How to develop an F1 car?

[adam2007]

right example. when people say how can a rookie devlope a car i mean when he reports back to engineers saying can lacks rearend grip i mean what else can he say? if u get what i mean?

[Scotracer]

Most drivers and engineers split driving up into two or three segments:

-Turn in (transient)
-Mid-corner (steady-state)
-Exit (transient)

The cars handle differently in each section so just saying it lacks rear-end grip doesn't get to the heart of the matter and will just result in more questions. It could be an issue of not enough rear wing or the rear ARB/Diff settings could be off. Or the engine-braking map could be off if it's on turn-in. So giving a specific condition in a specific region of a corner helps a lot.

The more detail the driver can convey quickly, the better.

[Giblet]

A good driver can tell the difference between not enough grip in the rear, or too much in the front.

Both conditions would feel the same to the uninitiated. A driver like Alonso engineers love, because if a new part is put on the car, and he goes faster, that part is a keeper.

Poor input from a driver, and poor problem solving skills can quickly lead the engineers down a performance path that is sub par to the optimum.

Drivers tend to go for the best balance and feel, but the really good developers know what is quicker as well, not just better for the balance.

Also, if the driver says one thing, and the telemetry says another, then what good is he?

EDIT: AM I the only one here with a spell checker?

Let's AT LEAST check the spelling in our topic subjects. Devlope. Heave.

[Jersey Tom]

Good, accurate, concise driver feedback makes the chassis tuning process MUCH easier and faster for the engineer, data or not. Consequently, inaccurate subjective feedback will screw you good.

"Car is oversteer" is crap feedback.

"Car is oversteer on entry" even is pretty much crap.

Good driver needs to identify what the car's doing, when, and why. For example:

"Car becomes free on entry from forward weight transfer under the brakes... mid corner balance is neutral with smooth off-to-on throttle transitions.. car tightens up slightly on-throttle until rear end breaks free. Car turns-in well on low speed corners in general, but is a little over responsive during small steering corrections at high speed"

Would be a start.

[Shrek]

i know in both dirt and circle track the driver usually says something like: "im a (1-10) loose (oversteer/push) in entry of turn 3 and a (1-10) tight (understeer) on the exit of turn 4" to get an even accurate feedback.

[speedsense]

right example. when people say how can a rookie devlope a car i mean when he reports back to engineers saying can lacks rearend grip i mean what else can he say? if u get what i mean?

As data acquisition has grown in the sport, there is much less importance on driver feedback, as all of the driver movements are tracked. The car's handling is tracked and handling problems are pinpointed, and don't have to be remembered by the driver. In fact with live telemetry, a driver/car is pretty much analyzed before the car returns to the pitlane. Even when a driver's comments are opposite of what the data says, the engineers will follow what the data says, rather than the driver because the drivers brain isn't physically capable in memory or speed of current data acquisition collection.

For instance, the driver movements (all of them) are recorded faster than the driver can react (20hz-50hz) and car's movement are recorded faster than the car can physically move. (100hz-1000hz).
A driver may make 300 decisions a lap from which only 10% percent of them will be remembered. A good driver with excellent feedback can remember some 20-30 decisions of a lap. The data system is capable of seeing them all, when the analysis (the data geek(s) ability to analyze) is capable of tracking them. Even current simulation programs which are run with current data, include driver analysis.

About the only thing left up to the driver is whether he likes the car or not and that the car can do the things that the driver demands of it. About the only thing that can't be tracked is how the driver "likes" the car or his/her desires.

It's not like the old days where it was up to the driver to define the direction of the car and it was important that a driver had really good feedback so that the car would be setup correctly. Now a days, the engineers, designers and setup engineers have a 100 times the information to replace driver feedback.

A rookie coming into F1, doesn't need feedback or setup skills, only driving skills, if the data is analyzed well. Drivers feedback is no longer a requirement for employment and is infact way down the list of abilities from where it used to be, in the pre-data era. IMHO

[Jersey Tom]

right example. when people say how can a rookie devlope a car i mean when he reports back to engineers saying can lacks rearend grip i mean what else can he say? if u get what i mean?

As data acquisition has grown in the sport, there is much less importance on driver feedback, as all of the driver movements are tracked. The car's handling is tracked and handling problems are pinpointed, and don't have to be remembered by the driver. In fact with live telemetry, a driver/car is pretty much analyzed before the car returns to the pitlane. Even when a driver's comments are opposite of what the data says, the engineers will follow what the data says, rather than the driver because the drivers brain isn't physically capable in memory or speed of current data acquisition collection.

For instance, the driver movements (all of them) are recorded faster than the driver can react (20hz-50hz) and car's movement are recorded faster than the car can physically move. (100hz-1000hz).
A driver may make 300 decisions a lap from which only 10% percent of them will be remembered. A good driver with excellent feedback can remember some 20-30 decisions of a lap. The data system is capable of seeing them all, when the analysis (the data geek(s) ability to analyze) is capable of tracking them. Even current simulation programs which are run with current data, include driver analysis.

About the only thing left up to the driver is whether he likes the car or not and that the car can do the things that the driver demands of it. About the only thing that can't be tracked is how the driver "likes" the car or his/her desires.

It's not like the old days where it was up to the driver to define the direction of the car and it was important that a driver had really good feedback so that the car would be setup correctly. Now a days, the engineers, designers and setup engineers have a 100 times the information to replace driver feedback.

A rookie coming into F1, doesn't need feedback or setup skills, only driving skills, if the data is analyzed well. Drivers feedback is no longer a requirement for employment and is infact way down the list of abilities from where it used to be, in the pre-data era. IMHO

Completely untrue, IMHO.

I spend a good chunk of time looking at state of the art racecar telemetry. Upwards of a million dollars on some cars.

Some data is essential and lives by itself. Potentiometers and accelerometers used to develop 7-post drive files for example.. are their own thing that are independent of the driver for the most part.

A lot of handling though is very subjective to the driver. Some stuff you can see in the data, some you can't.

Even something as seemingly simple as 'oversteer' can be incredibly difficult to really define accurately, in relating what the data shows to what the driver says.

[Scotracer]

Good, accurate, concise driver feedback makes the chassis tuning process MUCH easier and faster for the engineer, data or not. Consequently, inaccurate subjective feedback will screw you good.

"Car is oversteer" is crap feedback.

"Car is oversteer on entry" even is pretty much crap.

Good driver needs to identify what the car's doing, when, and why. For example:

"Car becomes free on entry from forward weight transfer under the brakes... mid corner balance is neutral with smooth off-to-on throttle transitions.. car tightens up slightly on-throttle until rear end breaks free. Car turns-in well on low speed corners in general, but is a little over responsive during small steering corrections at high speed"

Would be a start.

Oh oh oh I want to diagnose...

The front springs hardened to reduce weight transfer. Actually, since the rear breaks quite early you could soften the rears to achieve the same whilst improving rear traction. Reduce front wing to improve high-speed responsiveness. With the softened rear springs a possible softening of front ARB to counteract the possible increase in understeer.

amirite?

[speedsense]

right example. when people say how can a rookie devlope a car i mean when he reports back to engineers saying can lacks rearend grip i mean what else can he say? if u get what i mean?

As data acquisition has grown in the sport, there is much less importance on driver feedback, as all of the driver movements are tracked. The car's handling is tracked and handling problems are pinpointed, and don't have to be remembered by the driver. In fact with live telemetry, a driver/car is pretty much analyzed before the car returns to the pitlane. Even when a driver's comments are opposite of what the data says, the engineers will follow what the data says, rather than the driver because the drivers brain isn't physically capable in memory or speed of current data acquisition collection.

For instance, the driver movements (all of them) are recorded faster than the driver can react (20hz-50hz) and car's movement are recorded faster than the car can physically move. (100hz-1000hz).
A driver may make 300 decisions a lap from which only 10% percent of them will be remembered. A good driver with excellent feedback can remember some 20-30 decisions of a lap. The data system is capable of seeing them all, when the analysis (the data geek(s) ability to analyze) is capable of tracking them. Even current simulation programs which are run with current data, include driver analysis.

About the only thing left up to the driver is whether he likes the car or not and that the car can do the things that the driver demands of it. About the only thing that can't be tracked is how the driver "likes" the car or his/her desires.

It's not like the old days where it was up to the driver to define the direction of the car and it was important that a driver had really good feedback so that the car would be setup correctly. Now a days, the engineers, designers and setup engineers have a 100 times the information to replace driver feedback.

A rookie coming into F1, doesn't need feedback or setup skills, only driving skills, if the data is analyzed well. Drivers feedback is no longer a requirement for employment and is infact way down the list of abilities from where it used to be, in the pre-data era. IMHO

Completely untrue, IMHO.

I spend a good chunk of time looking at state of the art racecar telemetry. Upwards of a million dollars on some cars.

Some data is essential and lives by itself. Potentiometers and accelerometers used to develop 7-post drive files for example.. are their own thing that are independent of the driver for the most part.

A lot of handling though is very subjective to the driver. Some stuff you can see in the data, some you can't.

Even something as seemingly simple as 'oversteer' can be incredibly difficult to really define accurately, in relating what the data shows to what the driver says.

My career of some twenty years in pro racing, is looking at data. And yes you can clearly define oversteer (infact it's the easiest to define, one simple way is comparing steering to lateral g, though many more ways are possible) and understeer (much harder to define, as each driver the amount of it is different according his taste).
Techniques have developed, as has the software, that clearly define and quicken the analysis process. It is no longer just reading an absolute number. Comparitive and Separation analysis (seperating the driver from the car) makes tracking the driver technique very possible and comprehensive. Even the smallest technique change, such as how the driver "rotates" the car, can be measured, analyzed and calculated into the setup change.
Data Acquisition has become a hundred times more than just a numbers tool as it was some twenty years ago.

Micheal Schumacher, how the importance and growth of data helped him and how his record may get broken from the new drivers that have grown up around and using data acquisition (around 1:40 he talks about it)
http://www.youtube.com/watch?v=nGDgbJLc0U0

[speedsense]

Completely untrue, IMHO

I spend a good chunk of time looking at state of the art racecar telemetry. Upwards of a million dollars on some cars.

It's not how much you have, but how you use what you have. In F1, their data personel is divided into a fleet of engineers for specific areas of the car for clarity on the many meg's of data extracted from each session.

Some data is essential and lives by itself. Potentiometers and accelerometers used to develop 7-post drive files for example.. are their own thing that are independent of the driver for the most part.

I agree, on a seven poster, they are indepent of the driver though what's driving the seven poster? One thing you cannot tell on a seven poster is whether the movements are caused by driver input, car input or track input, except in some cases where it's obivious.
On a car, they are used to define not only chassis movements but also whether the driver is inducing a movement or the car is or if the track itself is causing it. The separation of these causes is possible in on board data and can be clearly defined as so.

A lot of handling though is very subjective to the driver. Some stuff you can see in the data, some you can't.

The only constraints are the causes of handling situations in data and whether the software package can make it possible to define, by separating the causes from each other. You can see each and every movement and it is only the ability to see the cause and it's source (driver, car or track). Very few instances, cannot be defined this way, though those tend to be beyond the scope of the driver's feel or recall. A data acquisiton software package, is capable of analysis of the car, driver and track details at a speed and comprehensive that the driver cannot physically comprehend. That is, provided the person doing the analysis has the experience to extract it.

Even something as seemingly simple as 'oversteer' can be incredibly difficult to really define accurately, in relating what the data shows to what the driver says.

(previous post)


AA IMHO...

[Jersey Tom]

Good data is essential for chassis and "race package" development (e.g. 7-post testing), but you're not going to win championships with it by itself. You need a good driver that gives good feedback, coupled with a good team of engineers.

I don't think 'oversteer' is as simple as you put it, particularly since 'oversteer' or a 'loose' racecar can encompass a lot of things when the driver communicates it. The Ackermann corrected steering angle thing is hokey at best, and pretty poor IMO aside from some specific instances or as a really quick and dirty tool.

For example, if the driver has the car mid corner and is slightly sawing at the wheel, your data trace for "oversteer/understeer" will be oscillating just as much as the steering trace is, even though the tires are saturated and the trim of the vehicle isn't really changing.

Or, in some forms of racing you might have an asymmetric vehicle setup where the car is at some non-zero sideslip angle while going down the straight, even so much that the wheel is significantly off center. In that event you have "static" oversteer even if the car is going down a straight line.. which is kind of an odd concept.

Or there's the question of separating 'text-book' slip angle oversteer, from just high yaw attitude, to over-rotation. Which is most important for the driver? How do you even accurately measure accelerations in the "ground" plane when you have chassis roll in combination with changes in corner banking, bouncing off curbs, etc? Separating out corner banking is.. a bitch. Can use GPS.. but then you're limited to 20Hz or whatever it is.

Not to mention many of the data channels inevitably get very noisy. What's real? What isn't? If you're creating a camber trace from kinematic data it's invariably going to be off. Sometimes very significantly. If you're measuring it with laser sensors.. how much of your "camber gain" is coming from movement of your bracketry?

Track testing time is always at a premium. IMO a cognizant driver with good communication skills will always be essential to developing the best race package, and getting the most stuff done in the shortest amount of time.

[speedsense]

BTW, I am not saying that driver's recall isn't used anymore, it is. Though in 1988 (assumed pre-data era), it was far more important that the driver's duties and abilities, were included at top of the list, recall.
As data analysis has progressed, recall and chassis setup knowledge has digressed as far as how important it is to have such an ability. It is the advancement of data analysis, simulations and technology that has removed that importance and replaced it to the point that a driver without good recall or setup knowledge can compete with those that do have that skill, provided they have a good backing in data analysis.

[Jersey Tom]

BTW, I am not saying that driver's recall isn't used anymore, it is. Though in 1988 (assumed pre-data era), it was far more important that the driver's duties and abilities, were included at top of the list, recall.
As data analysis has progressed, recall and chassis setup knowledge has digressed as far as how important it is to have such an ability. It is the advancement of data analysis, simulations and technology that has removed that importance and replaced it to the point that a driver without good recall or setup knowledge can compete with those that do have that skill, provided they have a good backing in data analysis.

Sure. But it's one thing to just compete with the top dogs.. to be in the same field. It's another to win races, and yet another to win championships.. especially with tires, aero components, etc changing every week.

The sharp driver / engineer combination will make progress faster than just engineers by themselves. The whole idea is to get ahead of your competition and win races. I say it's still absolutely essential to have the driver in the process.

Plus there are some pro series where you are banned from having ANY telemetry on race day, in which case you damn well better have good driver/engineer communication. If you come in expecting to just go off data.. you're hosed.

[Jersey Tom]

Good, accurate, concise driver feedback makes the chassis tuning process MUCH easier and faster for the engineer, data or not. Consequently, inaccurate subjective feedback will screw you good.

"Car is oversteer" is crap feedback.

"Car is oversteer on entry" even is pretty much crap.

Good driver needs to identify what the car's doing, when, and why. For example:

"Car becomes free on entry from forward weight transfer under the brakes... mid corner balance is neutral with smooth off-to-on throttle transitions.. car tightens up slightly on-throttle until rear end breaks free. Car turns-in well on low speed corners in general, but is a little over responsive during small steering corrections at high speed"

Would be a start.

Oh oh oh I want to diagnose...

The front springs hardened to reduce weight transfer. Actually, since the rear breaks quite early you could soften the rears to achieve the same whilst improving rear traction. Reduce front wing to improve high-speed responsiveness. With the softened rear springs a possible softening of front ARB to counteract the possible increase in understeer.

amirite?

I just made that scenario and comments up, but if I were to diagnose it...

Changing the springs won't change the load transfer, just how fast it occurs. Also, didn't quite say that the rear breaks away early.. just that it gets tight first until you get on it enough to break them loose.

If you have room to play, taking air out of the tires will make them less sensitive to load transfer (cornering stiffness will change less with dFz). Alternatively, if you have a Salisbury diff you can change the ramp angles to lock more on the 'off-power' side and less on the 'on-power' side, unless the rear really does spin early in which case maybe some more locking on-throttle would be good.

But yea I'd probably try less front or more rear wing to numb the response a bit at high speed.. or can be done with stiffer front bump stops if the car is riding on them, or even front/rear damper split.

Or at least that's where I'd start.

[Scotracer]

Good, accurate, concise driver feedback makes the chassis tuning process MUCH easier and faster for the engineer, data or not. Consequently, inaccurate subjective feedback will screw you good.

"Car is oversteer" is crap feedback.

"Car is oversteer on entry" even is pretty much crap.

Good driver needs to identify what the car's doing, when, and why. For example:

"Car becomes free on entry from forward weight transfer under the brakes... mid corner balance is neutral with smooth off-to-on throttle transitions.. car tightens up slightly on-throttle until rear end breaks free. Car turns-in well on low speed corners in general, but is a little over responsive during small steering corrections at high speed"

Would be a start.

Oh oh oh I want to diagnose...

The front springs hardened to reduce weight transfer. Actually, since the rear breaks quite early you could soften the rears to achieve the same whilst improving rear traction. Reduce front wing to improve high-speed responsiveness. With the softened rear springs a possible softening of front ARB to counteract the possible increase in understeer.

amirite?

I just made that scenario and comments up, but if I were to diagnose it...

Changing the springs won't change the load transfer, just how fast it occurs. Also, didn't quite say that the rear breaks away early.. just that it gets tight first until you get on it enough to break them loose.

If you have room to play, taking air out of the tires will make them less sensitive to load transfer (cornering stiffness will change less with dFz). Alternatively, if you have a Salisbury diff you can change the ramp angles to lock more on the 'off-power' side and less on the 'on-power' side, unless the rear really does spin early in which case maybe some more locking on-throttle would be good.

But yea I'd probably try less front or more rear wing to numb the response a bit at high speed.. or can be done with stiffer front bump stops if the car is riding on them, or even front/rear damper split.

Or at least that's where I'd start.

I didn't think of the tyres - oh yeah.

I really need to get better with dynamics and kinematics. I specialised in Thermodynamics and Fluid Mechanics (I'll sort any CFD problems you've got!) at uni so I really need to sort that --- out. I've got a few books I bought since I graduated to get myself back up to speed on it; Milliken & Milliken for the fundamentals, Tyre and Vehicle Dynamics by Pacejka to get to grips more with tyres [I just noticed - nice pun!]. Anything else worth reading? I should know this stuff really but I really got pissed off with dynamics at uni so dropped it whenever I could. Silly me.