about buying some very powerful panels. Gave me some new ideas.
Adding another charge controller, it would be possible to put any newer technology / high power panels on it.
(I left room for installing another CC, as I keep painting myself into corners, but am learning to plan further ahead.)

Another CC would avoid concerns about mixing PV panels on the same CC, and avoid any other issues with incompatible Voc/Imp.

All mine are DC coupled, no sell-back, so I would stay w/ DC Coupled panels.

Rough estimate, using 480 watt panels, two strings of 7S would fit on the remaining shed roof space available.
That would add approximately 6720 watts to the PV capacity, for a total of 15,480 watts.

After losses are accounted for, I would expect the actual PV output, under perfect insolation conditions, to be close to the 2x XW Pro's 14,400 watts output.

 

[400bird]

In the video he mentioned installing smoke detectors for the battery. Wireless. Inside a Faraday cage?
Although, door seal is plastic, does provide a slot for radio waves to pass through (polarizer).
Hope he tests signal strength/communication.

I've got my battery in an all steel job box with a wifi smoke detector and BMS inside.

BMS signal -58 dBm
Smoke detector signal -56 dBm
Edit: minus, not plus...

Smoke detector:

X-Sense | XS01-WX Wi-Fi Smoke Alarm/Detector

By checking your smoke alarm’s status via the smartphone app, you can act quickly to safeguard your home against potential fire hazards.

www.x-sense.com

Initially I was concerned that I would need it build/buy a tiny access point to put inside the batte box. But, it's been good enough without anything. The BMS occasionally drops Wifi, so I might still do something about it.

 

[Shimmy]

Initially I was concerned that I would need it build/buy a tiny access point to put inside the batte box. But, it's been good enough without anything. The BMS occasionally drops Wifi, so I might still do something about it.

I got a bunch of those mini-router 2.5gHz doodads for when I need wifi inside a metal box. I run a power-over-ethernet cable to the box, use a POE/5V splitter, and the router. I would prefer to hard-cable everything, but Espressif saw to it that this would not be possible...

 

[400bird]

I got a bunch of those mini-router 2.5gHz doodads for when I need wifi inside a metal box. I run a power-over-ethernet cable to the box, use a POE/5V splitter, and the router. I would prefer to hard-cable everything, but Espressif saw to it that this would not be possible...

That's my issue, these are ESP based devices. Got a link to your doodads?

 

[Hedges]

Actually, just a wire feeding through from inside to outside would act as a passive repeater.
More efficiently if optimum length (e.g. 1/4 wave) for an antenna. Two antennas connected by coax could be good.

 

[Shimmy]

Got a link to your doodads?

Not sure if the Amazon app does affiliate links or not; that isn't my intent, but here: Gigastone TR1 Mini Wireless Travel Router 150Mbps, Router, Repeater, WiFi Range Extension, 802.11N, Built with Qualcomm Atheros AR9331 WiFi Ethernet Processor https://a.co/d/hFb3ctC

These are the power splitters: UCTRONICS IEEE 802.3af Micro USB Active PoE Splitter Power Over Ethernet 48V to 5V 2.4A for Tablets, Dropcam or Raspberry Pi 2/3B+ (48V to 5V 2.4A)

These might not be the exact models I got; the names and brands seem to change constantly.