Wiring Plans For Modular System

Hello everyone!

So I am in the process of planning out a modular powerwall system. I would like to submit the following wiring plans for critique by this group.

Cost effectiveness is not a major concern for me. I dont mind over-sizing components to make things easier, safer, or more attractive. I am doing this as a hobby project.

Max peak planned current draw from each cell will be 1 Amp. Current Draw under normal circumstances will be about 250-500mA from each cell

The modules will each hold twelve 18650s in a parallel group. I will have a 3A fast-blow glass fuse on the positive terminal of every cell. I plan to use stranded 10-gauge wire (insulation mostly stripped off) as the positive buss wire and negative buss wire. These two buss wires will terminate at an XT-90 connector. Important to note that these modules are constructed of 3D printed cell holders using leaf-spring connectors. Fuse leads and the smaller connecting wires solder on the back of the leaf-spring connectors and to the buss wires.

My thought is that the 10-Gauge buss wires will be over-sized for the planned current draw. They should also be large enough to allow for easy soldering vs smaller wire sizes. The XT-90 Connectors should easily accept 10-gauge wire. Smaller XT-60 connectors probably do not.

Here is a simplified drawing of the proposed module wiring:






The modules will be interconnected in parallel and series using 10-Gauge wire. I think that I can just keep adding the 10-Gauge wire from each module into the bundle so that the total wire diameter keeps increasing as each additional module joins a parallel group. The stranded 10-Gauge wires will be stripped of insulation where they run together so that they are twisted and wrapped with the other wires. The following drawing shows an example using parallel groups of 4 modules in a series of 3. You should be able to see how the buss wire grows larger where the current is expected to be higher.





These are greatly simplified examples of the project. But, I believe the same model and electrical principles would apply with a larger group of parallel modules, and a larger group of series connected groups. I just keep adding the 10-Gauge wires to the bundle with every additional parallel module.

The actual scope of this project, or physical space I have for the project, calls for an 8S configuration. 12-cell modules can be 18 to each parallel group. So the end result would be 216P,8S (1728 cells). I can simply add the modules 8 at a time to that maximum space capacity.

Thoughts?

If cost effectiveness is not a major concern, why use such small format cells ?

Sean said:

If cost effectiveness is not a major concern, why use such small format cells ?

Click to expand...

Excellent point. I am harvesting the cells from laptop batteries and similar sources. So they are basically free except for my time. If I were to purchase new cells I probably would have gone with something like 26650's. Also free will be a 3/4 height network/server enclosure to enclose all of the modules. That is why I have a set space limitation.

Some potential issues I see with your design:
* Large number of small packs probably looks messy andleads to a lot of wasted space inside the enclosure
* A lot of work/plastic/wire/parts in general vs relatively few cells. There are good reasons why most people go with 7s60p ~ 14s200p
* Bundling up to 18x10 gauge wires iseasier said than done



10 gauge wire indeed does not fit into an xt60. But you can make it fit bycutting a few strands at the tip. Looks fine when filled with solder.

ajw22,

Thanks for the images of the XT-60 technique. I may try that instead. I have not purchased any connectors yet, so Ill try the XT-60's first and see how it goes.

I am considering your other points also. I know you are correct. I'm trying to figure out if the extra work/parts is worth it to me. I do enjoy a good project and challenge.

I will post more of the design soon. I think it will look neat and tidy if I put enough effort into the physical layout. And, even though I am talking about 144 modules max, I will just add them 8 at a time. I currently have about 500 tested good batteries. I probably have about 40 laptop packs I have not opened yet. I also get a moderate steady stream of them. I think it would realistically take me 2 years to actually fill the wall with all of the modules.

One other thing is the 3D printing time. Ill only be able to pop out a couple of modules and brackets per week.

As you probably know, I also use 3d printed parts for my powerwall. My initial plans actually also included PCB boards with spring loaded contacts, but decided against using those for the reasons I already mentioned.
Another reason was that it would obscure the ends of the cells, making visual inspections difficult. As it turns out, visual inspection is very important when one uses very old ebike batteries that crap out way quicker than laptop or medical pack cells. I think people often keep using them until the ebike battery dies, whereas the other ones are retired much earlier.
Also, in the highly unlikely case of a cell spewing fire, I thought it'd be better if there were at least a few cm of free space for the gas to disperse and dissipate (wishful thinking?) before hitting something that could melt or even catch fire.

I'd love to see your 12-cell holder design. You're going to arrange the parallel holders like books on a bookshelf (ie the cells standing upright)? And the 8 shelves are connected in series? Are you going to 3d print some sort of slots / spacers?
One annoyance I'm facing are the cables/connectors that get a bit in the way when taking out a pack for inspection. You're going to have 8x times as many.

Some tips to speed up 3D printing time:
Use 0.5mm nozzles and 0.33mm layer height, even if you have to slow down the print to make it work. Soooo much faster than the typical 0.4mm / 0.2mm setup.
Provided you can let the printer run unattended, try to print multiple parts at once. Reduces down time when you can't set up the next print. For smaller parts, I've done up to 32 pieces at once!

ajw22,

Ill take some pictures of the module design soon. Right now I have a rough prototype that I have not taken the time to clean up. It is not yet something that I am proud enough to show off. PETG is stringy.

Basically the module is made up of two 6x cell sleds. They are in a triangular configuration back to back. So, the cells are not quite vertical. They are leaning back slightly. There are spacers between the sleds that facilitate buss wire tie down and provide access to solder points on the top and the bottom. Negative connections are on top with the narrow space. Positive connections are on the bottom with the wider space. This gives me more room on the bottom for the fuses.

There is a separate mounting bracket that actually screws to the wall. The module simply slides onto the bracket and locks into position. The module interconnect buss wires will actually be behind the wall (hidden but fully accessible). Those interconnect wires run through the wall and mounting bracket to the front of the module where the XT connection is made.

So basically you load up a module with cells, slide it onto an empty mounting bracket, and connect the XT connector.. If you need to remove a module you just unhook the XT connector, and remove the module from the mounting bracket. The cells also pop in and pop out of the module because they are not soldered into place.

I will pre-wire all 144 of the mounting brackets with module interconnect cables. So they will all be sitting there waiting for me to add modules as I print them and bring newly tested cells online. I will print all of the module brackets first, and just enough of the actual modules to get me started.

All of this will be mounted inside of a network/server enclosure. I am going to build "walls" on the left side, and right side, that run from front to back of the enclosure. So when you look inside of the enclosure you will see 72 mounting brackets on the left, and 72 mounting brackets on the right. The inside of the cabinet is where you load and unload modules. If you remove the side panels of the enclosure, you will see the backs of the "walls" on the left and right sides. That is where all of the module interconnect wires will be run and hidden.

Basically I am turning a network/server enclosure the size of a refrigerator into a giant battery. It is all metal with ventilation on the top, bottom, front, and back. Casters on the bottom so it can roll. It locks with keys also. I have not figured out what I want to build the walls out of yet. At first I though plywood, but I would rather find a material that does not burn as easily. Maybe something like those plastic porch/deck boards??? Not sure.

Dont forget you'll need to manage the cell voltage every time you parallel connect it to an existing such that the voltage is exactly the same - you can't simply connect a full cell (or packs of cells) in parallel with an empty one.

Sean said:

Dont forget you'll need to manage the cell voltage every time you parallel connect it to an existing such that the voltage is exactly the same - you can't simply connect a full cell (or packs of cells) in parallel with an empty one.

Click to expand...

Yeah that would probably pop cell fuses. I would not want to do that.

Ill have this sitting next to a workbanch. I can make sure that I have charging hardware there to bring individual cells, or full modules up to charge matching the others.

I have been working through some technical difficulties with soldering the prototype. Basically I managed to start melting the plastic under one of the leaf springs because I got it too hot. I need to practice a bit. I may just twist the copper tightly in the little loop on the back of the leaf spring and then just put a tiny dab of solder.

The concept is a mount on the wall (with accessible rear side). The cables go through the mount to the back of the wall. The module slides onto the mount and is secured by some tabs on the mount. Connecting the XT connector will also secure it in place.

Here is a rendering of the mount.




Here are 2 renderings of the module. It is made up of 2 6x 18650 sleds with spacers in between them. The spacers will hold the buss wires in place. Negative on top and positive on the bottom:





Here are 2 renderings of the module sittingon the mount:





Printed Protype:



Leaf-Springs snap into place very securely. You have to pry them out with a flat-head screwdriver to get them off:






Buss wire mounted:



I will post more as I get time. I will work on this as I get time after work and stuff.

That's a whole lot of space and plastic to house just 12 cells. There must be a more efficient and compact design?
AWG 10 for just 12 cells is overkill especially if your absolute peak draw is 1A/cell.
How is the sled/spacer put together? Glue? Screw? Snap-on?
Do you have a shopping link forthe leaf springs?

I've had problems with zip ties allowing the straight cable to slide length wise, no matter how tight I made it. Simple solution: bend the end into a "key" hole.

ajw22 said:

That's a whole lot of space and plastic to house just 12 cells. There must be a more efficient and compact design?
AWG 10 for just 12 cells is overkill especially if your absolute peak draw is 1A/cell.
How is the sled/spacer put together? Glue? Screw? Snap-on?
Do you have a shopping link forthe leaf springs?

I've had problems with zip ties allowing the straight cable to slide length wise, no matter how tight I made it. Simple solution: bend the end into a "key" hole.

Click to expand...

That is a lot of space and plastic. For me that was the trade-off to allow simple insertion, removal, and inspection of each cell individually. I have laid this all out and can fit 144 of these modules into the server rack. I will alternately build a row of "service" mounts on my workbench. That way I can mount a module for pre-balancing and testing/troubleshooting. There may be a tighter design. But it would have to involve the spacers because the cell sleds cannot get any smaller.

My first iteration was a little tighter in a square format. But I found that I didn't really have enough room for the fuses on the positive side (working my fat fingers in such a small space) I widened out the bottom positive end to give me a little extra room. That also eased my paranoia of a cell working its way out of the module by tilting them backwards on an 80 degree angle.

The modules are 123mm Long (4.84") X 82mm Tall (3.23") X 91mm Wide (3.58") So what seems gigantic in the pictures is really not that large.


I am going to follow your advice on the 10 gauge wire. You were absolute right that it is difficult to work with. It is tough to strip and difficult to bend in a tight radius.It is also a pain in the butt to try to cram into an XT-60 connector. Ill retry this with 12 gauge wire. I do not want to go too small because the wire comes in direct contact with the spacer. The material is PETG, so more heat resistant than PLA. But I do not want to riska hot wire warping it.

The sled/spacers are put together with tiny wood screws (style not material). The heads are countersunk into the bottom of the cell holder so they do not come into contact with the battery. I print the spacers with a 75% infill. I drill a tiny pilot hole and do not have any cracking issues. I also get a surprisingly tight hold.


Leaf-Spring Negative End:
https://www.digikey.com/product-detail/en/keystone-electronics/209/36-209-ND/151583

Leaf Spring Positive End:
https://www.digikey.com/product-detail/en/keystone-electronics/204/36-204-ND/227461https://www.digikey.com/product-detail/en/keystone-electronics/204/36-204-ND/227461

At first I used the 209's on both ends. But the contact area on the 209 is very large. I did not like it on the positive end AT ALL! The contact area on the 209 is large enough to risk bridging the positive contact with the cell shoulder if not perfectly lined up. Switching to the smaller 204 on the positive end makes it closer to fool-proof.


Good call on the key hole for the cable end. I will work something like that into the design.