diy solar

diy solar

Adding Schneider XW Pro

I finally had some time to work on the new Bolt Battery modules. I got one of the 10S blocks cut into 4S/2S/4S and have mechanically spliced the 4S sections onto the end of each 10S block, here are a few pics.
IMG_9353.JPGIMG_9354.JPGIMG_9356 (1).JPGIMG_9357.JPG

IMG_9360.JPGIMG_9358.JPG

It is certainly a bit nerve racking when you cut into that plastic insert. There is more room than you think, but it still feels wrong to be pushing a saw into a lithium battery bank. I chose a little different approach to combining them.

The pack with all the cells the same direction, and no separator plate has the added 4S section upside down. So the negative stud is down instead of up. This way, I just need to make a single buss bar to go from the minus stud of the 10S pack to the tabs and buss bar right there in the 4S block. I had to get a little creative on the end of that pack. I used one of the center separator plates, then cut a plastic plate to fill in the back side where there is no cell, and used a heavy steel plate which is meant to hold ratchet strap in a box truck. I also made spacers to go around the 4 studs so the threaded rods are under tension clamping the corners, and the end plates are just held at the thickness of the battery pack with no bulging. Since the cells are very near fully charged, they are bulging a tiny bit, so here is a bit of pressure on them now, I really need to pull them down a bit.

The second pack actually turned out much nicer, and I knew it would when I chose how to arrange the batteries. I picked the full 10S that was the same polarity as the cut pack when the output studs are facing towards me and up. The factory GM/LG mounting plates are then at opposite ends just as they would be in the car, just missing a group of cells. The inner end of both packs has the cull extending out with the aluminum cover. I had a second of the center spacer plate from my first packs still. It fit perfectly between the cells and properly retains the tension just as it would between full packs in the car. I just had to cut new threaded rod the correct length, and it bolted up perfectly. Again, it just needs a single custom buss bar to go from the 10S pack stud to the 4S block tab. And then I have to make up the balance wire harness.

After trying to lug these beasts into my kitchen (ceramic floor) I was curious what they did end up weighing. One of the 14S bricks, 9 KWH's, came in just over 98 pounds. How does that compare with an LFP bank?
 

Attachments

  • IMG_9356.JPG
    IMG_9356.JPG
    198.3 KB · Views: 11
Yeah, it's not a lite pack...

Based on these random blue cells that seem popular, it think they would be at 115 lbs for almost exactly 9kwh. But, would still need to add compression.
But, those aren't a great direct comparison, as 16s is 14kwh at 168 lbs.
 
The Modbus command to force Bulk Charge is
Address 0x0165 uint16
Name Force Charger State
Set to 1 = Bulk, 2 = Float, 3 = No Float
Wonderful, thank you!

I had tried the XW modbus map earlier in my attempts, but was unsuccessful. Since my initial failure, I have been working from the Gateway modbus maps.

Now (mostly just after trial and error to figure out what I was doing wrong...) that I try the XW map and use the address you provided, I can read that register, but I get an error when I try to write:
Error: Modbus exception 3: Illegal data value (value cannot be written to this register)

Have you been able to change the charger status?
Are you using RS485 or TCP?
I have been using TCP, but I could give RS485 another go if it works on our set up.
 
I have not written that register yet. I used a software Modbus TCP program and have been able to write other registers, but the software was flakey. With my PLC that does modbus TCP, I have only been reading registers until I am confident in my handshake routine, which I am not yet. I will try to write soon, but have to do a register that is not important first. When I was writing with the PC software Modbus, it wrote an illegal value, and the XW locked up. I had to reinstall the firmware to get it back, so I have been a little reluctant to just poke around. This is the PLC board I have.
I am still having some issues with timing on the modbus TCP commands. The XW-Pro via the gateway is SLOW. If I send the connect request, then wait, then check the connect status, it says <connected>, but when I try to make the program do those in a ladder, it fails. I am to the point of just making the PLC wait for a few seconds between steps as my polling routine fails. I used to be good at learning a new programming language, but my brain is just not as flexible as I was 20 years ago. Stepping it manually through the commands, I am able to read any register reliably. So I know the communication is correct. I watch the battery voltage climb during charging and it works just fine. But I have to manually connect, check link, read port, disconnect.
What are you using to read/write modbus to the XW?
 
Wow, that is a step above my skills. Looks great.

I am using a Raspberry Pi 4 and Node-Red to send the commands via modbus TCP
I have an energy meter (software and hardware) running on the same pi, and also getting info from the BMS.
All of that is feeding into a InfluxDB and displayed using Grafana.
 
I figured it out, my problem was the difference between the registers starting at 1 or 0.
If I send the bulk charge command (1) to 357 it works! I can command the XW to charge.

For future reference of those going down this path:

Bulk charge command:
Modbus port 503 node 10 (or whatever node your XW is assigned)
Address 358 (possibly 357)

Charge command in watts:
Modbus port 502 node 1 (port on Gateway/insight)
Address 40210
 
Nice work. Now I need to get mine to work. After crashing my XW-Pro once, I am a bit gun shy about just sending commands at this point. When I have it reading flawlessly, I will try to send again. Thanks for confirming the force bulk charge command worked.
 
Nice work. Now I need to get mine to work. After crashing my XW-Pro once, I am a bit gun shy about just sending commands at this point. When I have it reading flawlessly, I will try to send again. Thanks for confirming the force bulk charge command worked.

Yeah, that is my concern and I am worried about it too. I sent a 1 to some random register. Luckily I was reading it at the time and changed it back to a 3 without issue as far as I can tell. Now I've gotten in the habit to double check everything before I click deploy or save.
 
Here is my new buss bar setup to tie the 10S and 4S sections together.
IMG_9369.JPGIMG_9368.JPGIMG_9366.JPG
Now I just need to make another pair of my clamps to hold these against the tabs/bus bar in the 4S block.
They only had 0.025 inch thick pure copper in stock, so I am using 2 layers. It should end up close to the total resistance of the factory buss bars. I am not worried at all about the current capacity. The new packs will never see more than 30 amps in real use.
 
Looks good. I like that.

Spreading the current between more packs is one of the reason I'm adding two more. When the current is at 100 amps I get some temperature rise out of the end cells vs. my temp sensor mid pack.
 
Wondering if anyone can share their AC1 or AC2 Advanced configurations? We've had some contiguous cloudy days here in So AZ (Monsoon season,) and I really need to get that generator online and stable. Here's what I have attached:
- AC Settings
- AC2 (Generator) Settings
- Generator Support
- Advanced Features (Generator Support Plus)

I can't think of anything else to show. My firmware: 1.11.00bn28 has a "Power Control Settings" section that isn't documented in the latest docs. It does have an "Import Active Power Limit" which I have set to 3600W. Everything else is disabled.

Any thoughts? I've opened up a ticket w/Schneider, but I won't be holding my breath for fear of entering a BLUE state. I've also put in a call to Alt-E where I purchased the XW Pro. They've been pretty helpful in the past.
 

Attachments

  • AC Settings.png
    AC Settings.png
    20.6 KB · Views: 8
  • AC2_Generator.png
    AC2_Generator.png
    309.1 KB · Views: 8
  • Generator Support.png
    Generator Support.png
    17.6 KB · Views: 8
  • Advanced Features.png
    Advanced Features.png
    41 KB · Views: 7
I'm not much help there, I don't have a generator.

I see you have the generator set as your grid priority, are you 100% off grid?

I'd widen out the disconnect (frequency and voltage) as far as possible as a test to see if you can keep the XW connected and happy.


One time, I couldn't get a setting to actually function, even through I could toggle it on and off in insight. I ended up switching the function on (then disabling charge and invert) and rebooting the XW, then leaving the XW alone for 5 minutes after the reboot. Then everything worked normally again and I could turn the function on and off. It might be worth a try.
 
I'm not much help there, I don't have a generator.

I see you have the generator set as your grid priority, are you 100% off grid?

I'd widen out the disconnect (frequency and voltage) as far as possible as a test to see if you can keep the XW connected and happy.


One time, I couldn't get a setting to actually function, even through I could toggle it on and off in insight. I ended up switching the function on (then disabling charge and invert) and rebooting the XW, then leaving the XW alone for 5 minutes after the reboot. Then everything worked normally again and I could turn the function on and off. It might be worth a try.
Any suggestion for how wide/tight I should make those?
 
After I had the issues with my generator disconnecting, I did spread out the numbers, but have not tested on generator since. I have the AC2 frequency limits set to 50 to 65 Hz and the voltage from 90 to 132 volts. My input breaker is set to 20 amps, that is the output breaker on my 5,000 watt generator. I may drop that setting to 15 amps and see if that helps as it should revert to batter to help keep the current under that setting. I do have the generator support setting at 15 amps, I may also lower that.
 
I took this weekend to mess around with the 6 extra cells out of the middle of the 10S pack I cut up. It was still all welded up as a 3P2S group, but I wanted 2P3S to run 12 volt devices. Most inverters seem to run down to 10 volts, so that should be fine.

Sorry I didn't get any pictures, but you have seen the cells separated to a point. Peeling one cell off of each end to make them just 2P was a lot tougher than I expected. Those spot welds are tough. And the tabs at the ends of the cells are very different. Even the welds are different. One side was not too bad to get off of the copper buss bar, but the other end hung on tight. I ended up tearing a little out of the middle of 2 of the tabs, but for my end use, I don't think it will cause a problem.

So I now had the 2 end 2P groups, the the center group has the two cells facing opposite directions. If the tabs at both ends touched together, it would be bad news. I put tape on the tabs of one cell for a little protection. The metal lid comes off of the 2 cell holder pretty easy, and with both cells tabs being free, the pair of cells basically falls out of the plastic holder, but between the two cells is a layer of foam. It is sticking pretty tight to the cells. Being pouch cells, I do not want to pull on them at all. I used 4 dull plastic knives. I very carefully worked them between one cell and the foam. IT does not seem to have any real adhesive, but it sticks like a million tiny suction cups. As I got it apart at one end, I just kept moving down and slid another knife in to keep them from touching again. After about 20 nervous minutes, the cells came apart. I looks at how the tabs were arranged and decide which way to position them, and carefully stuck the cells back together on the foam, and put the pair of cells back into the holder. Now both cells are facing the same way, and I am able to stick it between the other 2 pairs of cells. I did a final check with a meter and when I put the correct 4 tabs together on each end, I now have a 2P3S pack that measures 12.36 volts. I used some more of my aluminum strip but this time I did just drill through the tabs as well. I used the Ox Guard and cleaned up the aluminum bars and all of the tabs with sandpaper. I am not sure what chemicals are in the Ox Guard, but it very quickly eats the oxide layer off the aluminum. And of course, the sandpaper is actually aluminum oxide, and it pretty much removed that off the paper as well. When I built my first pack, I used a different brand called No Al Ox and I don't remember it being as messy as the Ox Guard, but I have to say this stuff really cleaned up the aluminum to a bright sheen. Of course, it was not 106 degrees out as I built that pack a year ago. The instructions say to coat the metal and brush it in and keep it coated as you assemble it. I have not pulled much current yet, but the connections look solid. The screws I used are long enough I can use a lug and a second nut for the balance leads to the BMS.

I also picked up an Apache case from Harbor Freight to house it and I will put the BMS and charger in there so I just have to plug it in. The inverter will probably bolt on the outside of the case. I am using the shut off switch and jumper cable leads from a dead jump pack I had laying around. I should install a diode pack. If I do jump start a car, the alternator voltage will try to go too high for the 3S battery. I think I have a few 150 amp 3/8 inch stud diodes in my old parts kit. That should hold up to a normal car start.

I'll post some pics as it comes together. I didn't want the 6 left over cells to go to waste.
 
Here is my new buss bar setup to tie the 10S and 4S sections together.
View attachment 53727View attachment 53728View attachment 53729
Now I just need to make another pair of my clamps to hold these against the tabs/bus bar in the 4S block.
They only had 0.025 inch thick pure copper in stock, so I am using 2 layers. It should end up close to the total resistance of the factory buss bars. I am not worried at all about the current capacity. The new packs will never see more than 30 amps in real use.
Nice and clean. Like how you doubled up the buss. Have you had a chance to bench test the module yet? I'd be interested to know how this solution works.
 
Sorry, I don't have a generator here, so I'm not help there.

What I do have is batteries outside in this 109° heat!

Last night around midnight the cells were roughly 100°, this morning with the battery box closed up, it was nearly the same.
With the lid open and some fans to move air around the cool air this morning, I was able to get it down to roughly 85°

The pack currently reads 114f, that is after stacking some shade junk around the box and cracking the lid to let the heat out once the inside was warmer than ambient.

I remember someone sharing a link or having one in their signature about calculating cooling load on an enclosure, but I can't find it now.
Does anyone know where I can find this or have any hints.
I'm going to be storing my battery bank in an outbuilding and when making the building, came up with this, which may or may not be useful to you: I have a very high water table in my area, so easy access to at least modest geothermal cooling via the ground water. I put a 1' dia. pipe vertically into the ground when making the foundation of the building. My plan, should I need to cool the batteries, is to run a closed loop into the pipe, heat exchanger coil submerged in the ground water, back up to a cheap car radiator with fan blowing into the battery box - so this would be added to what GXMnow suggests. Maybe overkill. Maybe too much work. Maybe too complicated. For my part, I am hoping I don't need to use it. Still, integrating a geothermal cooling solution could be one option rather than going refrigerant-based . . .
 
My current battery bank is a set of 3 year older Chevy Bolt packs. The same cell configuration. 6P14S of 60 amp hour Li NMC cells. They have been working perfectly for almost a full year now. They do not need any cooling at the current I am running, and that is with just the 18 KWH's of battery, supplying a maximum of about 4,000 watts. When I have the additional 18 KWH's connected, I could run the inverter a little harder. It is rated to 6,800 watts maximum. But to do it right, I will need to add the Watt-Node so that it will track my power consumption and not exceed my export limits. For now, I have it powering all my loads in my backup loads panel, and from 4 pm to 9 pm (our peak time of use rate) I also have it export 900 watts back to my main panel. When the air conditioner compressor is not running, almost the entire 900 watts is going back to the grid, but when the compressor kicks on, It draws 3,400 watts, so I am currently buying 2,500 watts when it runs. With the watt node, it would know to ramp up and push that 2,500 additional watts back to the main panel.

I currently don't have enough solar panel to cover that for the whole day. I have 4,800 DC watts, it produces up to a maximum of 3,800 watts for a bit around solar noon with Enphase iQ7 microinverters. I could charge more from that when the A/C is not running, but again, the XW-Pro does not know to ramp up charge power while exporting, so I just try to dial in the charge current so I buy a little power in the very early morning, but then it still exports a bit around solar noon. My current plan is to add about 2,000 watts of DC coupled panels to charge the battery bank directly. The house will run off of the AC solar when the sun is shining, and at the same time, the DC solar will charge the battery. Then at night, the house runs off the battery and I should need very little grid power. But I do not plan to go off grid. Even with the additional panels, I will have several days when my consumption will exceed production. Especially when I get a plug in car.
 
I made good progress last night and finally got active charge control figured out!
Today the battery charged with available solar power not used by my home loads.

I still have some tuning to figure out as there are a few spikes where the combination of changing load (it appears the old electric stove runs the burner on a very slow pwm) and my charge set up mean we pulled power from the grid when I didn't want to.

The labels in the key show "current" as in "value right now" not current as in amperage.
All the values are in watts or kW

Green is solar
Purple is battery charge
Orange is grid usage

I can see the coffee maker and house fan at roughly 7:30
I started the test run at about 8:15 when the battery first starts charging.
The instability from about 8:30-8:56 are kitchen appliances including the electric stove.
Right at 9:00 the car started charging on the slow setting on the slow charger (8 amps at 120 vac)
Wife left with the car at 11:00 and stopped the car charger.
Back just after 12:00 and the car started charging again.

I have it set to stop charging the battery at 4:00, I don't really know why it stopped charging after the AC kicked on for the first time at roughly 3:30. Those huge wattage spikes are the central AC cycling on and off.

In the afternoon, when the orange (grid) spikes deep into the negative that is pushing juice out to the grid with the AC off. When the purple spikes deep into the negative, that is the air conditioning powered by the XW/battery.

Overall, I think it's a pretty good first test of the system.
I do only have control of the charge rate at this point, I have not yet integrated the force charge command to start charging in the morning.

1626325309132.png
 

Attachments

  • 1626325309132.png
    1626325309132.png
    45.1 KB · Views: 0
Back
Top