MS16 Cell Selection
Note for People Just Starting
For an introduction to what kind of specifications are given for cells, read this: Battery Specification Guide (WIP). Additionally, take a look at the datasheet for the cell we used for MS15, LG’s INR21700 M50 cell: M50 Datasheet. Looking at what kind of specifications are given will give you an idea of what criteria can be used to choose a “good” cell for our pack.
Pack Constraints & Requirements
Constraints and requirements relating to the pack is important to keep in mind when doing cell selection. Consider what pack configuration would be appropriate for each cell, and what that implies for the pack.
Constraint/Requirements | Reasoning + Notes |
---|---|
Max Energy Capacity <= [FILL IN AFTER REG DROP] | Our maximum energy capacity is limited by ASC2026 regulations. We want our pack configuration to allow our energy capacity to be as close to this number as possible (if not exactly on this number). [1] |
Max Voltage < 165V | Constraint based on motor controller’s max cont. bus voltage [2]. Higher voltage is always better because we want to minimize current to reduce power loss from high current (P=I^2R). The max voltage is more-so a constraint on our series count. It’ll stay at 36 (37 is a prime number, 38 has 1, 2, 19, 38 as factors, so neither are very good for us when splitting the pack into modules) |
Minimize Weight | The battery is a big contributor to the car’s weight. The lighter it is, the more efficient our car will run. Additionally, a lighter pack is easier to remove, making serviceability and overall quality of life better. Minimizing weight relates to the cell’s gravimetric energy density and size (larger sizes are often more energy dense than smaller size, though this is not at all always the case). |
Minimize Volume | The battery takes up a considerable amount of space. As we aim for a slimmer car with MS16, minimizing volume is something to keep in mind. Minimizing volume relates to the cell’s volumetric energy density and size (larger sizes often have higher mAh capacity and so will allow us to use less cells for the same amount of energy → less volume taken up). |
Minimize Heat Generated | One area of improvement (or deletion) for the MS16 battery system is the removal of forced air cooling using fans. Minimizing heat generated from the pack may allow us to either move to passive cooling (cooling with no fans) or have no cooling at all. This can allow us to simplify design and also have less parts. Minimizing heat generated relates to the cell’s DC internal resistance and mAh capacity (because lower cell capacity means higher parallel count without decreasing series count → current draw is split across more cells → P=I^2R, less heat generated). |
Cell Selection Criteria
There are many criteria to select our cell from, I believe the ones shown below are the most important.
Specification | Why it’s Important |
---|---|
Manufacturer | We’re going to be buying hundreds of cells. In keeping with our goal of making MS16’s battery pack reliable, it would serve us better to buy cells from manufacturers that are already well-known (LG, Samsung, Molicel, Panasonic, Sony, Sanyo). |
Size (eg. 18650, 21700, 26500) | The physical size of the battery determines the volume of the pack. Larger sizes also correlate positively with higher gravimetric and volumetric energy density (not always true). For MS16 we will be sticking to cylindrical cells. |
Chemistry | Different chemistries have different properties. For MS16 we will be sticking to Li-Ion NMC/NCA cells. Good reading for common Li-Ion chemistries: BU-205: Types of Lithium-ion |
Gravimetric (Wh/kg)/Volumetric Energy Density (Wh/L) | Gravimetric energy density shows how much energy a cell can hold per unit mass (usually Wh/kg). Volumetric energy density shows how much energy a cell can hold per unit volume (usually Wh/L). We generally want to maximize this or at least keep it reasonably high such that it doesn’t affect the car’s performance (too heavy) or design (batteries take too much space) to any large extent. |
Capacity (Ah) | There isn’t much wiggle room for the pack’s series count due to the max voltage constraint given by the motor controller, we’re also keeping the motor the same. The parallel count will depend on the cell capacity. Note: Most manufacturers will give their nominal capacity at a 0.2C rating, which is standard. It would be ideal to also look at the cell’s capacity at higher discharge currents/C-ratings and see how close it remains to the nominal capacity. |
DCIR/ACIR (mOhms) | DCIR and ACIR technically measure different things, DCIR measures internal resistance and ACIR measures internal impedance [3]. DCIR is more useful to us since it’s more reflective of the actual use case of the cell under DC load, but ACIR is generally more standardized and is often what you’ll see in datasheets [4]. The two values do correlate though, so if there aren’t any DCIR values of the cell available from the datasheet or the internet, the ACIR still gives a good ballpark idea. [5] DCIR matters to us because the higher the internal resistance the more power will be lost to heat. This means more energy in the cell is wasted and the cell will get hotter easily. |
Max Charge/Discharge (A) | Max charge isn’t incredibly important because the only charging source for the pack is from the solar array, and it won’t be supplying very much current. [6] Max discharge current will be what determines our cell’s max power output capability. The max power for our motor is 5000W [7], but we’d benefit more from power consumption vs. speed driving test data from MS15 [I’ll include it here after we get that data]. |
Currently we are still waiting for ASC2026 regs to drop so that we can see what our max energy capacity is. It seems likely to me that the deciding factor between the top candidates will boil down to:
We can use the cell to configure the pack as close as possible to the reg’s max energy capacity. Because of the reg change, MS16 battery pack will have less energy than MS15 battery pack (likely by around 1kWh) [8]. If we’re going to have less energy this car, we want to have as much energy as we can still get.
The cell will strike a balance between lower capacity (which will lead to more total cells and a higher parallel count → less load for each individual cell → better thermals) and a “reasonable” weight/volume.
Relative to aerodynamic drag, weight has a small effect on the efficiency of the car (rolling resistance accounts for 14% of spent energy while aerodynamic drag accounts for 77% based on Forest’s calculation [9].
Cell Selection Organization + Timeline
We’ll be using Jira for frequent updates on progress and Discord #bat-cell channel for discussions. The timeline is shown on Jira as well, but I expect we should have a cell selected (based on the datasheet) no later than a week after ASC2026 regs drop (first week of November). Afterwards, we will acquire some samples from distributors to check that we’re not dropping >$1000 on knock-off cells that don’t match the datasheet.
To get started on finding cells, use this: Cell Sourcing MS15 Confluence page on cell sourcing and start looking around distributors for what kind of cells are available and read their datasheets.
If you’ve found a cell that looks good, add it to this Excel sheet: MS16 Cell Selection Candidates.xlsx
Resources
[1] Insert ASC2026 regs here after it drops
[2] DC Bus - Prohelion Documentation
[4] Measuring DCIR of Lithium-Ion Cells
[5] Internal Resistance: DCIR and ACIR - Battery Design
[9] https://docs.google.com/presentation/d/1C4GlFdyFLaSJIhRW-QnGZWU_KHUQxUT-ykzU_RGqzS4/edit?usp=sharing