Q1:
A1:
2 main reasons for the busbars - material and shape. We want a low resistance in the busbars so that there is less heat produced in them (P = I^2 * R). With less heat production, we can support a higher maximum current draw for the same temperature target (the max temperature of the cells in the datasheet) - i.e. if P is constant (based on our cooling design), and we can reduce R in our design (based on material and shape) then our I increases, allowing us to give more power to the motors. Or, if our max motor current (I) is constant, then is we reduce R then we can reduce P (the power lost in the busbars) and thus reduce the module temperature.
Resistance can be calculated from the equation:
http://hyperphysics.phy-astr.gsu.edu/hbase/electric/resis.html
A resistivity table with Copper, Nickel, and Aluminum can be found here: https://www.electronics-notes.com/articles/basic_concepts/resistance/electrical-resistivity-table-materials.php
Resistivity is a material property, so we had to choose the material carefully to get a low resistivity.
The busbar material that we used is the EMS Sigma-Clad 60. More details about Sigma-Clad here: Module Busbars - EMS Sigma Clad 60 (read the designer’s guide especially). Sigma-Clad has lower resistivity than nickel, but is easier to work with than copper because it can be spot-welded with simple tools and it has better corrosion protection than copper.
Length and Cross Sectional Area depend on the design of the battery module
We were careful to ensure that the current flows in a straight path from one of the module to the other and does not wind throughout the module (ensure that cell connections to module bolted connections have short length and wide cross-sectional area).
As highlighted in the Design Guides for Sigma Clad 60 on the page linked above, a material with a relatively high thermal conductivity, it reduces the temperature difference between the hot and the cold parts of the busbar by spreading the heat better than nickel.
Q2:
A2:
The risk of shorting the cells when terminal touched was not really part of this decision. In the figure shown in the question, there is adequate separation between all the cell terminals, with an insulating material in between, which is sufficient.
Cell Can Shorts
It is also good to know that the can (the outside wall) of the 18650 is also part of the negative terminal. It is normally covered in colored heat-shrink PVC tubing, but in the case where the pack is crushed or a sharp metal object is dropped into the pack (screwdriver, allen key, bolt, wrench, etc.) there is a risk. If the object falls between 2 cells and punctures the PVC heat shrink tubing around the outside exposing the metal can walls, and if the object touched both cans at the same time it will cause a short circuit is the cells are connected in series. For series cells that are directly beside each other in the MSXIV modules, we added a piece of ‘fish paper' insulation (or vulcanized fiber insulation, often used in telecom and battery applications). Some images below. That covers cell-to-cell shorting on the negative side.
TODO - add a picture of this here as manufactured.
Shoulder Shorts
Another type of cell shorting to be aware of is ‘shoulder shorts’. Because the cans of the 18650s extend all the way from the negative end to the positive end of the cell and are crimped around the positive end, a single conductive object placed across the positive surface of a cell and touching the edges will cause a short. Read the section on shoulder shorts in the doc below. In the image below, the metal tab is spot-welded to the positive terminal of the cell, and when bent downwards, broke the insulation and touched the negative
We wanted it to be impossible for a wrench dropped on the pack to cause a short, so we made sure to make covers for all the exposed conductors (wires, busbars, or metal components touching those).
In order to protect against shoulder shorts, some pack builders will use fish paper insulation rings as shown below on the positive terminals of their cells.
This works pretty well, but is only required because the plastic holders that are available off-the-shelf and often used do not cover the whole ring around the positive terminal that is susceptible to shoulder shorts - they only cover the corners, so the extra fish paper insulation is required.
We designed our 3D printed cell holders to fully cover the area susceptible to shoulder shorts, so the ring insulators were not required.
Thermal Runaway Propagation
Now, the real reason that we avoided this configuration:
When a cell goes into thermal runaway, it pressurizes the inside and then explodes. Some cells are designed such that when the pressure builds up inside, it will be released from a specific location. They do this by weakening the can in the location that they want to cell to ‘vent' and release all the hot gasses built up inside. Often, the positive end of the cell is the location that the majority of the Thermal Runaway energy is release through. The presentation slide from NASA shown below confirms this, that 76.1% of the energy contained in the cell will be shot out through the positive terminal.
When the energy shoots out the positive terminal, it will heat up anything in its path. If there is another cell in its path it will heat that cell up and cause it to go into thermal runaway, which causes a chain reaction if many cells are connected in this way (thermal runaway propagation). If however, we remove all the cells from the path of this thermal runaway energy, we should be able to reduce the chance of thermal runaway propagation. This was the motivation to not have any cells stacked vertically on top of each other.
Note then in the final module design, we had an acetal sheet placed on top of the cells, and some cells in the module had the positive end facing down, so there is more resistance to gasses trying to escape from the positive end since there are more things in the way. We also used a potting compound to cover the spot welds, and decided on a semi-flexible compound instead of a rigid compound to give less resistance to any TR ejections trying to escape the top of bottom. We did not test any cells going into thermal runaway, but my guess would be that there would be some cells that blow out the side and cause thermal runaway propagation due to the increase in resistance out of the positive end compared to in free air as the tests in the NASA research were done.
Another great presentation on TR propagation minimization:
Q3:
A3:
A few tradeoffs to consider here:
A manageable assembly size. A single module with 200 or 300 cells is pretty heavy and would need 2 hands to hold it, harder to move around and keep extra modules. At ~50g/cell, 200 cells weigh 10kg and 200 cells weigh 15kg. Smaller modules are easier to handle.
If a module fails, we need to replace the module. If we had only 4 modules in the pack, then we may only have the budget for a single spare module. Also, since there are more cells in a larger module, there is a greater chance that a cell within that module fails. Smaller modules means we can have more spare modules, and will statistically fail less often.
It is beneficial to make all modules exactly the same. Then we only have to keep 1 type of spare module, and any spare module can be swapped in for another. MSXII had 2 types of modules (and end module and a middle module) and we ended up running out of spare end modules, while still having a few spare middle modules unused.
More modules mean that we have more inter-module connections. This increases the number of failure point. Each inter-module connection also has a small resistance associated with it, so more connections means higher resistance in the pack’s current path, and more power loss (P = I^2 / R).
If you compare the weight of the cells in a module to the weight of the rest of the items (cell holders, support brackets, cables, bolts, etc.) then the weight of the extra stuff in a larger module will be smaller per cell (LM = Large Module, SM = Small Module):
(LM mass of extra items / LM # of cells) < (SM mass of extra items / SM # of cells)
The extra items can also be considered ‘overhead weight’ and would be roughly constant for all module designs. More modules = more extra weight.
The layout of the modules within the pack may also be easier when working with smaller blocks.
Q4:
In case the link ever gets taken down, here’s the PDF:
A4:
Phase change material! Yes - this stuff is awesome! All Cell Technologies is where I first heard of it when looking into cooling for MSXIV: https://www.allcelltech.com/
Here’s another company that we looked into as well - Outlast Technologies:
I would absolutely love to do this if we can figure out a way. Increased safety is the biggest reason to push for this as it limits thermal runaway propagation. NASA also have a ton of presentations on this - just search ‘NASA 18650 Thermal Runaway Propagation’ on google and you should get a bunch of results.
We did look in to this a bit, and determined that it would be practically impossible to acquire some material to use (we did send an email to All Cell Technologies but they said they don’t sell the material directly, they will only sell it as built into a pack that you buy from them, which was super expensive). I did look into what it would take to make some, and determined it to be too difficult of a technical challenge. We also wouldn’t really be able to test the thermal runaway propagation technologies since the university wouldn’t really like us burning up cells on purpose. There may be a few battery labs on campus that we can reach out to about spaces to do this testing though. The main reason we did not pursue this route is that we needed something to work and be complete in a short time frame. We did not know if making a phase change material from paraffin wax stuffed with some kind of thermally conductive powder would even work, so we decided not to invest the resources into it. On this new cycle though, it would be great to reconsider!
Other resources for PCM:
https://ctherm.com/applications/phase-change-materials/
https://endless-sphere.com/forums/viewtopic.php?t=97512
https://endless-sphere.com/forums/viewtopic.php?f=14&t=84857&p=1241253&hilit=Outlast+phase#p1241253
https://www.allcelltech.com/pcc