Our thermal electrical engineering consultant* updates us on his latest thoughts on the 4680 pack design.
Please note that Tesla has not told us yet how they plan to cool the new 4680 pack.
In our last article (ref1) we outlined what we thought the 4680 pack would look like. Specifically, how it would be cooled and how the cells would be electrically connected. Our thoughts are still that Tesla will switch from its their cooling snake design that it has used since day one to a flat plate cooled system.
The tabless design is the perfect candidate for flat plate cooling as it provides an excellent thermal connection between the inside of the cell and the proposed cooling plate. Seventy percent of the heat goes out that end of the cell. The old cooling snake design pulls the heat out the sides of the cell and it’s ineffective to pull the heat out the sides with these new large diameter cells.
We suggested the 4680 pack would incorporate both a top and bottom cooling plate and the cell parallel groups would flip polarity up and down and pass the current in a similar manner as the Tesla Model S and Model X. We now think there will only be one cooling plate on top of the pack.
Based on some photos from Tesla (ref2) and Sandy Munro’s presentation (ref3), we have changed our thinking. We now believe that Tesla will orient all the cells with positive nipple end down and negative smooth side up (just like Model 3 (ref4). The one (revised from 2) cooling plate then logically is on the top smooth negative end of the cell since the copper electrodes in the tabless 4680 design are connected to the negative end of the cell and that is where 70% of the heat is conducted.
We speculate all the electrical connections being made on the bottom of the pack just like the Model 3 and Model Y packs (ref5), with the top cooling plate bonded directly to the cells, as well as the (speculated) current collector board, which is on the bottom being bonded to the cells. The current collector board also being bonded directly to the vehicle’s cast aluminum body/belly pan thus providing excellent structural rigidity as Elon Musk explained in the Battery Day presentation.
As explained, we now speculate Tesla will orient all the cells in the same direction-negative smooth side up. We base that on some photos Tesla released showing the smooth side up and also a Tesla technician installing a cell into the pack.
Note that the negative smooth side still appears to have a subtle, slightly raised circular section in the center. This is different from the well-defined nipple positive connection at the other end that we are all used to seeing. We propose that a subtle raised circular section on the negative side could be used to position the cells on the flat cooling plate during assembly. The pack would be assembled upside down then flipped before going into the vehicle.
Tesla photo showing the 4680 cells with negative smooth side up
Tesla photo showing the 4680 cells with negative smooth side up
Tesla photo showing a cell being installed
This negative smooth side makes a great thermal connection with the copper electrodes inside the cell, and based on our previous analysis, 70% of the heat is conducted out this end of the cell ... thus the most logical place to put the cooling plate (on top).
Electrical connections to each cell
In our last analysis, we suggested that Tesla might make the electrical connection between cell groups in a manner similar to Model S and Model Y ... ie the parallel cell groups flip polarity up and down. This makes it easy to pass the current from one cell group to the next since you simply make a short hop to the next cell group connecting the negative end of one group to the positive end of the next.
However, we now think Tesla will orient all the cells positive nipple side down and that the connections will be very similar to the way Sandy Munro proposed in his video (ref6), with one minor difference.
Sandy proposed a multilayered current collector similar to an ICE head gasket as shown in the photos below.
Photo of how Sandy Munro thinks the current will be connected between parallel cell groups.
Photo of Sandy Munro’s multi-layered current collector concept
We propose a slightly different approach. Instead of the multi-layered “head gasket” design, we suggest a simpler single-layer “printed circuit board” type design.
Sandy's concept had the current collected from each group following a long zig-zaggy path across each cell group's cells' positive terminals before jumping to the next group at a single point. We think Tesla will make a short hop from the + of each cell to the adjacent cell's - connection in the next cell group, with a shared current collector "bar" connecting all the +'s of each group together to maintain a single group voltage.
This could all be fabricated on one giant sheet, with holes at the appropriate points for the spot-welders to access the current collector/cell connection locations. The holes would get filled with the structural epoxy adhesive that adheres the cells/collector sheet to the structural pack bottom plate. The epoxy would electrically-insulate the exposed weld tabs from the plate, just as the Model 3 module current collectors were isolated from the pack bottom via foam or epoxy embedment.
Proposed 4680 current collector design single parallel group
Proposed 4680 current collector design showing connections between parallel groups
Proposed 4680 pack design showing top flat plate cooling plate and bottom printed circuit board type current collectors.
Proposed assembly method for the 4680 pack
(pack is assembled upside down):
- Cell "bottoms" are dropped into the pack cooling plate (with either a thermally conductive insulation grid first installed over the cooling plate or individual cell insulation "doughnuts" first applied to each cell over the cell bottom buttons), with the indexing buttons fitting into the plate indents.
- A pack-sized jig aligns the cell tops to maintain the approx 1 mm can separations.
- Structural electrical insulating adhesive is injected into all gaps between all cells to bind the cells into a unified pack.
- Jig removed and pack current collector printed circuit board plate is laid on to the cell "tops". Pack is assembled upside down. Cell tops are the nipple positive end of the cell.
- Welders tack all connectors to the cell + and - cans.
- Final top layer of structural epoxy applied to connect cell tops to connector plate and seal/insulate all connector plate holes providing welder access to tack points.
- BMS connections, final power terminals, etc. installed, if not already incorporated into the "printed circuit board" current conductor plate.
- Assembled pack flipped over to insert into the cast body bottom panel.
*Credit: Keith Ritter PE