I've got to say, I'm expecting fairly drastic aero revisions/improvements.
I remember when we got to the bottom of what caused the high nose exploit, the impression I got from our conversations with Scawen was that he wanted to produce a much more complete aero simulation before revising the physics.

Just fixing the high nose bug would literally have been the work of an afternoon, if that. It's just a case of adding a calculation step to the way the force was resolved. But if they released a new physics patch every time a little bug was fixed, or littel feature added, the community would be a mess.

So I think we'll have much more sophisticated aero, with proper undertrays that react to the undercar distance and rake, as well as more complete and accurate wing simulation.

Its a good point. Although I've found an even better workaround is to start a real multiplayer internet game but set it to private. That way all your lap times get recorded in LFSWorld for handy reference.

I've been patient for 5 minutes, but it still aint moving. Ever heard of AVI's?

Quote from PLAYLIFE :

On the cover it says "West brothers talk", yet I could not find it in the mag. Which page is it??

Its some completely ureadable drivel written in deep south dialect hidden in the news section (page 11). As far as I could tell it contained no information whatsoever.

I love the principle of ASS and the fact that people invest time in producing it, but I really wish the writing was better, and shorter. Downside of PDF over print is that people waffle on for pages and pages about nothing and no-one ever seems to edit it.

Quote from Becky Rose :

It's most likely that the in-house tools comprise a 3d format exporter for their chosen modelling program and possibly a small utility for rigging the 3d model and setting collision zones and game data interactive elements (such as 'if a wheel touches here make a sand particle emit from the wheel' [that might also be zone with heirarchy data]) and editing the basic AI driving line.

As far as we understand, LFS is made entirely with custom built tools. A few times Scawen has referred to spending some time working on functions for the modeller for Eric.

So for them to release tools to the public, they'd either have to write new converters/utils from scratch or brush up the in house tools to a release standard. Neither of which are small jobs.

The other option of course is to just release the file format specifications along with S3 and let the community write the tools.

Yeah, the timing display options do need a complete overhaul. you should be able to get a +/- readout from your best lap during races/quali at each split, like every other sim ever.

I dunno about today, but I'm fairly sure there was an era when formula one engines could be either 3 litres or 1.5 litres with a turbo, and plenty of teams went with the crazy turbos instead of the 3 litres.

Indeed, or even any raciunmg.

Um, you obviously speak as a viewer rather than a competitor.

The guy in front doesnt really care about 'interesting' racing, he cares about winning.

Quote from skiingman :

Yeah, so I'm definitely NOT a NASCAR fan, but this is just wrong.

You need to do just that, except with 800bhp and a live rear axle and unsettling transitions from asphalt to concrete, all with 42 other cars nearby. Read: Bristol/Martinsville.

Fair enough, then I stand corrected, I wasnt aware there were oval tracks out there with hairpin bends and spoon curves.

Well clearly, the guys at the top in NASCAR and oval racing generally are enormously skilled at the range of skills that oval racing requires, and we should respect that.

It just seems to me that the actual range of skills is smaller and different, rather than a lower level of skill being required. Clearly in oval racing, you dont need to be able to trail brake from 150-50mph into a decreasing radius turn over a crest without losing the back end; you dont need to be able to put down 700bhp while exiting a 30-40mph tight bend twice a lap without spinning; you dont need to be able to outbrake another car into a sweeper, or ride kerbs through a chicane at high speed. etc.etc.

Of course, you need need to be able to ride at high speed, surrounded by other cars, maintaining peripheral awareness, controlling the car and formulating overtaking strategy all at the same time, which no doubt takes great skill and concentration.

Quote :

This is all I've ever been trying to say!

To quote my dear tutors at uni "So why didn't you then?"

I wouldnt bother really, I thought we'd already covered those things.

I would just say that maximising the area under a wheel torque/time curve, and maximising the area under a power/time curve (assuming freedom to gear appropriately), are effectively the same thing.

So really, some of you are arguing that the answer is 10+10, and the others that the answer is 40/2.

So, y'know, let it lie.

Blimey. Here we go again. I'll be honest and say I didnt have the patience to read all those posts in detail, but it just seems to me that everyone is coming to a roughly equivalent understanding, but are still arguing about minutae that got posted ages back.

I also think its difficult to have a meaningful debate because the definitions haven't been sorted. Also, I think that for myself as well as others, existing pieces of knowledge are preventing us from properly assimilating other bits of knowledge, and also we are making statements that apply to a particular context without necessarily qualifying that context.

When I step back from cars and engines and gearings for a moment, and just think about raw physics, it all seems blindingly obvious.

If over period of time t, an average power of p is expended, the amount of kinetic energy added to the object will be p*t.
So any change in the average power over the period of time will result in a proportional change in the amount of kinetic energy added.

So lets ignore transmissions or drivetrain losses etc. and say that our CVT produces a dead flat power curve (line?) at max power, so the average power for the time t = max power.
In any manual transmission, the power curve will rise and fall depending on gear changes, but will always be below or equal to max power, so average power over time t < maxpower.

So that means that after time t, less kinetic energy has been added in the second example, which means that any time our engine is below maximum power, we are adding less kinetic energy.

Now i'm starting to state the obvious now, but if we have initial velocity u, our final velocity v will always be higher when more kinetic energy has been added. And if our final velocity is higher, we have 'accelerated better'.

So, for a given engine, with complete freedom to gear appropriately, maximum final velocity will be acheived when the engine spends as much time as possible as close as possible to maximum power.

Just for completeness, I going to make a few statements that have come up in the debate, together with the context needed to make them correct.

1. For a given fixed gear ratio, the highest instantaneous acceleration acheivable in that gear will occur at the moment when the engine is producing maximum torque.

2. For a given road speed, the highest instantaneous acceleration acheivable at that speed will occur if the gearing is such that the engine is producing maximum power.

3. Over a given period of time, ignoring any gearing limitations, the greatest overall increase in velocity will occur when the area under the power/time graph is maximised.

And thats it from me.

Although on the plus side, LFS is the only sim I've ever played that gives you ALL of its setup parameters in proper, physical, SI units, so you can actually make meaningful calculations.

Well when you say an object is accelerating at 10 m/s/s you are only referring to its acceleration at that moment in time (or to put it another way, the gradient of the velocity/time graph at a given point).
You arent saying that it will continue to accelerate at the same rate at any point in the future. But i suppose the units suggest a kind of prediction. An acceleration rate says "at this rate of acceleration, in 1 seconds time, this object's velocity will have changed by this much".

Imagine, a graph of speed/time. If your acceleration rate is constant, the graph will simply be a straight line going off at an angle to the top right. The gradient of that straight line is your constant acceleration.

Now imagine a real life speed/time graph like you'd see in F1 Perfview, it starts off kind of steep, when your acceleration is high, but the faster you go, the more it flattens off, because your acceleration is lower.
At any point on that graph, the gradient of that point is the current rate of acceleration at that moment in time.

This is interesting stuff actually, you're kind of discovering the edges of calculus by yourself.

Quote from Hyperactive :

What is this? Who writes the longest post -competition .

So when do I see the power and torque curves of the GTRs?

If you download Bobs GRC2 you can see some estimated torque curves. But we dont really need them.

Basically, the XRR acheives its peak power by producing more torque at lower revs (6278 rpm), and its peak torque point is only 1500 revs lower. So the torque curve is quite peaky, as you'd expect from a turbo.

The FZR acheives peak power by producing less torque but at higher revs (8000+ ), but its torque peak is a full 3000 revs lower, which means basically that it has a wider powerband. So, because the power comes on earlier in the rev range, over the course of an acceleration, the area under the graph is greater, so the car covers the distance in less time.

Another factor to think about is because the FZR has a higher rev limit, it can always be in a slightly lower gear than the XRR and FXR, which will compensate for its slightly lower torque output.

It seems to me that if physics processing has any chance of becoming mainstream then it has to get incorporated into gaming graphics cards, and a standard API for it incorporated into DirectX 11 or whatever.
Nobody in there right mind is going to buy another addon card for a miniscule number of games from developers crazy enough to support a miniscule number of users.
If its a gaming related function, which it is, then it should be incorporating into our gaming addon cards, and then a generation or two down the line, developers can assume it will be present in the same way they can just assume that pixel and vertex shaders are present now.

Wow, this is getting confusing. BBO, you seem think that I've 'gone over to the torque side'. Not at all, i'm just trying to explain the way the two things interellate.

Your expression of it in terms of work is just as correct as my expression of it in terms of the aggregate of instantaneous forces over time, the two methods of explanation are not mutually exclusive.

I agree with Skiing man that peak power is a much more useful and relevant figure than peak torque in estimating the performance potential of an engine, but I'll repeat till I'm blue in the face that the torque/power curve is the be all and end all descriptor of the performance of an engine, and once you have that, all other relevant information can be derived (in theoretical performance terms at least, it wont tell you if the thing drinks oil, or blows up if you dont get it serviced every 1000 miles ).

Last edited by colcob, Wed 22 Mar 2006, 22:17.

Actually, having read through both your posts again, I think you just both fundamentally dont get it, but in different ways.

Oh I give up.

BBO, I was with you up to a point, but obviously you havent studied maths or physics to a high enough level because although your instincts are good, you're struggling with the maths.

The concept of rate of change in an instant is fundamental to the entirety of calculus. In a sense you've been quite astute because the instant is actually defined as an infintestimal quantity of time. But if you take a graph, the gradient of a graph at any point is its rate of change.

Clearly you can point to a discrete spot on a curve and say that it has a gradient. This is what calculus is all about in a way, its a mathematical method of determining what the gradient is at a given point on a curve (and loads of other things as well, but lets not complicate it).

In the question of the CVT transmission, you are making the same mistake as Tristan in confusing the force from the engine with the force applied at the tyres.
A car accelerates ONLY by applying force to it. Acceleration is directly proportional to force (a = F/mass). So the torque applied at the wheels is the only determining factor (ignoring mass) as to how fast the car will accelerate at a given moment.

The torque generated by the engine is multiplied by the gearing factor to give the torque at the wheels. So if an engine puts 100Nm at 1000RPM into a gearbox with a ratio of 1:2 (or 2:1, I get confused) the wheels will apply 200Nm of torque and turn at 500RPM.

So your CVT gearbox basically calculates the correct ratio to ensure that for the current road speed, the engine will be turning at maximum power RPM, and this will result in a higher torque at the wheels than would be the case at peak torque, because the gear is lower.

An example.

the FXO gives 225lb ft at 4338RPM, and 234bhp @ 6365.

Using the formula BHP = torque * rpm / 5252 (yes, torque and bhp are directly related)

Torque at peak BHP = 234 * 5252 / 6365 = 193 lb ft.

Lets say we are travelling at 25m/s (about 50mph) and the circumference of the wheel is 1.7m, that means the wheel is rotating at 15rps = 900 RPM.

If the CVT matches our speed to peak power revs, that means the gear ratio will be 6365/900 = 1: 7.07.

So if we multiply our torque at peak power (193 lb ft) by our gear ratio, we get:

torque at wheels = 1365 lb ft

If our CVT matches the speed to our peak torque rpm, the gear ration will be 4338/900 = 4.82.

So the torque at the wheels will be 4.82 * 225 = 1084 lb ft

Voila. The reason a CVT produces greater acceleration at peak power is because it lets you use a lower gear which gives greater torque at the wheels. This is fundamental even ignoring CVT's. The reason more power gives you better acceleration over time is because it lets you stay in a lower gear for longer, thus giving you more torque at the wheels for longer.


And finally, as the astute will have notices above, torque and power are directly mathematically related to eachother. power = torque * rpm*constant depending on units used.

So all this nonsense about how power is more relavant than the torque curve, or peak torque is more relavant than peak power, is all just tripe.

In the beginning there was the torque curve. A dyno measures torque. From the torque at each given RPM we mathematically derive the power at each given RPM using the simple formula above. The point at which peak power occurs is a product of the torque curve, or the other way round depending on how you look at it.

The point is that the two curves are not independent, their shapes are inexorably bound to eachother by maths.

I suppose that in theory, it would be possible to get the information out in real time with some memory sniffing techniques such as those used in LFSTweak. The downside being that every new patch you have to scour the memory for the right offsets all over again.

No. You change gear when the red light comes on. Simple as that.

LFS calculates when the next gear will give you more acceleration than the current one, and thats when the light comes on.

What is happening here is a clash of definitions.

There are two ways of defining acceleration in this debate.
The strictly scientifically correct definition, which is the rate of change of velocity at an instantaneous moment in time; and the more practical but semantically innacurate definition of 'how quickly can we go from point A to point B'.

In the first case, the highest instantaneous acceleration will always occur when the longitudinal force is highest, which is at the torque peak in any given gear. Of that there is no doubt.

But when people talk about the 'acceleration' of cars, they tend to be actually thinking of how long does it take to get from point A to point B, or even how long does it take to get from velocity A to velocity B.

In these cases we get back to thinking about the area underneath graphs, rather than the maxima of them.
If you take a graph of acceleration over time, the total area under that graph is equal to the average velocity over that time. So anything you do to increase the overall area under the tractive effort curve gives a faster average velocity over the course of the acceleration. Which means that your final velocity will be higher and your overall distance travelled will be greater.
Ergo, in laymens terms, the 'acceleration' is better.

So basically, the performance potential of an engine is dictated by the area under the torque curve. Now before someone jumps in and says 'but what about peak power', I should say that peak power is dictated indirectly by the area under the torque curve, the longer the torque curve stays above a certain point, the higher the peak power will be.

An excellent illustration of all these points is the restricted V10's that Torro Rosso are using in F1 this year. They have been rev and inlet restricted to give supposedly the same peak power as a V8, but they have a greater area under the torque curve due to those 10 cylinders pulling before the restrictor kicks in, which means they have a better performance potential than they should.

So in summary, acceleration performance is not dictated by peak power only, or peak torque only, but by the total area underneath the torque curve (within the rev range used by the current gearing setup).

OMG, KILL ME NOW.

Not this again, please for the love of all that is holy.

*col goes searching for the TORQUE VS HORSEPOWER THREAD OF DOOM.....*

(..and you're still wrong Tristan, a CVT car will accelerate quickest at peak power because by definition, the CVT will adjust the transmission to give greatest torque AT THE WHEELS. You're confusing engine torque with wheel torque again)

Last edited by colcob, Tue 21 Mar 2006, 20:46.