Model S Performance vs BMW M5

[chrisn]

The confusing part about the limp mode thing is that it is so confusing. The car's UI should convey enough information so the driver knows what action/condition caused the limp mode. And Tesla the company should proactively (e.g., in the manual and website) describe the designed behavior under sustained heavy loads.

Another weird item is that car magazines will surely call this out in their reviews-- wouldn't Tesla want to control the "spin" proactively. Imagine if the performance version of Panamera, 5-series, S5, CTS-V, E-class, etc. went into limp mode during a couple of hot laps. The car would get crucified in the reviews....

If my M5 could not complete a 20 minute session of hard driving without going into self protect mode, I would seriously insist that BMW gave me my money back.

 

[Tommy]

If my M5 could not complete a 20 minute session of hard driving without going into self protect mode, I would seriously insist that BMW gave me my money back.

You wouldn't start your 20 minute session with only a gallon or two of gas in the tank and then insist that BMW give your money back because the car ran out of fuel during your spirited run. Effectively, that is what occurred with the S... a low battery charge triggered the reduced power; that's the hypothesis as I understand it with the S's reduced performance.

 

[eledille]

I would definitely add ambient temperature to the mix, if you create an experiment. If you could magically monitor the pack temp, and so therefore see the delta, then you'd maybe know if a throttling was caused by thermals rather than voltage.

Hypothesis 3: MSP quickly limits power on a race track due to high temperature somewhere in the system.

Exp 3.1: See if it cuts early with a fully charged battery. If it doesn't, then it would have to be a thermal problem that only arises at low pack voltage. Sounds unlikely to me, everything is sized for amps and I doubt they will pull more amps to counter voltage dips. That would deepen the dip. It's possible, but due to the very high minimum allowable cell voltage the difference between a deep dip and a very deep dip might be something like 0.3V/cell, so somewhere between 5 and 15% - shouldn't make that much of a difference anyway.

Exp 3.2: At the same SoC, see how long before it goes into reduced power mode on a hot and a cold day. If there's little difference, then the hypothesis is falsified.

Exp 3.3: Glue/tape temp probes to the radiator inlets and see if there's a difference between high SoC and low SoC laps. If not, then temperatures can't explain the difference in behaviour.

The confusing part about the limp mode thing is that it is so confusing. The car's UI should convey enough information so the driver knows what action/condition caused the limp mode. And Tesla the company should proactively (e.g., in the manual and website) describe the designed behavior under sustained heavy loads.

Agree, I just posted this on the thread about the performance of the cabin heater. They have all the data on temps, voltage, current, etc, but they don't display them. A "tech-head" tab in the UI that displays stats and warnings would be really nice.

If my M5 could not complete a 20 minute session of hard driving without going into self protect mode, I would seriously insist that BMW gave me my money back.

The battery is not like a gas tank at all. Its performance will vary according to state of charge. I'm guessing that the car will perform as expected down to about 50%. That should give you 8 minutes of continuous full power (it will last longer on a real track, of course).

 

[chrisn]

You wouldn't start your 20 minute session with only a gallon or two of gas in the tank and then insist that BMW give your money back because the car ran out of fuel during your spirited run. Effectively, that is what occurred with the S... a low battery charge triggered the reduced power; that's the hypothesis as I understand it with the S's reduced performance.

So I'd need to trailer my Model S to the track to ensure enough SOC to drive normally for more than one session (versus normal four or five sessions)? I think EV owners know they might need to top-off before driving home from track (depending on distance and charging facilities (Supercharger nearby?), they might need to "stay for dinner" or longer), but come on-- don't be a fanboy here. Most road course tracks are at least 50-100 miles away from city centers. I had assumed that you could drive your MSP to the track, have essentially a normal track day (if you could 240V recharge between sessions), then drive home (perhaps after recharging some more). Based on information disclosed byTesla, my assumption seemed reasonable until this user-discovered car behavior was revealed.

I wish there were more data from experiments, but ideally Tesla would have simply disclosed the design parameters in the first place.

It seems likely that the "performance" (little "P") you get with "Performance" option is akin to an "over boost" mode that really pushes things beyond design limits and is not indicative of sustained load in a performance driving context.

 

[Tommy]

Trailering a car to the track isn't all that unusual. I am going to wait until more data is available via other driver's reporting similar experience and also hear from Tesla themselves on the cause and their "fix" before attributing the track limitations to inherent design limits. And if the design limits prove to be true, not the car to take to the track.

 

[DaveN007]

You wouldn't start your 20 minute session with only a gallon or two of gas in the tank and then insist that BMW give your money back because the car ran out of fuel during your spirited run. Effectively, that is what occurred with the S... a low battery charge triggered the reduced power; that's the hypothesis as I understand it with the S's reduced performance.

The Roadsters were going into "limp mode" with fully charged batteries. It was a heat issue. On a 100 degree day, my Cayman barely registers warm after 20 minutes of HARD driving. That is what a hundred extra years and a quadrillion more miles of development get you. Better performance under conditions that don't matter to 99 percent of the motoring public. ;-) These conditions seem to get a lot of attention, however. Give Tesla 20 years and we will finally get our hoverboards and personal electric turbofan jetpacks.:love:

 

[Doug_G]

Any car on the track consumes far more energy. My S2000 is fairly fuel-efficient on the highway, but consumes nearly 3 tanks of gas on a track day. Also I can't take it out with anything less than half a tank or I'll get into trouble due to the fuel sloshing. So there are some parallels...

My Roadster will only do three laps before a power limit due to motor temperature. The Model S has liquid cooling so I am hopeful it will do better. One thing I will point out, though, is when I take my Roadster to the track I fully charge it before starting, keep it plugged in whenever it's not on the track, AND spend 2 hours recharging so I can get home. I would expect the same for the Model S.

 

[chrisn]

Trailering a car to the track isn't all that unusual.

Really? Really? Again with the apologist/fanboy stuff. Trailering a track car/racecar to the track is not unusual, but trailering a stock sedan to the track is laughable. I have never seen it. Have you? What would be the point? I've seen stock show/shine Ferraris roll out of trailers, but that is because the owners are avoiding $10/mile depreciation (and/or own a matching Rosso Corsa tow vehicle and trailer).


Anyway, I agree that no one knows the real deal yet, so it is too soon for firm conclusions, but if the conclusion is: "If you want to push your Model S to the limits on the track, you'll need to trailer it and only drive for one or two sessions," that would be pretty embarrassing, IMHO. Academic and not relevant to street driving-- but embarrassing.

Reality: in the $100K performance car category, lap times (and other silly stats) impact sales.

 

[eledille]

I wish there were more data from experiments, but ideally Tesla would have simply disclosed the design parameters in the first place.

I don't remember seeing Tesla claiming MSP would be a good track car?

It seems likely that the "performance" (little "P") you get with "Performance" option is akin to an "over boost" mode that really pushes things beyond design limits and is not indicative of sustained load in a performance driving context.

No, there's no "over boost" mode, it's just that you didn't know how batteries work. No RC car enthusiast would dream of entering a competition without the battery hot off the charger, because that's when they perform best. Performance drops with SoC, li-ions drop off less than other chemistries, but they still drop. And during heavy load, you will hit the voltage limit before the battery is empty. This is just the way it is.

The Roadsters were going into "limp mode" with fully charged batteries. It was a heat issue. On a 100 degree day, my Cayman barely registers warm after 20 minutes of HARD driving.

I believe that was a completely different issue, and one that Tesla has fixed.

If I'm right, you will be able to charge the MSP in range mode, drive nicely for 50 miles to the track, and get maybe 15-20 minutes of good performance before you need to charge. If that's not good enough for you, then don't buy MSP.

That is what a hundred extra years and a quadrillion more miles of development get you. Better performance under conditions that don't matter to 99 percent of the motoring public. ;-)

On the other hand, an EV can get you vastly better performance on the road... and how long did it take them to achieve that? :biggrin:

 

[Anzir]

The Model S was not developed as a car to take to the track and toss around. Given that, it's impressive that we can even consider it in the same breath as the M5. But it is going to require a lot more development to become the kind of all-around beast that the M5 is. The kind of car you drive back and forth to the track while embarrassing Corvettes on and off the track.

 

[stopcrazypp]

The Roadsters were going into "limp mode" with fully charged batteries. It was a heat issue. On a 100 degree day, my Cayman barely registers warm after 20 minutes of HARD driving. That is what a hundred extra years and a quadrillion more miles of development get you.

The Roadster issue clearly has to do with air cooling (I'm pretty sure that was already mentioned) and nothing to do with years or miles of development. Tesla never designed the car with the track in mind, so it's a different engineering issue. The Model S we don't know yet (esp. when fully charged), but I would be surprised if it's still a cooling issue given Tesla switched to a liquid system.

 

[chrisn]

No, there's no "over boost" mode, it's just that you didn't know how batteries work. No RC car enthusiast would dream of entering a competition without the battery hot off the charger, because that's when they perform best. Performance drops with SoC, li-ions drop off less than other chemistries, but they still drop. And during heavy load, you will hit the voltage limit before the battery is empty. This is just the way it is.

I actually am an R/C enthusiast and do know how batteries work. With LiPos, even under heavy load, the voltage during discharge is very flat (voltage versus SOC). Obviously, voltage drops more with more load, but at heavy loads (I am looking at 25C curves) this drop is approximately constant (very stable voltage with load) between about 90% SOC and 15% SOC.

You certainnly get an "extra punch" from being nearly-fully charged (100% down to 95 or 90%). This is the abnormal part of curve. From there, under heavy load, performance is VERY flat until you get near the exponential decay part (where Low Voltage Cutoff would kick in).

So, in reality, purely from the perspective of battery performance, the story is MUCH more about "not being topped off" versus "being nearly empty." Characterizing performance based on tests conducted with SOCs from 100% to 90% (or whatever) isn't really fair.

The "wrongly/dumbly extrapolated range bug" hypothesis is possible, but heat related issues seem just as plausible.

I actually think the "overboost" analogy might be very apt. It implies a burst performance that it outside the real envelope for sustained load. It would enable 0-60 and 1/4 mile tests (apparently only one or two at peak performance), but would not allow sustained performance at this load. This could either be due to loss of the "extra punch" of being near 100% SOC or could be due to heat issues-- or a combo.

 

[Robert.Boston]

I agree with Jay Leno on this: let's all drive electric during the week so that we can take our ICEs out to play on weekends at the track. There's simply no disputing that gasoline has much higher energy density than batteries, and that a gas tank can deliver a higher-rate, continuous stream of energy-potential than a battery. Car buyers (looking for non-track vehicles) will simply need to learn that they shouldn't be selecting cars based on how well they perform on a racetrack. Farmers never used quarterhorses to do useful work. Likewise, I question why comparisons to an M5 are relevant for all but a tiny slice of potential buyers--perhaps they've been told it's relevant by the automotive press, but it's not.

 

[Doug_G]

I agree with Jay Leno on this: let's all drive electric during the week so that we can take our ICEs out to play on weekends at the track. [snip] Car buyers (looking for non-track vehicles) will simply need to learn that they shouldn't be selecting cars based on how well they perform on a racetrack.

+1 Robert. If you've set up a car to work really well on the track, it will be awful to drive on the street. (Maybe there's a sports car out there that can do both equally well... but I am dubious.)

Driving a Roadster got me interested in motorsports, but I soon picked up a ten-year-old S2000 to use for that purpose. My S2000 is stock but with good tires and a mildly aggressive alignment it's pretty darn good on the track. I can keep up with much more powerful cars because it's so darn good in the corners. But it's an hour and a half drive to the track, and that trip is kinda tedious! On the street it feels annoyingly busy and bouncy. On the track it's incredibly responsive, you can really feel what it's doing, and you can take it to the limit of grip and feel confident doing it. The car is designed to do that. It's not really designed for a comfortable Sunday drive.

Sure, I'll take my Model S to Shannonville to see how it does. Really curious to find out. But that will be a very occasional thing. I'll keep the S2000 (or some successor) for the track.

 

[eledille]

I actually am an R/C enthusiast and do know how batteries work. With LiPos, even under heavy load, the voltage during discharge is very flat (voltage versus SOC). Obviously, voltage drops more with more load, but at heavy loads (I am looking at 25C curves) this drop is approximately constant (very stable voltage with load) between about 90% SOC and 15% SOC.

Compare curves for 1A and 5A here.

5A for a 3400mAh cell is approximately 1.5C. Tesla reserves some capacity at both the top and the bottom end, so assume the MSP battery is really a 100 kWh battery with 15 kWh hidden for protection and longevity. If so, then 310 kW is slightly more than 3C, so about twice as fast a discharge as the 5A curves in that link.

I don't think they're all that flat... these are high-capacity cells. They're "very flat" compared to e.g. NiMh, but they're not flat.

Assume the trigger point is 3.2V cell voltage and check where the curves cross the 3.2V line in the various charts. What would the 10A curve look like, if there was one? You would get cutoff somewhere around 50% SoC, which seems to fit very nicely with my hypothesis.

If this is the case, then it might be possible to reduce max power smoothly instead of suddenly cutting power in half. That would give better and more predictable performance, and likely generate fewer negative surprises too.

 

[jerry33]

+1 Robert. If you've set up a car to work really well on the track, it will be awful to drive on the street. (Maybe there's a sports car out there that can do both equally well... but I am dubious.)

I've never seen one either. Every person that I know or have heard about who has tried to do this through after-market modifications has a) Spent a ton of money b) Gotten rid of the car after about three to six months actual driving and c) Become wiser about spending money on aftermarket modifications.

 

[chrisn]

I don't think they're all that flat... these are high-capacity cells. They're "very flat" compared to e.g. NiMh, but they're not flat.

Dude, those are laptop/flashlight batteries and charts have auto-fit origins. Use a 0,0 origin and they will be flatter-- especially cells engineered to be high discharge cells (like an R/C pack and presumably the new Tesla batteries). Laws of physics/chemistry of dictate that they not "really" be flat forever (that would be magic), but, until you get to LVC range, there is very little fall off. Contrast the significant "boost" / kick you always get (on extended range full charge) at the BEGINNNING of the curve.

Observation: Battery will perform BETTER THAN NORMAL for first ~5% of SOC, then will performance will be essentially stable until you get close to end of usable charge window. Therefore, the representative testing should be done at around 80% SOC. A drag test at 100% SOC is not a fair representation of performance.

Edit #2: I was imagining that WOT was more than 3C for MSP, but I agree with your math. But even at lower (but still significant) loads, it just takes a bit longer (maybe 80% instead of 90% SOC) to get to "flat-ish" part-- but when you get there, it is pretty plat. The WOT performance of MSP should be characterized in the 80%-20% SOC range to be fair.

Edit #1: See curve below from R/C pack designed for 40C discharge (not origins aren't even zero on this chart):




- - - Updated - - -

I've never seen one either. Every person that I know or have heard about who has tried to do this through after-market modifications has a) Spent a ton of money b) Gotten rid of the car after about three to six months actual driving and c) Become wiser about spending money on aftermarket modifications.

I don't folks here are talking about using the MSP or M5 as a "donor car" to build a race car. I thought we were talking about the ability to drive your bone stock luxury car to the track, hang with sports cars and push your car to the limit, then cruise home in comfort.

As an aside, the newer Porsches with adjustable shocks/suspension can be made to be very quick on the track while retaining stock drivability around town. Of course, the "baseline" comfort/drivability of a Porsche (apart from Pan) is not the same as an MSP (but I suspect better than a Lotus/Roadster).

 

[jhs_7645]

I don't folks here are talking about using the MSP or M5 as a "donor car" to build a race car. I thought we were talking about the ability to drive your bone stock luxury car to the track, hang with sports cars and push your car to the limit, then cruise home in comfort.

As an aside, the newer Porsches with adjustable shocks/suspension can be made to be very quick on the track while retaining stock drivability around town. Of course, the "baseline" comfort/drivability of a Porsche (apart from Pan) is not the same as an MSP (but I suspect better than a Lotus/Roadster).

I've been lurking in this thread since the beginning because I thought it was fascinating that the MSP could even be in the same conversation as an M5. But I've come to a personal conclusion that chrisn is playing the roll of a troll. Yes, he throws the occasional bone to the Tesla bunch but overall, if you look back at the posts, he is just that. No matter what MSP folks say, he's going to re-define the argument. Yes, the M5 with decades of engineering and design and running on gas will out last an MSP on the track. I, with a bicycle and a sufficient supply of water and food, will also beat the M5 given certain constraints. The argument has become pointless. It's beyond academic at this point.

 

[chrisn]

The OP and early commenters framed the central Q as: which is faster around the track? I have not tried to redefine that question. I have played all with various excuses/explanations, but have not tried to dodge the central question. I'm the one providing real world data.

I sincerely struggled with MSP v M5, and made my choice (I also got a Volt for DD), so obviously have a bias. That's all.

 

[stopcrazypp]

I've been lurking in this thread since the beginning because I thought it was fascinating that the MSP could even be in the same conversation as an M5. But I've come to a personal conclusion that chrisn is playing the roll of a troll. Yes, he throws the occasional bone to the Tesla bunch but overall, if you look back at the posts, he is just that.

Chrisn seriously considered the MSP before deciding on the M5 (look at his posting history). Yes, his current ownership of the M5 does given him some bias (just like owners of an MSP will have bias), but I think he's genuinely just trying to contribute to the MSP vs M5 comparison given he owns an M5 (by actually driving at Laguna Seca to get data on the M5).

I think the central question by the OP was initially framed as MSP vs M5 on a drag strip/stop light. They are almost neck to neck on that front (depending on launch control usage, SOC etc.). It deviated to other tracks after a few posts. I think there was agreement that any track with a top speed much higher than 100mph will have the MSP obliterated. There was not much agreement on which would perform better on slower tracks.

But he's right that the facts so far show the M5 is faster on Laguna Seca than the MSP (with the caveat that we are comparing a prototype with production car and different drivers and that we don't have much data on the MSP). And Laguna Seca isn't really that fast of a track in terms of top speed.