Interesting Chris Harris review (of Cali T) - sheds light on other Turbo options

[Boss330]

Originally Posted by CanAutM3

Or they could have simply selected gear ratios that maximize the power plateau of the S55 as it is (you know, that long term gripe of mine ). This would require 7300rpm shifts in every gear to maximize acceleration. Isn't that what Ferrari is trying to mimic?

It would seem so, yes

[ASAP]

LOL, Cobb AP for an N54 and N55 has offered maps like this for years... as have any of the piggy tunes... Ferrari witchcraft... LOL

The moment you drive these linear maps... u immediately want your turbo tq back.

[bruce.augenstein@comcast.]

Originally Posted by ASAP

...The moment you drive these linear maps... u immediately want your turbo tq back.

Agreed. My only experience with actively managed power/torque was with my son's '05 GTO. GM supplied that 400 HP car with "Torque Management" software that limited things in first and second gears. Tromp on it in first and it felt sort of lively, but also prompted a "Where's the 400HP?" question. Pretty similar in second, but then torque management went away in third, which seemed to accelerate just as hard as second gear. Felt weird.

One of the first things he did was to send the ECU away for a reprogramming, and in addition to having some extra power (355 HP on a Dynojet), all was well with driver expectations, as Torque Management was done away with.

Next move, documented here, was a supercharger, and thinking back, maybe reinstating Torque Management would have been a good idea, as unless it was very hot and you were on clean asphalt, full throttle was a no-no until third. Even then, my theory was that you would need to be drag racing on the surface of the sun to actually use full throttle in first.

In any event, I know that Ferraris have been top-end-rush cars since the Beginning of Time, so they have something to protect here - but I'm skeptical of the entire process, and thankful BMW hasn't resorted to something similar. I know that the E9X faithful are skeptical/pissed off/outraged about the new car and its relative lack of top-end rush, but there'll be a bunch of converts after the new car blows some exhaust in their grilles. That's when the siren song of 8400 RPM goes a little sour.

Bruce

[myzmak]

Originally Posted by bruce.augenstein@comcast.

Agreed. My only experience with actively managed power/torque was with my son's '05 GTO. GM supplied that 400 HP car with "Torque Management" software that limited things in first and second gears. Tromp on it in first and it felt sort of lively, but also prompted a "Where's the 400HP?" question. Pretty similar in second, but then torque management went away in third, which seemed to accelerate just as hard as second gear. Felt weird.

One of the first things he did was to send the ECU away for a reprogramming, and in addition to having some extra power (355 HP on a Dynojet), all was well with driver expectations, as Torque Management was done away with.

Next move, documented here, was a supercharger, and thinking back, maybe reinstating Torque Management would have been a good idea, as unless it was very hot and you were on clean asphalt, full throttle was a no-no until third. Even then, my theory was that you would need to be drag racing on the surface of the sun to actually use full throttle in first.

In any event, I know that Ferraris have been top-end-rush cars since the Beginning of Time, so they have something to protect here - but I'm skeptical of the entire process, and thankful BMW hasn't resorted to something similar. I know that the E9X faithful are skeptical/pissed off/outraged about the new car and its relative lack of top-end rush, but there'll be a bunch of converts after the new car blows some exhaust in their grilles. That's when the siren song of 8400 RPM goes a little sour.

Bruce

Well said.

[swamp2]

Very interesting move by Ferrari. I would have liked to hear more about the acoustics from someone who loves the typical V8 Ferrari "bark". I've often (who doesn't) hold Ferrari up on a pedestal and I took them at their word when they stated they would go with turbo charging when all negatives with turbos (from the NA perspective) have been eliminated. It seems clear, that at least with the Cali-T this isn't really the case.

On the flip side as long as your car has plenty of torque to spin the wheels easily/on command at a wide variety of rpm/speed/load combinations in low gears, this particular "witchcraft" seems like it could have a role to play - in both improving feel and improving performance (perhaps traction control is already doing enough in the nanny on mode). Unfortunately they also did not eliminate the turbo lag. It's probably similar to the M4 and I would expect them to be using something like BMW's "pre-tensioning" strategy, perhaps like in the M4 only in particular modes. Either way it is quite positive that Harris overall calls the entire engine/software combination "brilliant".

Also, fortunately, as Harris noted, this is sort of "NA character with a turbo" v1.0 for Ferrari. I'm hoping they deliver improvements in sound, lag and linearity in the 458 replacement and expect they will do so. Is that overly optimistic?

[swamp2]

Originally Posted by CanAutM3

Or they could have simply selected gear ratios that maximize the power plateau of the S55 as it is (you know, that long term gripe of mine ).

I really can't believe after our extensive debate on this topic that you still cling to this false idea. You've been completely unable to demonstrate that the car is doing anything but very nicely staying in it torque plateau (or very nearly so) across a wide range of gears. You've also been unable to supply a set of ratios and FD that provides any significant performance improvement (via simulation of course).

[onatuesday]

Why are OEMs trying so hard to disguise that these cars are turbocharged? Why not embrace it by showing off the turbo sound rather than trying to hide it? Cars like the Pagani Huayra pump up the volume on all the whistles, snorts, BOV releases rather than quieting them. Ferrari F40, supra, 300zx, etc all sounded great. Boggles me why OEMs are having are a hard time now with the sound?

[myzmak]

Originally Posted by trinim3

Why are OEMs trying so hard to disguise that these cars are turbocharged? Why not embrace it by showing off the turbo sound rather than trying to hide it? Cars like the Pagani Huayra pump up the volume on all the whistles, snorts, BOV releases rather than quieting them. Ferrari F40, supra, 300zx, etc all sounded great. Boggles me why OEMs are having are a hard time now with the sound?

I think you just need to read this forum for a spell to see why.

In the olden NA days, to squeeze power out of engines it required high revs. That sound/phenomenon led people to associate the sound with 'speed' and 'fun' (the proverbial dinner bell for Pavlov's dog).

Turbos block the NA sound and don't require the high revs. So, people who were brought up to think 'high rev and only high rev = fun/excitement' have a hard time adjusting. As I've said elsewhere, it is all nostalgia. You remember your first sexual experience in detail, not because it was the best but because it was formative. We always measure the present against those first experiences. With NA being the 'standard' until recently, lots of folks have a hard time letting go and embracing the new.

So, some manufacturers try hard to give those folks some of what they think they want (NA 'sound' or 'power curves') while actually doing what they need to do in this day and age (FI, lightness, etc).

But, as anyone who isn't Pavlov's dog will tell you, you don't need a f------g bell to eat supper. You can still eat - and eat just as well or better - with a different noise.

[solstice]

They should just let the turbo cars do what they do best. Deliver brutal power and TQ in small cheaper packages. The F40 is one of the best loved and iconic Ferraris of all time. It's not becuase it's so driveable or acts like a NA car. It's because it's such a raw, direct and simple car with brutal power delivery. Like a go-cart with a jet engine. The power of a monster FI engine corrupt you. It's hard to go back to less power once you are used to it. It's also hard to go to an FI engine when you are used to the sharp throttle, high end rush and scream of a performance NA engine. Don't try to make one act like the other they deserve to have their own strenghts and weaknesses and if they are one of the best of their kind we'll get used to it and like them for what they naturally are.

[myzmak]

Originally Posted by solstice

They should just let the turbo cars do what they do best. Deliver brutal power and TQ in small cheaper packages. The F40 is one of the best loved and iconic Ferraris of all time. It's not becuase it's so driveable or acts like a NA car. It's because it's such a raw, direct and simple car with brutal power delivery. Like a go-cart with a jet engine. The power of a monster FI engine corrupt you. It's hard to go back to less power once you are used to it. It's also hard to go to an FI engine when you are used to the sharp throttle, high end rush and scream of a performance NA engine. Don't try to make one act like the other they deserve to have their own strenghts and weaknesses and if they are one of the best of their kind we'll get used to it and like them for what they naturally are.

Agreed.

Frankly applies to not just turbos, but any car. When we stick too hard to one thing we know/love (NA, for example) we risk being stagnant. Sometimes we need a challenge (fuel economy and CO2 requirements, for instance) to make truly spectacular new engineering.

Without those challenges, we wouldn't be living in a year where the 918, LaFerrari and P1 all get released. Those powertrains will all drive differently from a traditional NA engine, so lets learn what they are like and learn to appreciate what they give us - just like the Turbo does.

[CanAutM3]

Originally Posted by swamp2

I really can't believe after our extensive debate on this topic that you still cling to this false idea.

Very simply because it isn't false

I understand from our previous debates that, from your simulation in CarTest, you don't believe gear ratios have an impact on acceleration performance. If it were truly the case, why would so many racing teams bother with adjusting gear ratios going from one track to another?

Originally Posted by swamp2

You've been completely unable to demonstrate that the car is doing anything but very nicely staying in it torque plateau (or very nearly so) across a wide range of gears.

It is not the torque plateau that is important, but rather the power plateau that needs to be optimized (I guess it is typo on your part). And yes, I have demonstrated that the current DCT gearing does not keep the engine in the power plateau after 1-2 and 2-3 shifts (not very difficult to do).

Originally Posted by swamp2

You've also been unable to supply a set of ratios and FD that provides any significant performance improvement (via simulation of course).

Simulation is only as good as the model that is used. While CarTest is able to reasonably predict overall acceleration by "best fitting" the acceleration curve using the different input parameters, I don't believe the sub-models within CarTest are precise enough and well calibrated enough to do more in depth analysis. For example, the tire grip/wheel spin sub-model is certainly not optimal since it grossly overestimated the low speed (off the line) acceleration of the F8X. I also remember that you mentioned that CarTest uses a linear regression relative to engine speed for drivetrain losses, this is also inaccurate. IIRC, you also mentioned that CarTest uses fixed/constant (not user input) inertia numbers for the engine an drivetrain components, which also introduces an element of error.

So it is not because CarTest gives a certain result that I take it for granted .

That being said, for the topic at hand, wider gear spacing in the upper gears (3-7) would have made for a more entertaining and engaging powertrain where (at least) 7300RPM shifts would have been needed to extract the maximum performance of the powertrain (rather than having one than can be short shifted). This would yield the same result on the S55 as the "in gear torque limting" Ferrari approach. That was the point of my reply to Boss330.

[swamp2]

Originally Posted by CanAutM3

Very simply because it isn't false

It is false until you can prove it. Come on, in any reasonable debate the onus is on the one making the claim to prove the point and you absolutely haven't proven squat.

Originally Posted by CanAutM3

I understand from our previous debates that, from your simulation in CarTest, you don't believe gear ratios have an impact on acceleration performance. If it were truly the case, why would so many racing teams bother with adjusting gear ratios going from one track to another?

No, you are generalizing and mis-stating my beliefs. You are also falsely abstracting from the specific to the general.

But it does basically come down to power, no more power no more performance. Whether it is FD or individual gear ratios. Ratios have very little effect on overall performance, generally a FD modification will provide a very small advantage in some particular contests (some times to speed, some times to distance, time for speed XX to YY, 1/4 mi time, trap, etc.) at the same time though the change will cause declines in other particular contests. Now as to specific individual ratios in the very specific case of particular sections on a particular track, yes of course, this is simply a special case of the general statement I have made above. And in such situations, fractions of a second can matter. Capiche?

Originally Posted by CanAutM3

It is not the torque plateau that is important, but rather the power plateau that needs to be optimized (I guess it is typo on your part). And yes, I have demonstrated that the current DCT gearing does not keep the engine in the power plateau after 1-2 and 2-3 shifts (not very difficult to do).

Yes, typo. And no you certainly have not demonstrated this, without some representation of wheelspin your prior analysis is utterly flawed.

Originally Posted by CanAutM3

Simulation is only as good as the model that is used. While CarTest is able to reasonably predict overall acceleration by "best fitting" the acceleration curve using the different input parameters

You have a massive and fundamental misunderstanding of what CarTest does according to the statement above. There is no curve fitting. CarTest uses basic vehicle dynamics, similar to that in "Fundamentals of Vehicle Dynamics" by Gillespie, accounting for an enormous range of parameters to model the real world. The basic physics being used is F=ma, but many subtleties such as exact power curve, shift points, wheelspin, weight transfer, drive train interia, etc. added to the equations. "Curve fitting" is categorically false.

Originally Posted by CanAutM3

I don't believe the sub-models within CarTest are precise enough and well calibrated enough to do more in depth analysis. For example, the tire grip/wheel spin sub-model is certainly not optimal since it grossly overestimated the low speed (off the line) acceleration of the F8X. I also remember that you mentioned that CarTest uses a linear regression relative to engine speed for drivetrain losses, this is also inaccurate. IIRC, you also mentioned that CarTest uses fixed/constant (not user input) inertia numbers for the engine an drivetrain components, which also introduces an element of error.

No simulation is perfect and CarTest isn't either. That though in no way whatsoever justifies "throwing out the baby with the bathwater". It has enough of the basics and advanced functionality/capability to perform very accurate overall simulations. It certainly has the ability to precisely test things like gear ratios. Again arguing from isolated cases of specifics to then make an overall conclusion simply is not valid. I had hoped to not get into specific as we've done so before but it is in part warranted.

1. CarTest wheelspin models are not perfect. Then again none are. It overestimates tractive force under wheelspin for a small faction of second at launch and perhaps during wheelspin from shift to shift. The only inaccuracies from this are the actual peak values of acceleration vs. time (g foces) curves over a very small duration in time. Also 60' times are underestimated by perhaps a couple of tenths of a second. Uh oh, the sky is falling... Not.

Relevance to CarTests ability to predict overall performance and performance changes from gear changes. Almost nothing. Well unless one is drag racing professionally...

2. Losses: Losses are not "linearly interpolated". Losses are a function of both rpm and speed which nearly all sliding rotating frictional losses are fundamentally known to be. This comes from both P=F*v and the fact the F=-c*V (or their equivalent rotational analogs). You have a purely speculative doubt here not consistent with the literature nor demonstrated predictive capability.

Relevance to CarTests ability to predict overall performance and performance changes from gear changes. Almost nothing.

3. Drivetrain inertia: It appears some component use fixed inertias. I do not have all of the details. However, this is certainly more accurate than using mass factors. I've played with the latter in gory details using my own simulator and it just is not as accurate.

Relevance to CarTests ability to predict overall performance and performance changes from gear changes. Almost nothing.

Originally Posted by CanAutM3

That being said, for the topic at hand, wider gear spacing in the upper gears (3-7) would have made for a more entertaining and engaging powertrain where (at least) 7300RPM shifts would have been needed to extract the maximum performance of the powertrain (rather than having one than can be short shifted). This would yield the same result on the S55 as the "in gear torque limting" Ferrari approach. That was the point of my reply to Boss330.

You have already given it one shot with some gearing you said would be better. Unfortunately, you can't draw curves vs. rpm ignoring drivetrain inertia, losses (drivetrain and tires), wheelspin and aerodynamics and make any reasonable predictions whatsoever. You simply MUST work in the (real) time domain and you must work across gears. This then gives something that can be compared to real world observations.

You can continue to nitpick CarTest but it has been well validated, both by me and by others like regular_guy. You even like to selectively quote the excellent agreement between the CarTest results for the M4 and observed or quoted figures for the car. What is it, works or doesn't. I guess only works when personally convenient for you?

The options here are quite limited...

A. Build modified transmissions and test them yourself. Obviously not going to happen. Even if you could, some significant repetition would be needed as changes would be fairly minor.
B. Perform some realistic simulations that work in the time domain and cover the basics mentioned above. This is highly non-trivial. Again I have written my own simulator and even with 10 or so Excel sheets, thousands of lines and dozens of formula, I can barely match CarTest and when I throw it a "curve ball" like an gutless old car it doesn't exhibit the robustness of CarTest.
C. Get over the nitpicking with a tool that clearly has the plenty of capacity and accuracy to answer these very simple questions. Give me some "improved" gear ratios and I will be happy to simulate them for you, again proving my point - there is little to no margin left on the table here by BMW.
D. Buy CarTest yourself, it's not expensive. I'd be happy you help you through the learning curve with it.

Ugh...

[swamp2]

Originally Posted by myzmak

I think you just need to read this forum for a spell to see why.

In the olden NA days, to squeeze power out of engines it required high revs. That sound/phenomenon led people to associate the sound with 'speed' and 'fun' (the proverbial dinner bell for Pavlov's dog).

Turbos block the NA sound and don't require the high revs. So, people who were brought up to think 'high rev and only high rev = fun/excitement' have a hard time adjusting. As I've said elsewhere, it is all nostalgia. You remember your first sexual experience in detail, not because it was the best but because it was formative. We always measure the present against those first experiences. With NA being the 'standard' until recently, lots of folks have a hard time letting go and embracing the new.

So, some manufacturers try hard to give those folks some of what they think they want (NA 'sound' or 'power curves') while actually doing what they need to do in this day and age (FI, lightness, etc).

But, as anyone who isn't Pavlov's dog will tell you, you don't need a f------g bell to eat supper. You can still eat - and eat just as well or better - with a different noise.

Calling 8000-9000 rpm production vehicle engines (less the famous Honda S2000) nostalgia is pretty well nonsense. Most of the highest performing and most rewarding production sports cars of today and of all time are NA. That being said holdouts like Ferrari are certainly going down the path to green... Perhaps in a couple more generations (car generations) say maybe 10 years the term nostalgia will be more appropriate.

Furthermore denying the importance of vehicle acoustics is borderline delusional. Why do you think the new BMW Ms have the acoustics enhancement nonsense. Well the more I understand it the less I believe it is nonsense. Fake yes, but useless or poseur, nope. It is to make one able to better interpret what the darn engine is actually doing. Acoustics is important both emotionally and practically. I suppose you can also go tell all F1 fans and ever their CEO also that acoustics "doesn't matter".

Although the Pavlovian analogy is clever, I almost want to agree with it, it's still nonsense in my humble opinion.

[Soorena]

Another day, another pissing contest in M3P. Nothing changes.

[myzmak]

Originally Posted by swamp2

Calling 8000-9000 rpm production vehicle engines (less the famous Honda S2000) nostalgia is pretty well nonsense. Most of the highest performing and most rewarding production sports cars of today and of all time are NA. That being said holdouts like Ferrari are certainly going down the path to green... Perhaps in a couple more generations (car generations) say maybe 10 years the term nostalgia will be more appropriate.

Furthermore denying the importance of vehicle acoustics is borderline delusional. Why do you think the new BMW Ms have the acoustics enhancement nonsense. Well the more I understand it the less I believe it is nonsense. Fake yes, but useless or poseur, nope. It is to make one able to better interpret what the darn engine is actually doing. Acoustics is important both emotionally and practically. I suppose you can also go tell all F1 fans and ever their CEO also that acoustics "doesn't matter".

Although the Pavlovian analogy is clever, I almost want to agree with it, it's still nonsense in my humble opinion.

You are misinterpreting some of my post.

I didn't say there was no benefit to a high revving motor. My point was this is no longer 1980 when the *only* motors with real benefit were high revving. Used to be you needed one (pavlov's dog only ate when the bell went). But now, you don't.

Now you can have a very good (equally good? Better?) experience with FI. Or with other technologies (crazy hybrids ala P1 etc). The world no longer relies on one setup to make for great driving.

My point was people who were raised to believe that ONLY high revving is good will have a tough time letting go of that. Perfectly natural reaction, but that doesn't make it right.

Also didn't say sound was of no consequence. I do think the high rev crowd overplays it a bit with 'oh the sound! The sound!' dramatics but absolutely feedback under load is important to the driver. I am ok with the active sound because I expect 99.9% of drivers couldn't tell it is amplified and wouldn't know it was if they hadn't been told. Sadly, all the other exterior sound dampening technologies (so we don't hear road noise, etc) make it necessary.

Lastly, it isn't just about going green, as though everyone is becoming a tofu eating hippie. The costs to car companies of consumption (in USA) and C02 (in Europe) are very significant. No way corporate management can justify spending hundreds of hundreds of millions of dollars in fines to keep NA when they can do as much or more with different tech. (And I'm not just talking M and AMG here. There is a reason you see so many more diesel cars in North America now....). For better or worse, like getting rid of leaded gas and putting noise limits on cars, it is the way of the world.

What good car cos are doing is not pining for love lost, but figuring out how to make the brilliant car of TODAY, not 1955, 1975 or even 2007. Here and now.


Edit: on topic, an article from Globe and Mail on tomorrow's Canadian GP. Interestingly, Jacques Villeneuve (the track being named after his father and being a former F1 champ himself, so someone who knows something) likes the sound, though some dude from my town (who likes Harley's....nuff said) doesn't.

http://www.theglobeandmail.com/sport...ticle19060380/

It isn't as black and white as suggested and, as noted, the genius in all of this is what teams do with the new world and how they win.

[myzmak]

Originally Posted by Soorena

Another day, another pissing contest on the internet.
Nothing changes.

Fixed it.

[CanAutM3]

Originally Posted by myzmak

You are misinterpreting some of my post.

I didn't say there was no benefit to a high revving motor. My point was this is no longer 1980 when the *only* motors with real benefit were high revving. Used to be you needed one (pavlov's dog only ate when the bell went). But now, you don't.

Now you can have a very good (equally good? Better?) experience with FI. Or with other technologies (crazy hybrids ala P1 etc). The world no longer relies on one setup to make for great driving.

My point was people who were raised to believe that ONLY high revving is good will have a tough time letting go of that. Perfectly natural reaction, but that doesn't make it right.

Also didn't say sound was of no consequence. I do think the high rev crowd overplays it a bit with 'oh the sound! The sound!' dramatics but absolutely feedback under load is important to the driver. I am ok with the active sound because I expect 99.9% of drivers couldn't tell it is amplified and wouldn't know it was if they hadn't been told. Sadly, all the other exterior sound dampening technologies (so we don't hear road noise, etc) make it necessary.

Lastly, it isn't just about going green, as though everyone is becoming a tofu eating hippie. The costs to car companies of consumption (in USA) and C02 (in Europe) are very significant. No way corporate management can justify spending hundreds of hundreds of millions of dollars in fines to keep NA when they can do as much or more with different tech. (And I'm not just talking M and AMG here. There is a reason you see so many more diesel cars in North America now....). For better or worse, like getting rid of leaded gas and putting noise limits on cars, it is the way of the world.

What good car cos are doing is not pining for love lost, but figuring out how to make the brilliant car of TODAY, not 1955, 1975 or even 2007. Here and now.


Edit: on topic, an article from Globe and Mail on tomorrow's Canadian GP. Interestingly, Jacques Villeneuve (the track being named after his father and being a former F1 champ himself, so someone who knows something) likes the sound, though some dude from my town (who likes Harley's....nuff said) doesn't.

http://www.theglobeandmail.com/sport...ticle19060380/

It isn't as black and white as suggested and, as noted, the genius in all of this is what teams do with the new world and how they win.

I think Villeneuve was a bit sarcastic, he says the sound is good because you can bring your kids to the race without fear of damaging their hearing and that you could have phone conversation during the race. I don't read this as praise for the new sound...

[myzmak]

Originally Posted by CanAutM3

I think Villeneuve was a bit sarcastic, he says the sound is good because you can bring your kids to the race without fear of damaging their hearing and that you could have phone conversation during the race. I don't read this as praise for the new sound...

Yeah, not the spin the article gave it, but I see how you can read that... (frankly, as a parent and a fan, I don't think not deafening my kids is the worst thing in history.....)

Either way, noise aside, this is the most exciting F1 season in many years (what a finish yesterday!) The challenge the something new brings (ie: new engines, torque not just HP) is surely a major contributor to that.

[CanAutM3]

Originally Posted by myzmak

Either way, noise aside, this is the most exciting F1 season in many years (what a finish yesterday!) The challenge the something new brings (ie: new engines, torque not just HP) is surely a major contributor to that.

Agreed

[swamp2]

Originally Posted by myzmak

You are misinterpreting some of my post.

Fair enough. Thanks for the well stated reply, most of which I agree with.

[myzmak]

Originally Posted by swamp2

Fair enough. Thanks for the well stated reply, most of which I agree with.

No worries man - always nice to have dialogue on here!

[CanAutM3]

Note that my original intervention in this thread was only to point out that BMW could have provided a more engaging powertrain by giving us wider spaced gear ratios in the higher gears that would require higher RPM shift points.

That being said

Originally Posted by swamp2

You even like to selectively quote the excellent agreement between the CarTest results for the M4 and observed or quoted figures for the car. What is it, works or doesn't. I guess only works when personally convenient for you?

No simulation is perfect and CarTest isn't either. That though in no way whatsoever justifies "throwing out the baby with the bathwater".

I certainly am not "throwing out the baby". I did say that CarTest provided reasonably accurate prediction of overall acceleration. It is a very good tool when there are no or little actual test samples out there. Or to validate if a given test data set makes sense.

Originally Posted by swamp2

But it does basically come down to power, no more power no more performance.

Completely agree.

Originally Posted by swamp2

Whether it is FD or individual gear ratios. Ratios have very little effect on overall performance

This is where I dissagree. If, by optimizing ratios, the powertrain is able to produce more average power during an acceleration run, especially early in the acceleration run, it will provide an advantage in terms of time to distance travelled.

Originally Posted by swamp2

You have a massive and fundamental misunderstanding of what CarTest does according to the statement above. There is no curve fitting. CarTest uses basic vehicle dynamics, similar to that in "Fundamentals of Vehicle Dynamics" by Gillespie, accounting for an enormous range of parameters to model the real world. The basic physics being used is F=ma, but many subtleties such as exact power curve, shift points, wheelspin, weight transfer, drive train interia, etc. added to the equations. "Curve fitting" is categorically false.

I think I do understand how CarTest works. I do understand that it is a physics based model. I understand that it relies on a series of sub-models to modelize tire grip, rolling resistance, inertial impacts, aero drag, drivetrain losses, etc...

As with any analytical model, it needs to be calibrated by adjusting some assumed parameters to yield representative results. The various sub-models therefore need to be calibrated for the overall model to function.

My point about "curve fitting" touches the calibration of the individual sub-models. An error in calibration of one sub-model can be made up by the error of another sub-model. By adjusting the various parameters, the overall model can be made to closely match real world figures even if the sub-models are not fully accurate individually. Hence my "curve fitting" reference.

For example, if Cartest overestimates the grip at the launch of a quarter mile run (as we know it does), to be able to match a real world ET, it has to slow the car down elsewhere in the model (drivetrain losses? aero drag? rolling resistance?).

As another example, if the losses in DCT are underestimated relative to 6MT, introducing shift time for the DCT (period of time with no tractive force during the shift) can bring the numbers back in line.

Originally Posted by swamp2

2. Losses: Losses are not "linearly interpolated". Losses are a function of both rpm and speed which nearly all sliding rotating frictional losses are fundamentally known to be. This comes from both P=F*v and the fact the F=-c*V (or their equivalent rotational analogs). You have a purely speculative doubt here not consistent with the literature nor demonstrated predictive capability.

Bad choice of words on my part. By linear regression I did not mean interpolation. IIRC from previous disscussions, you mentioned that Cartest works with a formula that linearly reduces drivetrain losses with reduced engine RPM:

Originally Posted by swamp2

Let me share the way both CarTest and I have implemented this. The value one enters as a fixed percent, say 15% (loss) is scaled according to rpm and redline. Let's say we call the 15% number the loss, ℓ

Thus ηtf = (1 - rpm x ℓ/redline) (efficiency = 1 - loss)

This is simply not accurate. Torque transmission efficiency depends on a variety of parameter including load and speed. Just check any chassis dyno chart and compare it with manufacturer torque charts, the efficiency does not improve that much with lower engine speeds.

Further, most of the drivetrain rotates at a speed that is constantly proportional to the drive wheels, not engine speed. Only part of the gearbox varies with engine speed.

Originally Posted by swamp2

Relevance to CarTests ability to predict (...) performance changes from gear changes. Almost nothing.

Also have to dissagree here. If the sub-models are not accurate, it can have a non-negligible effect on the impact of gear ratios.

For example, if Cartest overestimates tire slip traction at launch, it will not see the benefit of a longer first gear.

As another example, if Cartest underestimates the impact of losses at lower RPM, it will also underestimate the power gains that could be made by keeping the engine closer to the power peak at higher RPM. Same applies to inertia factors.

Originally Posted by swamp2

You have already given it one shot with some gearing you said would be better.

I remember those gear ratios very well (they are the only ones I ever provided). They were not established for optimal acceleration, but rather for redline shifts. They were based on fictive power/torque curves before the official BMW numbers were out. Any exercise with those ratios is utterly useless.

Originally Posted by swamp2

Give me some "improved" gear ratios and I will be happy to simulate them for you, again proving my point - there is little to no margin left on the table here by BMW.

As I stated in the other thread, I don't have the car data nor the analytical models to attempt this exercise. And if I would, based on my points above, I would certainly not trust Cartest with validating them.

Originally Posted by swamp2

D. Buy CarTest yourself, it's not expensive. I'd be happy you help you through the learning curve with it.

I am very tempted to, since I like this kind of stuff. But I know myself though, and I am concerned it will suck up too much of my time and get the wife upset .

Originally Posted by swamp2

Get over the nitpicking with a tool that clearly has the plenty of capacity and accuracy

Ugh...

Don't be too exasperated with me. I spent a big portion of my engineering career proving analytical models wrong. It is just professional deformation if I have my doubts about any analytical model