TMC is an independent, primarily volunteer organization that relies on ad revenue to cover its operating costs. Please consider whitelisting TMC on your ad blocker or making a Paypal contribution here: paypal.me/SupportTMC

Air cooled battery modules

Discussion in 'Model 3' started by Swampgator, Jun 20, 2016.

  1. Swampgator

    Swampgator Member

    Joined:
    Apr 27, 2016
    Messages:
    65
    Location:
    Florida
    Randy Carlson (seeking alpha) is convinced this is the tech Tesla will be using for Model 3.
    His recent article points out several advantages to doing this, and the main one seems to be increased speed of battery balancing (initial) and drastically reduced cost.

    If you have not read his articles, I can't recommend them highly enough. Most of the stuff Goolle feeds me is click bait, but this guy in an engineer and writes his articles based on sound engineering analysis of the underlying technology.
    He was almost 100% correct a year ago on M3 specs.

    http://seekingalpha.com/article/3983102-teslas-gigafactory-christmas-july
     
    • Informative x 2
  2. JeffK

    JeffK Active Member

    Joined:
    Apr 27, 2016
    Messages:
    2,211
    Location:
    Indianapolis
    I can see zero advantages of using air cooled batteries in an EV.

    Which costs more liquid cooling or replacing batteries?
     
    • Like x 2
  3. Jayc

    Jayc Member

    Joined:
    Mar 28, 2016
    Messages:
    295
    Location:
    UK
    Would second @JeffK. I can't see Tesla diverting away from their current recipe especially for Model 3 when they would want to minimise R&D costs by re-use of proven technology further optimised for cheaper & more efficient production.
     
    • Like x 1
  4. Swampgator

    Swampgator Member

    Joined:
    Apr 27, 2016
    Messages:
    65
    Location:
    Florida
    I'm just curious if either of you read the article?
    If so, other than your general assertion of opinion, what specifically do you disagree with? That air cooling will NOT reduce costs? That the new load balancing software systems cannot combine with the new architecture to make building packs faster and cheaper? Do you agree that the packs shown on the launch video are not consistent with Tesla current liquid cooled methodology?

    I'm not trying to argue, I really was hoping to start a more robust discussion of the possibilities.
    I'm not an engineer so hopefully some on TMC are and can add their thoughts.
     
    • Helpful x 1
  5. juanmedina

    juanmedina Member

    Joined:
    Mar 31, 2016
    Messages:
    67
    Location:
    sc
    Never going to happen....

    End thread/
     
    • Like x 2
    • Dislike x 2
  6. Vitold

    Vitold Member

    Joined:
    Aug 10, 2015
    Messages:
    987
    Location:
    NM
    Air cooling is too energy intense to be practical since air has 4x less heat capacity than water and is a bad conductor of heat. Air would have to be circulated very fast through radiators to approach effectiveness of liquid cooled solution.

    If Tesla went with air cooled battery it would have to have different chemistry to be practical and perform as good (battery fade) as current Tesla battery pack.
     
  7. dandurston

    dandurston Member

    Joined:
    Jul 16, 2015
    Messages:
    192
    Location:
    Victoria, BC, Canada
    I read the article (or at least a lot of it)

    He might be right about some of the cost savings but most of these have to do with balancing the cells in the packs instead of in a separate step. It appears this could also be done with liquid cooled batteries and packs with a bad cell would just need to be recycled instead of fixed (as opposed to replacing individual cells). So I could be missing something but I don't see this as a big argument for air cooling.

    He's right that the size of the Model 3 battery does look unusually large. This could be due to air cooling but there are many other plausible design explanations such as structural considerations, more of the electronics are stuffed in here, or Tesla has plans for a really high capacity version (perhaps in a CUV using the same platform).

    Tesla will never do a poorly cooled pack (a la Nissan Leaf) but active air cooling is plausible if it can be done well. There is obvious weight savings from doing this, but other than that I don't see a compelling reason to do it. Then again, I don't really understand this stuff.
     
    • Like x 2
  8. Booga

    Booga Member

    Joined:
    Apr 21, 2016
    Messages:
    362
    Location:
    Ohio
    That was a great read - thanks for sharing. I can't say I have the knowledge to argue the author and I think he's got to be right. It gives us a big explanation for how cost savings will be achieved and it actually demystifies things tesla has said.
     
    • Like x 1
  9. Genshi

    Genshi ermagerd, lorst mah reservation!

    Joined:
    May 4, 2016
    Messages:
    83
    Location:
    Portland, Oregon
    I'm about halfway through the article - and even though I'm on Seeking Alpha, I totally missed this article somehow. Thank you very much for posting it.

    I'll give my thoughts on how plausible it all seems after I finish it and have time to reflect on it, but at first glance I have to say this is an impressively well thought out and researched argument. And it would certainly help to explain WHY everyone in the industry is so dubious about Tesla's claims for their battery costs, and how they could be so incredibly wrong about the true cost, if they are. It also jibes with some other articles I've read arguing that the cost of batteries can't come down until such time as we find a better way to cool them, because liquid cooling brings so many costs of its own.

    That said, and with the option to revise my opinion later, I'm with dandurston on one thing. There could be other explanations of why the Model 3 battery looks unusually large, but we can be certain of one thing: Tesla won't botch the battery cooling, no matter what solution they implement. If anyone can come up with an active air cooled system, it's Tesla & Panasonic.
     
    • Like x 4
  10. jackbowers

    jackbowers Jack Bowers

    Joined:
    Aug 23, 2009
    Messages:
    116
    I agree with everything the author says except for the presumed advantage of single-cell replacement. I just can't see where that's a necessary or practical need in a production process where robots are placing 5+ million cells a day. Tesla's liquid cooling system only adds a few hundred pounds to the vehicle's weight. To replace it with an equally effective air-cooling and air heating system would yield negligible weight savings and might dramatically alter the look of the vehicle in favor of massive air flow, as Faraday has done.
     
    • Like x 1
  11. 22522

    22522 Member

    Joined:
    Jun 6, 2016
    Messages:
    327
    Location:
    Texas
    Thank you for posting the link. There is a lot of good information in there. Especially the cell balancing. Some of the arguments about replacing cells and then not needing to replace cells seemed, maybe circular. But I think I learned something.

    As far as cooling goes, air is not how I would do it - mostly for super charging and time constant reasons.

    If concerned about the electrical conductivity of the coolant, use phase change wax instead of air. Flow the coolant on one side of a pan with accordion pleats that show up between each row of cells. Or run radiator tubes if you like. The main point being that charging speed matters. And charging speed is limited by cell temperature rise during charging. If you immerse the cells in a phase change wax, the cells will stay cool during charging and the cooling system can be designed for a much lower (normal driving and braking) rate of cooling, and thus be less expensive. Wax is pretty light.

    Phase change wax makes the whole system more forgiving and keeps the batteries in the narrow range of temperatures that they prefer. Maybe it could be added as a last step.
     
  12. JeffK

    JeffK Active Member

    Joined:
    Apr 27, 2016
    Messages:
    2,211
    Location:
    Indianapolis
    It's never going to happen for a few simple reasons:

    The volume of the 20700 cell is larger and it's therefore harder to cool then current 18650 cells which already use liquid cooling.

    If you live in a hot and dry environment air holds even less heat and it'd be difficult to cool the battery no matter how much air you're forcing over the battery. Forget supercharging at a decent range.

    The biggest enemy of Lithium Ion batteries is heat, why in the world would anyone choose to use a less efficient cooling mechanism, especially if they plan on giving a warranty on the battery.

    Make no mistake cooling will be more efficient on the Model 3 than it is on the Model S.
     
    • Like x 3
  13. diamond.g

    diamond.g Member

    Joined:
    Nov 5, 2015
    Messages:
    522
    Location:
    Fredericksburg, Virginia
    Yeah I immediately thought, how do you cool these while supercharging. The water pump/fan is already quite loud (admittedly from Youtube videos) while supercharging to dissipate the heat. Could low pressure R134 work? Though that still seems pretty expensive...
     
  14. stopcrazypp

    stopcrazypp Well-Known Member

    Joined:
    Dec 8, 2007
    Messages:
    7,038
    #14 stopcrazypp, Jun 20, 2016
    Last edited: Jun 20, 2016
    I remember someone bringing up the air cooled theory in another thread and the counter-intuitive suggestion that a larger volume pack means air cooled. Usually it is the opposite, air cooled packs tend to be smaller because they don't have any coolant paths between cells, but rather rely on heat exchangers between closely packed cells to remove heat.

    Just for reference, here is the Leaf pack with no active air cooling:
    http://c1cleantechnicacom.wpengine.netdna-cdn.com/files/2016/01/Nissan-LEAF-battery-pack.jpg
    Here is the e-NV200 pack with active air cooling (main difference being the addition of a radiator and fan, plus coolant hoses):
    Does Nissan e-NV200 Show Change of Policy on Battery Heating, Cooling?

    To go back to the article, I think the summary is:
    1) Larger cell spacing suggests air cooled battery for Model 3
    2) Air cooled battery is cheaper
    3) Air cooled battery allows for individual cells to more easily be replaced (based on his theoretical buss bar design)
    4) Instead of doing cell match/aging/testing, he suggests making the modules directly, having a dynamic balancing system make up for cell variations and also do the testing in-module. This will save money on separate cell matching/aging/testing equipment.
    5) When a defective cell is found in the built module, it can be individually removed and replaced.

    There are several issues I found:
    1) There are other simpler explanations for the larger cell space (which may not necessarily be true in the first place): it could simply be Tesla allowing more room to increase capacity of the pack in the future.
    2) While a densely packed passive air-cooled pack like the Leaf's is cheaper than a liquid cooled one, it is unclear a loosely packed active air-cooled pack (esp. one with his special buss bar design and dynamic balance circuit) is cheaper. While there won't be coolant hoses etc in the pack, the pack enclosure overhead is larger (need larger enclosure for given kWh) and you need fan(s) and heat exchanger(s) in the pack.
    3) Tesla can design a liquid cooled pack with a similar buss bar design that allows easier individual cell replacement, so it does not follow that this feature is a result of an air cooled design.
    4) Dynamic (aka active) balancing systems are significantly more expensive/complex because the circuit has to shuttle charge between cell groups (rather than just dissipating it using resistors). Basically you are saving a one-time cap-ex factory equipment cost for cell testing equipment and offloading it into per-battery equipment. I'm pretty certain that the costs increases there per pack would actually be more.
    Also, the system he suggests (which is at the granularity of a cell group, not individual cells) is unable to find individual cell failures. That means Tesla would potentially be shipping packs with defective cells in them. Even if Tesla is willing to do so, I'm not sure Panasonic would be willing to do so. A pack with a defective cell that underwent thermal runaway (something that had so far never happened in an EV) would not be good PR, and his suggestion would drastically increase the risk of this.
    The other alternative of adding per-cell circuits would be ridiculously expensive (for obvious reasons, since you basically would be duplicating the same per cell testing equipment except shipping it with the car each time instead of keeping it in the factory).
    5) Suggested cell replacement would also have a cost (regardless of it being made easier) vs just having cell QC take care of it in the first place.
     
    • Like x 2
  15. JeffK

    JeffK Active Member

    Joined:
    Apr 27, 2016
    Messages:
    2,211
    Location:
    Indianapolis
    I can't imagine a labor bill at a service center if someone were to suggest replacing a single cell by taking apart an entire battery, then taking apart the affected module, then testing every cell in the group affected, replacing a single $2 defective cell, and putting it all back together again...

    I'd rather just get a new battery under warranty and Tesla can take the old battery do whatever they want with it on their own time. Dissect it, test it, turn cells into a powerwall, whatever. Tesla has shown they can swap an entire battery in just a few minutes.
     
    • Like x 1
  16. Booga

    Booga Member

    Joined:
    Apr 21, 2016
    Messages:
    362
    Location:
    Ohio
    I think the idea would be that they would provide more cells than needed for the battery pack you offered to account for any that could be defective. It's a question of whether it's cheaper to test each one or just provide 10% extra and let the battery circuitry figure it out and adjust power specs before it gets to the customer. If you put extra cells, it won't matter it 5% of them are bad for example.
     
    • Like x 1
  17. Cosmacelf

    Cosmacelf Active Member

    Joined:
    Mar 6, 2013
    Messages:
    3,399
    Location:
    San Diego
    First, the article author's battery cooling system has nothing in common with how the Leaf's air cooled system works. Unlike the Leaf, he proposed a sealed environment where air (or more likely a special gas) circulates at high speed and is chilled via an air conditioner heat exchanger. I have no doubt this could be as good if not better than our current liquid cooled pack (air, an insulator, can go places that liquid, a conductor, cannot).

    Second, the major innovation he proposed in how the gigafactory and battery pack works isn't tied to an air cooled pack. He could be wrong on this point. Regardless, his idea of not needing the huge equipment equipment required for sorting and initial powering of the cells is very interesting. That innovation alone could save 15% of battery packs costs.

    Third, that such a set up could ship cars with shorted cells isn't a problem (assuming they didn't in fact do cell replacement) since the cells would be electrically isolated due to the per cell fuse (thin wire).

    Fourth, anyone know the failure rate of our current battery packs due to the liquid system failing?

    Fascinating idea. Rather than dismiss it out of hand (and please read the article before commenting people, it is more complicated that this thread title would suggest), it would be more interesting to try to find corroborating evidence (or counter evidence).
     
    • Helpful x 1
    • Like x 1
  18. Booga

    Booga Member

    Joined:
    Apr 21, 2016
    Messages:
    362
    Location:
    Ohio
    The other thing I wonder about is that Elon said ludicrous mode will be an option on the car. Surely, that will generate more heat. If they can design it for ludicrous mode, then my normal driving won't be an issue, but it does still make me wonder just how they've figured out air cooling to work so well, and in a multitude of climates.
     
  19. JeffK

    JeffK Active Member

    Joined:
    Apr 27, 2016
    Messages:
    2,211
    Location:
    Indianapolis
    A liquid can typically carry more heat than a gas... And with the current setup you're moving heat between the motor, battery, and HVAC. What the author is suggesting adds an unnecessary level of complexity.
     
    • Like x 2
  20. Frank Schwab

    Frank Schwab Member

    Joined:
    Apr 7, 2016
    Messages:
    71
    Location:
    Arizona
    I dunno. He makes some strong mechanical arguments for an air-cooled pack (though I quibble with his serpentine air path through the pack), and there's no reason it wouldn't work.
    From a temperature-control standpoint, moving air doesn't transfer heat as well as moving water. But, a cooling system for air with the same cooling capacity as Model S cooling system has for liquid could easily keep the batteries cool, assuming sufficient airflow (his whole argument). If you make the air cooling system reasonably closed loop (i.e. doesn't have to be airtight, but should keep 95% of the air inside the battery packs), you essentially have a refrigerator, but one designed to move a hundred times more joules of heat than your refrigerator.
    This should make the car cheaper, with no reduction of battery life - you're still effectively cooling the batteries, even in hot weather (and it's currently 113F outside right now).

    His whole argument about cell balancing is vastly overblown. Each cell group (batteries in parallel) will stay explicitly balanced - an out of balance cell will immediately be balanced by the other cells in the group. To balance between cell groups, you only need to be able to inject current between cell groups - every RC hobbyist's Lithium Polymer battery charger already has a "balance" port designed to do precisely this. For the Model S pack (96 cell groups), that means that you need 96 connections to the on-board battery charger, which needs the ability to monitor the voltage on all 96 of those connections. If each battery module is 24 V, that implies 6 cell groups in each module; you'd only need a 6 pin connector coming out of each battery module to do this. The electical hardware to do this is basically a $5 microprocessor to monitor voltages, and a $1 FET on each of the 96 connection points to switch in charging voltage.
    Cell balancing doesn't need to be done on every charge cycle. He was unnecessarily worried about Supercharging an unbalanced pack. Based on my hobby experience, you could go dozens of partial charges with no significant unbalance occurring; if the battery charger noticed a significant unbalance (say, you only charge at superchargers, never let the charge complete, and do so for months), it's easy for the supercharger to balance the pack BEFORE charging it - you simply use the 96 connections to each cell group to individually charge cell groups to the same voltage, THEN start charging the entire pack at full speed. This might add 10 minutes to your charge time in this corner case.

    I'd give the air-cooled pack a better than even chance of being correct.
     
    • Informative x 1
    • Like x 1

Share This Page