Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Why does the P3D have better acceleration than the LR3D (current versions)?

This site may earn commission on affiliate links.
There are three different rear drive part numbers in the parts catalog. Two of them say mosfet in the part number, one says IGBT.

Yes, those new ones (in addition to the one original one that's been there all along) were only added to the catalog shortly before the MR/SR/SR+ models came out.

But up to at least a month or two ago, none of the new PNs were actually being used in cars.

Because there's a whole thread where everyone posted pics of their DU part numbers, and RWD, AWD, P, and even MR all got the exact same rear drive units

There was rumor one guy in europe finally got one of the "other" ones but don't recall if it was ever confirmed or anyone else ever saw one.

Tesla uses silicon carbide mosfets in some of their drive units, very efficient but very expensive. In others they don’t.

Again though, up until at least a month or two ago (and possibly still through today at least in LR cars) they all use SiC in the rear motors for the 3.



I thought it was not the drive unit that differed but the rear inverter power?

The inverter is part of the drive unit- hence that IGBT drive unit PN (even though they don't appear to have ever actually installed one in any LR 3 so far)
 
  • Informative
Reactions: GSP
Not to be a spoilsport but realistically, the answer is "it doesn't matter" UNLESS Tesla offers an AWD to P upgrade turning all the AWDs to 3P-.

Anything is possible but I would not count on this for planning.

Musk's tweet back in Summer of 2018 for part binning could bear zero relevance today. They could have done it for just 10 cars and his tweet would still be "truthful".

My theory which I think is the most reasonable is that there is no difference between a 3LRD and a 3P+ besides advertised features like wheels, spoilers, pedals, etc.

There have been a few anecdotes that support this:

Customer orders a 3P- but got a 3LRD instead. Reports that their car was reflashed on the spot to be a 3P-.

Rumors where people mentioned but were kept down low that they ordered 3LRD cars but got Performance instead.

The MOST sense is to engineer everything to the same standards/tolerances and control all of it through software.
 
Some references and dates and data that might help folks out are over here:

Model 3 Motors on the Tesla Parts Catalog

That first referenced post is from Oct 29, 2018...mentioning the parts catalog contains only ONE rear DU- for all versions of the model 3 (it specifies with 75kwh pack- but that's because at the time Tesla though they'd have a "totally different" smaller pack for SR, which didn't end up working out)

It also mentions only one front DU, which is for all AWD versions of the Model 3.



In December 2018 AlanSubie4Life updates the thread to mention there's now 3 dufferent rear DUs (still only a single front one)

ASY,3DUR,MOSFET (NOTE: This is the same one as the Oct catalog that came in every car)
1120980-00-D

ASY,3DUR,IGBT
1120970-00-D

ASY,3DUR,AWD,MOSFET-LC
1120990-00-A

Sadly Tesla has now removed a lot of the description/detail from the public catalog but someone managed to grab a screenshot before that covering at least 2 of those parts-

One is the 980 everyone had been getting- same as before, for VINs with 75kwh battery

Other is for the IGBT one, listed as to use with Model 3 50 kWh or 62 kwh RWD only- these are presumed to be the SR/MR vehicles.

No more info for the 990 motor- it says MOSFET-LC (versus just MOSFET) and says AWD but no content beyond.


Anyway- what follows is months of folks guessing about what anything is... and folks crawling under cars to see what they have....and literally everyone having the 980 rear motor- P, RWD, AWD, MR, doesn't matter. At some point they actually stop listing all of these in the catalog too.

The 3 listings come back by early March around launch of SR/SR+ models...but the 990 no longer says AWD, in general they've pulled a lot more detail out- and up to that point still everyone had only seen a 980 DU in the rear of any version.


Then March 16, one guy (with 3 total posts to this day, all in that thread) reports a 990 motor in his Swedish delivered 2019 AWD non-P. No picture to verify and nobody else ever reported one.

AFAIK nobody's ever reported the 970 one in anything.



TL;DR version-

At least through February 2019 all Model 3s came with exactly the same rear DU as every other one- P, AWD, RWD, MR, didn't matter.

One person reported, but nobody has verified or also reported, getting a different DU in a single European March 2019 delivered LR AWD car.
 
Creating differentiation and ramping up production are contradictory objectives.

Unless substantial savings can be done that paid for the logistics and scaling challenges.

Tesla giving up the true 3SR also speaks against differentiation.

Tesla tries to survive until Model Y and redesigns the Model 3 to be a low cost car that they can actually make money off of.
 
BTW, if I were to speculate- LC on a PN usually means Low Cost... given the fact they didn't bother having a distinct PN for the P vs AWD for a good while after production began I could see something like this being true:


A) Elon says "We will test and bin the motors and the best ones will meet the spec for P cars!"

B) In actual testing it turns out they're ALL good enough for P cars. Tesla overbuilt the thing. They don't bother actually binning (which would mean giving it a different PN... see also CPU makers who bin chips into different PNs based on tested performance).

C) Someone points out "Given the situation, we can probably whip up a cheaper downgraded version that's still good enough for non-P cars and costs less- given right now we're putting the most expensive motor in EVERYTHING since that's literally all we have, maybe we should do that!

D) They do... and starting March 2019 the "low cost" but still MOSFET (now MOSFET LC) rear DU appears in at least 1 european AWD non-p, allegedly.


Still unknown (if the above were correct)- are they only doing that in Europe? Are they only doing it in one batch to see what fleet performance/reliability looks like before widespread use? And where the heck is the IGBT part that was supposed to go into all the non-LR cars but hasn't been seen by anybody so far?



One upside for Tesla if the above is true though- once they widespread-deployed the LC rear they'd be able to offer a for-cash "Upgrade your pre-March-2019 AWD Model 3 to a P3D- for the low low cost of $6000!"


This would:

Bring in a big pile of cash to Tesla.
NOT cannibalize P sales since "new" AWDs wouldn't ever be able to do it.

And bonus since that's the effective price premium P3D- buyers paid anyway, they couldn't complain about being screwed with an early adopter tax....and P3D+ owners could still tell everyone how awesome their boat anchors and inconvenient brakes are :p
 
  • Like
Reactions: omkar
Even if you bought the upgrade, it would not change the brakes or front sway bar (which I understand is thicker) or wheels and higher performance tires. Usually when a car Mfr increases HP in a model, they make other changes to accommodate the increase in power. For all I know the stock tires do not have the same speed rating as the MP4S that come on the performance model. Yes the top speed could be limited.
 
I've asked each time for anyone to suggest any practical reason or benefit for doing it via SN, and been met with total silence every time.
Part numbers are the reference for the engineering definition and manufacturing process. Products have the same part number because they are the same design with the same design tolerances, are physically interchangeable and meet a minimum performance level. Due to uncontrolled variability in the manufacturing process, some products will yield different levels of performance. That performance is only able to be measured after manufacturing is complete and the product is put on a test stand. Test results, which would include a large dataset for each production unit, needs to be tracked and associated with each unit. This must be done based on something unique from the part number. A serial number is used.

The "practical reasons" are many. There's often motivation to track product performance or maintenance over time, in field use, to help isolate variability in the production process. Serialization permits a traceable feedback loop from field experience to manufacturing with the goal to further reduce variability. Serialization tracking is expensive and complicated, but is important to make complex products in a repeatable fashion. A long-term goal would be the removal of product serialization after that product is considered more a commodity. This can happen when the manufacturing variability has been well characterized and controlled. Serialization also allows part utilization to be optimized; meaning higher performing products are selected for more demanding installations.

We have been told by a Gigafactory engineer that the DUs are serialized and sorted prior to installation, as they should be. Product serialization is of course extremely common. Serialization to differentiate components based on tested performance is less common but still pervasive. Other examples are optics, diamonds, specialty pharmaceuticals and jet engines. These products are generally produced to the same spec, tested, binned and installed based on test results.

TL;DR--Tesla serializes their Drive Units. They would be acting irrationally if they didn't. They do this to improve their manufacturing processes using serial number tracking and feedback about Drive Unit performance over each unit's useful life. Tesla also sorts the DUs by serial number prior to installing them in different car configurations to maximize the likelihood of individual units reaching their intended lifespan.
 
I can't seem to find a definitive guide to the differences in the P3D compared to the LR3D.
Why does the P3D have much higher acceleration?
Does it have a larger battery, more current available from the existing battery, better motors, something else?
Or is it only software? The Tesla website only lists the brakes, wheels, suspension, etc.
Tesla Model 3 Performance vs. Dual Motor - Dynojet [ NationalSpeedInc | 28 sec]

AWD = 367HP, 394ft-lb,
PERF= 464HP, 496ft-lb

But here's the thing. ICE manufacturers measure their vehicle performance at the engine crank,
not at the wheels, so if you want to do an apples-to-apples comparison,
add 15% to those numbers (and that's being conservative):

AWD = 422HP, 450ft-lb
PERF= 533HP, 570ft-lb

The corrected numbers above are a more representative figure of what an electric motor's "at the crank" power output would be,
which is what the ICE companies put on their sales brochures.
This is a big distinction, since the crank loses power by having to turn a bunch of other things like a transmission, A/C, alternator,
drive shaft, etc. whereas an electric motor's mechanical output only turns the wheels.

For example, the Mustang GT is listed as 460HP. The dyno output at the wheels comes in around 370HP.

 
  • Love
Reactions: diamond.g
Tesla Model 3 Performance vs. Dual Motor - Dynojet [ NationalSpeedInc | 28 sec]

What you're looking at in the dyno graph below is an overlay of the two vehicles
- in Performance Mode (Red and Blue Graphs),
- followed by the graphs of the two vehicles in "Chill" mode.

It is amazing that two vehicles can go from such a wide variance in power
to a nearly perfect overlay with the push of a button.

Coming in at an absurd 463.71awhp & 496.28lbft,
Hoang's Model 3 in Performance Mode is definitely no slouch!

Peak numbers aside, check out that torque curve...
it's nearly vertical at tip-in! 0 to 500 lb ft instantly. Insanely-cool.
2018 Tesla Model 3 Dual Motor vs. Performance .jpg
 
Tesla Model 3 Performance vs. Dual Motor - Dynojet [ NationalSpeedInc | 28 sec]

AWD = 367HP, 394ft-lb,
PERF= 464HP, 496ft-lb

But here's the thing. ICE manufacturers measure their vehicle performance at the engine crank,
not at the wheels, so if you want to do an apples-to-apples comparison,
add 15% to those numbers (and that's being conservative):

AWD = 422HP, 450ft-lb
PERF= 533HP, 570ft-lb

The corrected numbers above are a more representative figure of what an electric motor's "at the crank" power output would be,
which is what the ICE companies put on their sales brochures.
This is a big distinction, since the crank loses power by having to turn a bunch of other things like a transmission, A/C, alternator,
drive shaft, etc. whereas an electric motor's mechanical output only turns the wheels.

For example, the Mustang GT is listed as 460HP. The dyno output at the wheels comes in around 370HP.


Back around 2004, the new SAE J1349 standard for HP rating was established. I know that dropped some of the manufacturers numbers, but I'm not sure if it retained crank vs rear wheel or not.
 
Even if you bought the upgrade, it would not change the brakes

Good, I wouldn't want the + brakes anyway.

If I tracked the car (the only reason the stock brakes wouldn't work as well or better than "upgraded" ones) I'd go with MPPs better offering anyway.


or front sway bar (which I understand is thicker)

yes, by 1 mm.

I doubt you'd notice, but if you think you would, even thicker bars are available in the aftermarket.


or wheels

Again, good- I wouldn't want the boat anchor 20s (can you tell I find the P3D- the much superior car to the P3D+ yet?)


and higher performance tires.

PS4S tires are of course available in the stock size for the 18" wheels too.


Usually when a car Mfr increases HP in a model, they make other changes to accommodate the increase in power. For all I know the stock tires do not have the same speed rating as the MP4S that come on the performance model. Yes the top speed could be limited.

The P3D- came with the same wheels/tires as the regular AWD. The MXM4s.

Which are speed rated to 168 mph, higher than the highest top speed on any Model 3.


I believe there's some bigger issues, such as the VIN that comes into play and that then relates to insurance.

Nope- same VIN P or AWD.[/QUOTE]
 
Part numbers are the reference for the engineering definition and manufacturing process. Products have the same part number because they are the same design with the same design tolerances, are physically interchangeable and meet a minimum performance level. Due to uncontrolled variability in the manufacturing process, some products will yield different levels of performance. That performance is only able to be measured after manufacturing is complete and the product is put on a test stand. Test results, which would include a large dataset for each production unit, needs to be tracked and associated with each unit. This must be done based on something unique from the part number. A serial number is used.

That's completely untrue.

See again CPUs- where PNs are assigned after testing.


Faster CPUs get assigned higher-end PNs, and are sold at higher prices.

Slower ones get assigned lower-end Pns and lower prices... unless there's a glut of higher performance ones in excess of demand, then sometimes they'll sell a "better" part under a cheaper PN.... (one company in that position once famously drilled holes through more expensive parts to sell them under a cheaper PN because yield of the better one was so high and demand for the cheaper one so high)


But they always BIN them by part number because that makes tremendously more sense for supply chain, repair parts, and more.



The "practical reasons" are many. There's often motivation to track product performance or maintenance over time, in field use, to help isolate variability in the production process. Serialization permits a traceable feedback loop from field experience to manufacturing with the goal to further reduce variability.


You seem to be confusing two different things.

The "many" reasons you give- which really was just one reason BTW- is a reason to HAVE serial numbers.

it's not, at all, in any way, even slightly, a reason to use them instead of PNs when binning a product after testing it in the factory.

In fact using SN, as described, for that purpose, is demonstrably inferior to PN, as already explained in some detail.


Let's go back to CPUs again.

intel CPUs have SNs. But they're not BINNED by them. After binning they get PNs based on what they tested at.

The SN still exists- and is still tracked- but NOT for things like replacing a part under warranty.... where PN is a much better way to do it- and requires all parts with the same PN to be functionally swappable.

SN remains useful for things like knowing which CPUs, specifically, to recall if they find out there was a manufacturing defect during a specific production run. Also for knowing when a specific part was sold for warranty reasons (this one is less useful for a car maker of course).

But using SNs to differentiate performance? That's nonsensical and you've still given -0- practical reasons to do that instead of PN like everyone else does.


Serialization also allows part utilization to be optimized; meaning higher performing products are selected for more demanding installations.

So does PN binning.

And without making your replacement parts ordering system, or off-shelf-parts-picking a nightmare.


We have been told by a Gigafactory engineer that the DUs are serialized and sorted prior to installation, as they should be. Product serialization is of course extremely common.

Sure, just as mentioned for CPUs.

But they don't use the same PN for parts that aren't entirely interchangeable in repair.

And nothing in the like 4 words the GF employee said tells us otherwise.


If an AWD cars DU dies, or a P cars DU dies- the same PN is ordered to replace it-or the same PN is pulled off the shelf if the SC has, as Elon claimed, begun actually stocking some parts- since that same PN was what was in the car from the factory anyway (that one weird european car notwithstanding).


Inventing a whole secret inventory system where replacement parts have a second, not-listed-in-the-actual-parts-catalog, tracking/sorting method, that appears to make everything harder/more complex, yet does not provide a single benefit anyone can articulate simply falls apart to occams razor here.
 
yet does not provide a single benefit anyone can articulate
As previously stated, they do this to improve their manufacturing processes using serial number tracking and feedback about Drive Unit performance over each unit's useful life. They obviously cannot do this by part number.

Besides tracking cars by VIN, which is of course a serial number, I expect they also track the battery (pack and modules) plus select electronics by SN.

How else would they do this?
 
As previously stated, they do this to improve their manufacturing processes using serial number tracking and feedback about Drive Unit performance over each unit's useful life. They obviously cannot do this by part number.

If they're all the same PN, all built to the same spec, but some randomly test out better, how does tracking the SN "improve the manufacturing process", specifically?


Certainly SN helps them get feedback over the life of the car- but that has literally nothing to do what we're actually discussing.


If they actually bin the motors then they would use a different PN on the P motors, because they would be different parts and when it came time to replace one, they could insure they use the correct replacement part based on the part number physically visible on the part

Like everyone else on earth does.


Just like Intel tests a batch of CPUs they get different PNs based on how they test out. They also know each ones SN, but they don't use that for replacement parts because doing so makes no sense at all- and you've yet to offer a single example where it does.





I expect they also track the battery (pack and modules) plus select electronics by SN.

I'm sure they do. As I said that's helpful for things like knowing which batteries need to be replaced if they had a bad batch (this actually happened, there's a service bulletin out there about it with impacted VIN range)

But when they need to replace one of them guess what they use to order the replacement ? part number not serial number.
 
Last edited:
Ok, I see what you're saying. It seems like Tesla doesn't want to reassign the Drive Unit PNs after testing. I can't explain why that would be the case.


Probably because there's no actual DU difference in what they put in an AWD and a P (at least up until a month or two ago) and thus no reason to use a different PN :)

Same motors and inverters in both, and only "performance difference" is the software. Hence also why numerous folks reported getting an AWD car at delivery and the dealer could just software flash it to a P on the spot.


Where life gets interesting is going forward, now that there IS a different PN in the catalog for the LC rear DU for AWD cars.... even though it's only been reported as used in a single (non confrimed) case thusfar.... plus the different PN for the IGBT rear DU for SR/MR cars (that doesn't appear to have actually been used in anything yet).

If those DUs start actually showing up in built cars then there might be practical reasons going forward from that point that the performance of the cars differ other than software, but that hasn't been the case so far.