New Pack Architecture that is Likely Used in P100D and New TE Products

[vgrinshpun]

It looks like the new battery architecture that was mentioned by Elon and JB used in P100D pack is outlined in US Patent Application Publication US 2015/0244036 A1 titled ENERGY STORAGE SYSTEM WITH HEAT PIPE THERMAL MANAGEMENT.

I made an attempt at comprehensive search of all Tesla patents, but could not find any other that would be consistent with what we know about the P100D:

• Significantly increased speed of supercharging. Both improved cooling due to use of heat pipes and reduced heat generated by the cell interconnection using flexible printed circuit
• Elimination of cooling ducts between the rows of the cells which allows for more cells within the same size pack. This is consistent with the new heat pipe based cooling design.
• Improved performance of the P100D could be attributed not only to improved cooling, but perhaps in greater degree to the reduced resistance of the battery call interconnections due to use of the printed circuit technology

Salient points of the Patent:

1. Heat Pipe technology originates from the aerospace industry. The technology based on the application of the hollow pipe (not necessarily round in cross section and not necessarily having just one hollow channel) which contains phase change fluid. At the evaporative end (hot end) of the heat pipe fluid changes state from liquid to gas. At the condensation end (cold end) of the heat pipe fluid changes state from gas to liquid. The circulation of the fluid within the heat pipe is occurring due to pressure differential between the gas and liquid states and therefore does not require any external power, has extremely high reliability, and very robust heat transfer properties – all contributing to original use in the aerospace industry. Video illustrating technology. Another video
   
   
2. The advantages of heat pipe based cooling system for automotive and stationary energy storage systems are listed in the patent as follows:

• Elimination of the internal fluid connections, thereby avoiding leakage
• Closed loop cooling system reduces pressure drop losses with regard to an overall cooling system
• Elimination of the external cooling tube assemblies
• Rapid fluid migration can be provided that keeps cells at even temperatures
• Cooling tube sections between rows of cells can be eliminated, thereby allowing more cells to be packed into a given space
• Even if rapture occurs in one of heat pipe lumens, significant cooling/heating can nevertheless be provided by way of other undamaged lumens within the heat pipe

3. The arrangement of cooling pipes at the bottom of the cells requires both positive and negative terminals of the cells to be located at the top of the cells.


4. Interconnections between the cells using printed circuit technology not only allows for efficient highly automated manufacturing, but also ostensibly allows for a reduction of the resistance associated with the interconnections, improving voltage profile at high discharge rate, allowing for increase of power without increase of the maximum current – IMO secret sauce allowing improvement of the performance in P100D and improvement in the speed of supercharging.

This is truly a breakthrough stuff. Extremely impressive. Especially considering that the application was filed in February of 2014, and the detailed design, manufacturing design, testing, tooling, work with suppliers, and final production of the new battery pack took only 22 month. I am wondering how many other innovations Tesla is working on, unbeknown to casual observers.

Here are couple of illustrations from the Patent with my annotation:

 

[Waiting4M3]

Even if rapture occurs in one of heat pipe lumens

Thanks for the info. I'm looking forward to my Tesla M3 rapture

 

[sandpiper]

Excellent post! I've been very curious as to what what the differences were.

I wonder if they modified the P100D cells so that they have both terminals at the top, or if they came up with some sort of hybrid, using the current 2-end cells? Neither of these would seem to be an easy thing to do.

If they're modifying the cells, wouldn't that mean a fairly significant change to the large 18650 cell plant(s) in Japan to support the 100D, and perhaps versions of the smaller pack using the same module? And I wonder if they would actually do that, given that the 18650 cells are going to drop out of production within a year or so?

 

[scaesare]

I had the same question about the 18650's... it almost feels odd to change them at this point given how close we (assume) we are to the 2170's coming online.

I wonder if this implies either: the P100D's have secretly had 2170's all along, or that Tesla expects 18650's to be used for the foreseeable future in some applications.

Is it wrong to hope somebody wrecks one so we can get a salvage teardown soon?

 

[AudubonB]

Excellent input - thanks. Will even permit the religious reference to stand (I'd noticed it too!).

I'm moderately surprised that a separate patent or patent swarm pertaining to a double-terminaled-top 3.7-4.2V cell also hasn't surfaced. Even though that has been standard configuration for decades for, e.g., 9V cells, there must be significant internal changes that both encompass this AND make efficient the "bottom-centric" cooling this supports - and demands.<===See? I'm now deferring to this.

 

[Stirthepot]

Don't want to see someone wreck one, but if a tree or other large object falls on a parked empty p100d and totals it that would work.

 

[JRP3]

Interconnections between the cells using printed circuit technology not only allows for efficient highly automated manufacturing, but also ostensibly allows for a reduction of the resistance associated with the interconnections, improving voltage profile at high discharge rate, allowing for increase of power without increase of the maximum current – IMO secret sauce allowing improvement of the performance in P100D and improvement in the speed of supercharging.

I'm not understanding how this would work vs the current cell interconnects. Currently they are welded wires attached between the cells and the cell bus bars, with the wires acting as fuses. How does using printed circuits create lower resistance?

 

[jbcarioca]

Now, with this very exciting news, What developments there have been in increasing silicon content in the anodes? Have there been other patents related to manufacturing process?

This strongly implies substantial increase of energy density at pack level, because of the weight/complexity reductions implicit in both the passive cooling and printed circuit connectors. How much might that mean?

So, we suspect much of this (all of it?) may have been employed with the 100 kWh battery pack, but, it seems quite logical that the energy density of the entire installed pack would improve by another substantial margin by using the 2170 format with consequent tighter cell packing and reduced total number of connections.

Assuming the last sentence logic is valid we might have the 100 kWh pack as a production-level proof of concept, with the larger gains expected with 2170's.

Am I reading this correctly?

 

[scaesare]

Don't want to see someone wreck one, but if a tree or other large object falls on a parked empty p100d and totals it that would work.

Works for me... as long as we aren't looking at bodily injury...

 

[dhcp]

Excellent write-up vgrinshpun, thank you.

 

[scaesare]

Excellent input - thanks. Will even permit the religious reference to stand (I'd noticed it too!).

I'm moderately surprised that a separate patent or patent swarm pertaining to a double-terminaled-top 3.7-4.2V cell also hasn't surfaced. Even though that has been standard configuration for decades for, e.g., 9V cells, there must be significant internal changes that both encompass this AND make efficient the "bottom-centric" cooling this supports - and demands.<===See? I'm now deferring to this.

Hmm, I had thought that top-mounted double terminals were already a thing. My Streamlight battery pack works this way, but now that I look at them a bit more closely, it appears that it's a pack implementation, and not a cell implementation:

So perhaps Tesla is the first to be making 18650 and/or 2170 form factor cells with double top terminals...

 

[MitchJi]

So, we suspect much of this (all of it?) may have been employed with the 100 kWh battery pack, but, it seems quite logical that the energy density of the entire installed pack would improve by another substantial margin by using the 2170 format with consequent tighter cell packing and reduced total number of connections.

Assuming the last sentence logic is valid we might have the 100 kWh pack as a production-level proof of concept, with the larger gains expected with 2170's.

The gains due to using 2170's are the extra length (about 8%), and cell chemistry changes, probably another 8-10%. A total of about 15%.

 

[vgrinshpun]

Excellent post! I've been very curious as to what what the differences were.

I wonder if they modified the P100D cells so that they have both terminals at the top, or if they came up with some sort of hybrid, using the current 2-end cells? Neither of these would seem to be an easy thing to do.

If they're modifying the cells, wouldn't that mean a fairly significant change to the large 18650 cell plant(s) in Japan to support the 100D, and perhaps versions of the smaller pack using the same module? And I wonder if they would actually do that, given that the 18650 cells are going to drop out of production within a year or so?

I was having my doubts on the change required to be made for 18650 to make the bottom cooling work, and was able to convince myself that it does make sense. Here are the contributing thoughts/facts.

First, I had an exchange with Jeff Evanson, Tesla IR VP on the subject back in August, and I think that it is notable that he did not use "same cell" in his response, rather "essentially the same cell.

Here is the exchange:

Hello Mr. Evanson,

There is a lot of discussions in Tesla investors and enthusiasts community on whether larger capacity of the P100D battery pack is achieved using additional cells, or using optimized BMS and cooling to safety extract more capacity from the same quantity of cells as in P90D pack.

Could you clarify this?

As a side note, this seem as a purely technical/engineering question, but if advanced design of the battery pack allowed Tesla achieve higher battery capacity with the same quantity of the same cells, it would indicate a very important technological break through, very important as a differentiator among the potential competitors.

Thank you in advance for you response."


"Elon and JB did a call with some media yesterday, and it was a good discussion, but this specific topic was not covered.

They did say its essentially the same cell, packed more closely together, but also in the same case size and connection points, so its possible some might infer more cells, but we were never asked that specific question directly.

In addition there were numerous hardware and software changes to the entire system.


Jeff Evanson
Tesla IR"

Second I think that focusing on how long Tesla will be able to use modified 18650 with top electrodes is wrong way to look at this. They need to continue to increase MS and MX sales before the GF has capacity to ramp up production of 2170, and getting there requires cash. So in this respect it is irrelevant that 18650 will be phased out in one year or so - modification of the terminals is necessary for supporting increased sales.

Third, this is the pack architecture that will most likely be used in M3 with 2170 cells, so flashing out any potential problems or possible improvements to the manufacturing process is better achieved on lower volume platform (MS/MX) than on mass produced M3.

An finally forth, the new architecture requires new suppliers, bot for heat pipes and flexible printed circuit. Once again it is beneficial to work out these new relationships on a lower volume platform before the mass produced M3.

Yet another useful, I think, angle is an example with Alien Dreadnaught 0.5 starting to produce M3, and then version 1.0 arriving year or so later. So if we have similar set of mind, logical question would be why not wait and do the AD 1.0 to begin with? I did not ever see this question asked, and it does not seem logical to ask it knowing what we know about company's plans. I do not think that question about expending resources for modifications to 18650 not to long before their phase-out should be any different.

The bottom line to me, that all of this boils down to the "velocity of innovation". Whatever seem to be less logical for the "normal" velocity, over a sudden becomes much more logical once velocity becomes supercharged, as it is with Tesla.

 

[JRP3]

The gains due to using 2170's are the extra length

And diameter.

 

[vgrinshpun]

I'm not understanding how this would work vs the current cell interconnects. Currently they are welded wires attached between the cells and the cell bus bars, with the wires acting as fuses. How does using printed circuits create lower resistance?

The evidence of this is circumstantial, but I think that the secret of the lower interconnection resistance is due to welding of the terminals to the conductive layer (plate) in the flexible printed circuit, rather than spot welding individual wires to the terminals and to the current collecting busses inside the pack.

I was also wondering how they got rid of the individual wires if they were providing the individual cell protection. My assumption that this role is now played by spot welds between the conducting layer of the printed circuit and cell terminals.

 

[vgrinshpun]

BTW, this photo from the GF opening event seem to indicate that 2170 have both terminals on the top of the cell. This would be an indication that new 2170 based pack architecture to be used for M3 and TE will indeed use bottom cooling via heat pipes as outlined in the subject Patent.

 

[stopcrazypp]

BTW, this photo from the GF opening event seem to indicate that 2170 have both terminals on the top of the cell. This would be an indication that new 2170 based pack architecture to be used for M3 and TE will indeed use bottom cooling via heat pipes as outlined in the subject Patent.

View attachment 205444

That looks similar to what a bare cell looks to me (below is an unwrapped Panasonic NCR18650B). I don't think that is necessarily the top terminal version (or at least a special version; you can somewhat use existing cells as "top terminal" given the body is already negative, while the middle button on top is positive).

 

[vgrinshpun]

That looks similar to what a bare cell looks to me (below is an unwrapped Panasonic NCR18650B). I don't think that is necessarily the top terminal version.

Yes, it would help to see the bottom of 2170 to have a final conlusion.

 

[AudubonB]

I was also wondering how they got rid of the individual wires if they were providing the individual cell protection. My assumption that this role is now played by spot welds between the conducting layer of the printed circuit and cell terminals.

My knowledge of this field of electronics is experiential only; not academic. With that, I would suggest that the "circuit breaker" function of wire welding (prior) or spot welding (as Vlad's assumption) is absolutely suboptimal* and could more elegantly, effectively and cheaply be done through pre-formed breakers - such as they are - for every single cell, within the flexible printed board.

*suboptimal is the polite word. Insert your own alternative

 

[jbcarioca]

The gains due to using 2170's are the extra length (about 8%), and cell chemistry changes, probably another 8-10%. A total of about 15%.

My question is about energy density per kg.
Historically the primary gains due to format changes were brought about by two things: 1) reduced weight of packaging since a higher proportion of total weight is made up of electrolyte and anode as the format increases in size; 2. Reduced weight and complexity of electrical connections and battery management components.

Gains from chemistry changes have brought gradual improvements also, but not so dramatic as many had hoped.

Physics has recently helped change both sides of the equation, especially nanotechnology which has acted to vastly increase the total area available for ion exchange. So far it seems that most such gains have not been economically producible, but they will be soon...

Then the other big leap is silicon anodes, which have not been generally practical due to silicon tendency to expand. Tesla did take the first production baby steps in this direction with the 90 kWh pack, and have been (I have been told by a 'usually reliable source') further developed for deployment in 2170's.

Probably we all agree that the topic of this thread, the new pack architecture, thanks to vgrinsphun, is a major advance in practicality and cost effectiveness, not to mention both charging and discharging velocities. Now we wait to find out all the other nice advances we'll soon hear about.

I hope I'm not too optimistic. I take Elon at his word when he says 100 kWh might be about the maximum. if weight, efficiency and cost drop significantly he can mean what he says while still having continuously improving performance and range.