Sunpower 6.5kwh Guts

[hogback]

Hi all,

A friend had some sunpower battery systems that he modified to basically keep the xwpros that are in there, and asked if I wanted to try to use the batteries. They are mostly in the 52V range sitting there. They have some sort of internal relay that activates when you push the power button, and ethernet type ports to daisy chain the batteries, and I gather once you terminate the chain, there is another port that goes to the inverter or insight facility module, which I also have. The only documentation I found online for the pcs (they call it) refers to grid sell which doesn't apply to me, and also that it isn't compatible with dc coupled systems with sccs like I have. Does anyone know if any part of this system is supposed to control/limit current going to the individual batteries when they are all wired in parallel? Any thoughts as to if there would be any difference between double-terminating each battery, having them operate blind (like my simpliphis do now), and actually daisy chaining the communication cable between the batteries? Basically, does this daisy-chaining actually do anything if the last battery is NOT connected back to the inverter/insight?

Thanks!

hb

 

[hogback]

It's these batteries (images taken from web):

 

[hogback]

I do realize that the schneider bms communication does not work with
-more than one xwpro
-dc coupled systems

I'm imagining that the batteries themselves for some reason cannot work open-loop. They just power off after about ten minutes of being on. I also wonder if they require some charge or discharge to remain on, but 5A from my small charger didn't help. Can anyone imagine why they would power off like this? Thx,

 

[Crowz]

I just spent a good bit surfing for info on that battery and pretty much found zilch for useful info.

It seems to be something your not supposed to have unless you bought a package from a dealer and they installed it. I did find it for sale used a bunch of places though. So hardly any info out on it. Sorry I couldn't find anything.

 

[hogback]

That's my understanding too. Thanks for having a poke around. I'll report back if I can find anything. I think my friend still has the shells that all the stuff was in. The inverters were already out, so it was just a mess of wires. I should have taken a closer look at how it might have been wired.

 

[hogback]

The appearance and specs look identical to one of the growatt offerings. The manual for that specific model indicates that it will shut off if it doesn't hear anything from the pcs. The wiring instructions show it connected to a growatt inverter with an rj45 comm port. Anyone ever run a growatt battery on its own? I suppose the next step is to read about the growatt inverter to see what it is doing with that port.

 

[hogback]

Not sure if I'm following their language here, but do you think that if I put a few volts (5? 12?) between pins 1 and 2, it would keep the batteries on? Would this just be a momentary pulse, or might the V need to remain there? Thx,

 

[hogback]

Although the battery 'wake' seems to be pins 7 and 8....

 

[hogback]

Looks like the pcs is connected via com, not rs485. Would there be any harm in trying this and seeing if I can get the batteries to stay on? Might the insight facility see them on their own, or would an xwpro be needed in the xanbus network? Thx,

 

[42OhmsPA]

I'd pull them open and replace the BMS so that you can use them pretty much any way you want.

 

[hogback]

That's a thought. I have a lot of knowledge about the electrical and electrochemical properties of materials, but am not actually that savvy with electronics or programming proper. I might have some more reading to do....

 

[hogback]

I wired the insight facility to a battery - CAN H and CAN L only. Not sure if the ground is needed. This is assuming the battery is equivalent to the growatt. No dice. Insight local does not detect anything. Perhaps it is because I have no inverter connected? I suppose what is curious is that the battery needs to "see the PCS." I can't see what the battery would need to do with information from the rest of the system.

On a separate note, I took the top and side off the battery. I could probably cram a diy bms in there, and leave all the factory wiring in place. I'll try to grab a few photos later. I wonder if the T-sensors they have in place are of the same sensing modality as the ones that come with, say, the JBD/overkill.

 

[Lt.Dan]

I would rip it all out, scavenge the cells and put a JK or JBD BMS on it and run it, if you can

 

[hogback]

If I do my own bms, it would at least be nice to salvage the case. All the bus bars and voltage monitoring wires are in place. Plus, some of the bus bar jumpers look integral. In the close view, the cable bundle upper left is all the cell voltage measurement and temp sensors. On the right bus bar, there appears to be a shunt. The large black circle below each bus bar is a relay, turned on by an unknown signal. I'd have to jump these or figure out what closes them. That is, if I wanted to use the factory amphenol connectors on the front of the case, which would be nice. The bms board is on the left side of the top of course. The rest of the small wires seem to be for V measurement around each side of the relays, and of course all the communication stuff for daisy-chaining multiple packs, and whatever was supposed to go to the insight facility. The opening at top is 4" wide, so I'd have to tilt a jbd diagonal to get it to fit, which seems like a good option. The shunt looks so small and fragile. I think I only see three copper hooped bus bars (left image), and the rest of the jumpering seems integral somehow. Like 12V integrated packs. Ever seen that before? I have a few of these to play with. Would be cool to make a second, independent system for the garage.

 

[hogback]

Each 'cell' looks like a tesla pack form factor. There's alternating gold and silver sawtooth zig-zag markings on each end, which seems to correlate with +/-.

 

[hogback]

I had the wires on the cut/stripped end of the homemade cable between the batteries and 24pin connector on the insight facility wrong. Fixed, and also now running an xwpro on xanbus with the facility. Good news is that I can see each battery (on their own) as a bms device in insight local. But I can't daisy chain more than one battery via the link in/out connectors. I do terminate on each end. When I go to turn on one battery (which is supposed to turn on all batteries in the daisy chain), it turns on the one with the PCS connection to the insight facility, but the other one doesn't fully turn on. The lights blink, but the top one just flashes. The battery doesn't ever show a V between the battery connectors. The insight facility doesn't see the other one.

I wonder if I need to set each one up individually with a device number and device name, and then try the daisy chain. The BMS is recognized as Sunpower v1.2. Also, when one battery is connected, it stays on indefinitely, NOT shutting off after 25minutes like it did without the proper PCS connection.

 

[hogback]

Do far no luck. I did determine that it's not the battery that has the pcs connection back to the insight facility that determines which battery powers up - it's which is master (powers up fine) or slave (lights on battery flash, but doesn't actually turn on). You determine which is mater or last slave by which port (in or out) the terminator goes on. With the terminator or COM/LINK in, that battery becomes master. Terminator on COM/LINK out, that battery becomes 'last slave. With more than two batteries, all non-terminated batteries become slaves.

Something about the slaves is not letting them turn on all the way.

 

[hogback]

Not much progress here. I did learn that the cell configuration in these batteries is actually 16s 2p. Would a diy bms work with this? How does this configuration remain safe if you're not checking the potential across all 32 cells?

I'm not sure when exactly this happened, but now all the xwpro fields in insight local are greyed out - can't change any of the settings.

Another question I have is about where the command comes from to shut the whole bank off if a cell is out of range. If I hack these things, with their current internal bms, and the system relies on an external signal (from the insight facility somehow?) to switch off, this wouldn't be safe as no signal would ever come.

Also, in addition to the growatt batteries that look identical to these, there is also a 6.6kwh Chelion battery that also appears to be identical. The Chelion documentation is marginally better than the growatt. This is where I learned of the 16s 2p configuration.

I did have a few back and forth emails with growatt about their battery, but they were not much help. They say the WAKE+/WAKE- pins need to be connected to make the slave batteries work, but not how/why.

 

[hogback]

Attaching the Chelion documentation:

 

[hogback]

A correction - I was using the abbreviations incorrectly. These units are 2p16s.