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 and becoming a Supporting Member. For more info: Support TMC

Automobile Mag: Quick Drive: Tesla Model 3 Performance AWD

Discussion in 'Model 3' started by Funkmobile, Jul 24, 2018.

  1. Knightshade

    Knightshade Active Member

    Joined:
    Jul 31, 2017
    Messages:
    4,491
    Location:
    NC
    Is there any evidence the part number is different? (obviously the P gets one for binning purposes, but if there's no physical difference from RWD to AWD other than the software there'd be no different PN)
     
    • Disagree x 2
  2. spesler

    spesler Member

    Joined:
    Jul 6, 2018
    Messages:
    290
    Location:
    Atlanta, GA USA
    Someone on reddit commenting that they have access to https://epc.teslamotors.com/#/login and “confirmed” it. Hardly bulletproof, I know.
     
  3. mongo

    mongo Well-Known Member

    Joined:
    May 3, 2017
    Messages:
    9,158
    Location:
    Michigan
  4. Knightshade

    Knightshade Active Member

    Joined:
    Jul 31, 2017
    Messages:
    4,491
    Location:
    NC
    what always seemed weird about that (it's what started the whole 500 amp AWD rumor) is that it's from a RWD teardown 5 months ago gaining access to factory mode.

    So the only possible place the rear inverter could say it was limited to 500 amps that he would've seen it then is... in software.

    Not hardware.


    And yet people seem to keep suggesting it's evidence that the difference is hardware, not software.
     
    • Like x 1
  5. mongo

    mongo Well-Known Member

    Joined:
    May 3, 2017
    Messages:
    9,158
    Location:
    Michigan
    Please elaborate.
    All systems I have worked with that have variations due to component tolerances have calibrations in their software. No on board potentiometers or jumpers, just a limit value in software. Also worked on systems that traded off lifetime for performance. Exact same hardware.


    Where the real limit is, I don't know. For example, a nominal MOSFET based driver could be rated to 650 Amp. A 10% worse rds_on (20% power dissipation increase) makes the worse case 520 A (for same dissipation) so they might derate to 500 A for standard use. A 10% better value in rds_on would allow it to operate at 20% higher current or ~780A.

    Seems like they would pre-bin the switching devices also, especially to match switching times, rather than only doing a post assembly measurement.
     
  6. GSP

    GSP Member

    Joined:
    Dec 28, 2007
    Messages:
    2,463
    How does 187.5 kW battery power line up with 211 kW motor output power? Is the motor more than 100% efficient? :)

    Same for getting 335 kW output from 300 kW input. I must be misunderstanding what you mean in your post.

    GSP
     
  7. Knightshade

    Knightshade Active Member

    Joined:
    Jul 31, 2017
    Messages:
    4,491
    Location:
    NC
    This one is pretty simple. A lot of modern HW, especially HW controlled by SW, tells the software what it's capable of. The RAM in your PC tells the BIOS what speed it's designed to run at- so does your CPU, GPU, etc... so does your TV tell that to your AV receiver (or a PC if it's attached to that).

    So if you're going to bother to make a physically different inverter it's trivial to have it able to tell the computer what its capabilities are to insure the software knows the limits of it- having this info burned into an eprom of some kind on the board for example.

    As a bonus it could include a serial #, build date, or other identifying info that makes later diagnostics easier since you can get all that data from the computer instead of needing to physically remove and inspect the part.



    On the other hand- if you're NOT going to make physically different drive units at all- then the logical, in fact the only place you can make the AWD perform lower than the P, is by putting in software limits in the computer.

    And the only evidence so far given for any differences in AWD and P has been...

    Software limits in the computer (of a RWD car no less from 5 months ago).



    But Tesla has been consistently making 800 amp RWD units since last year. So we know for a fact they can reliably produce those by the tens of thousands.

    So it doesn't make much sense they'd magically now be so UNABLE to get reliable yields that high that they suddenly need to stamp 500A max on a bunch of them going into AWD cars

    Elon specifically said it was DUs, the full assembly, being binned... apart from which the earlier and more often in the process you do it the more time, cost, and complexity you add to both manufacturing (as you're adding steps, sorting, logistics of differentiated parts, etc) and supply chain (as you're creating more part numbers to track, have where they need to be in sufficient numbers, have to ship as replacements, etc)


    24 SiC FETs BTW, according to a teardown of a RWD Model 3
     
    • Informative x 1
  8. mongo

    mongo Well-Known Member

    Joined:
    May 3, 2017
    Messages:
    9,158
    Location:
    Michigan
    Yeah, my pre-coffee mobile posts are not so understandable...:oops:

    As in, between the two numbers when derived from assumptions and internet info, they line up better in this configuration, that to say one motor pulls 800 amps on its own. 30kW (187<->211, 300<->335) off is much closer than 89kW and148kW off (300<->211, 187<->335).
     
    • Helpful x 1
  9. mongo

    mongo Well-Known Member

    Joined:
    May 3, 2017
    Messages:
    9,158
    Location:
    Michigan
    By the same point, it is trivial to test a drive unit to see what its limit really is and then program that value into the drive unit. Physically the same or different, they each get unique traceability data. The drive unit is a full embedded microprocessor system with all the limits and SN and PN and such stored internally.

    Are we crossing wires in terms of what we mean when we say stored in the computer? From the S/X tear down/ independent development, the rear drive unit was the master computer for motor power. For a given varient, a central computer would theoretically have the same limits that the installed drive units would (along with a check to ensure the correct drive unit was installed)


    If heat is the issue, a lower performance DU could handle the same high current, but not for as long. Could the 2 x burn be partly for thermal stabilization given parameters shift due to temperature also?

    Right, that is what he said. I'm just hypothesizing that they would get better P DU yields if they pre-sorted the the FETs rather than seeing what the combined 24 randomly selected parts averaged out to (like gain testing parts when building DIY high power audio amps).

    Super, thanks for that piece of information. That supports the idea of switching device heating as the power limit, since the loss is I^2*rds_on. So a 10% part variation yields a 20% power dissipation variation.
     
  10. M109Rider

    M109Rider Member

    Joined:
    Apr 8, 2018
    Messages:
    519
    Location:
    Kitchener, Ontario
    Interesting thread for sure.

    It would be disappointing (because we all want bigger better parts, not smaller parts), if the AWD has a physically different/smaller motor.

    However at the end of the day;
    “The AWD model is slower than the AWD P model, but faster than the RWD model. “

    This will likely always be the case, and should be, since we pay for what we get.

    Still a very interesting thread, and I’m still interested to know the facts.

    After reading everything, it seems to me it is a physically smaller motor, as opposed to just software, but I hope that’s wrong, since I have and AWD on order. :)

    Cheers.
     
  11. Knightshade

    Knightshade Active Member

    Joined:
    Jul 31, 2017
    Messages:
    4,491
    Location:
    NC
    Sure- but since they've been spitting out 800A RWD DUs since last year it seems unlikely they'd suddenly find they have significant #s of units not capable of 800A.

    Hence making it far more likely the AWD rears are the exact same thing, exactly as capable of 800A, that they simply programmed to run at lower power to differentiate between AWD and P.

    Well, the data where the 800a/500a info came from is from hacking into Factory Mode on a model 3-

    Tesla Model 3 gets hacked, reveals more details and great potential for dual motor/ performance versions

    It's possible it's pulling info from diferent computers/chips in the car... but if it was only pulling info from a self-identifying drive unit, which was an 800a RWD one (and shows on the screen as such), where would he have gotten the 500a number on the rear AWD unit?

    Only place that would make sense is if there's some programming in the cars computer systems that sets a software limit for the DUs based on the "type" of car.... since it's impossible for it to have set that limit from a drive ID- since no such drive ID existed in that car.



    Entirely possible, but that would only explain RWD vs P differences despite the physical part being the same. It wouldn't explain the AWD vs RWD one.



    Do they actually do that level of assembly themselves at Fremont? I would expect those boards to come already built by the time they get to the actual plant, where they're doing the sorting.
     
    • Helpful x 1
  12. mongo

    mongo Well-Known Member

    Joined:
    May 3, 2017
    Messages:
    9,158
    Location:
    Michigan
    I expect GF1 gets the prebuilt PCBs from a supplier. So, were it to be done, it would need to be at the component vendor or as a pre-population step. Again, not saying they are doing this, only that it could increase yield, if needed.
     
  13. GSP

    GSP Member

    Joined:
    Dec 28, 2007
    Messages:
    2,463
    I would guess that Tesla is depopulating the inverters from 24 FETs down to 18 for the AWD rear drive unit. That is a simple change that would save a decent amount of money on every non-performance AWD car sold.

    As long as they are software limiting current, then the extra silicon is just wasted. SiC FETs likely don't come cheap, even at M3 volumes.

    GSP
     
    • Like x 1
  14. Swampgator

    Swampgator Member

    Joined:
    Apr 27, 2016
    Messages:
    948
    Location:
    Florida
    So somebody with skills could add the missing FETs back on the inverter to create a P? I guess the motor control software would need to be hacked as well?
     
    • Like x 1

Share This Page

  • About Us

    Formed in 2006, Tesla Motors Club (TMC) was the first independent online Tesla community. Today it remains the largest and most dynamic community of Tesla enthusiasts. Learn more.
  • Do you value your experience at TMC? Consider becoming a Supporting Member of Tesla Motors Club. As a thank you for your contribution, you'll get nearly no ads in the Community and Groups sections. Additional perks are available depending on the level of contribution. Please visit the Account Upgrades page for more details.


    SUPPORT TMC