35 replies to this topic

[seryt]

I'm curious as to whether understeer/oversteer balance in modern F1 cars is totally dominated by downforce or whether the amount of load transfer at the front relative to the rear still plays a major role. Is there a particular ratio of load transfer that would be a target? I understand that in a non-aero car the amount of load transferred at the front relative to the rear should not depart too far from the weight distribution in order to get close to neutral steady state behaviour. The front would perhaps be a little stiffer than this in order to account for reduced lateral grip at the rear due to traction. With cars having downforce it seems greater flexibility in approach would be possible, for example would there be any advantage to having a very much greater front roll stiffness and using additional downforce to overcome the natural understeer?

[Goran Malmberg]

We are having a centre of gravity location on the car. We are also having a centre of aero force. Normally the centre of aero should be aft of the centre of mass location, in order to create stability for the car. This has been metioned as a comparsion with an arrow having feathers in the back.
Goran Malmberg

[Fat Boy]

Originally posted by seryt
for example would there be any advantage to having a very much greater front roll stiffness and using additional downforce to overcome the natural understeer?

You always end up juggling the mechanical and aerodynamic balance of the car. You can use one to compensate for the other a small amount, but generally speaking, you've got to satisfy both in order to make a fast car. The trick comes when the means to get a balance in one aspect hurts the balance in another. To make a stable aerodynamic platform, you may want stiff front springs. This may be counter to what it takes to get a good mechanical balance. Compromises in aerodynamic and mechanical balance always need to be made in order to make the fastest car around a certain track on a certain day.

If you ever get to the point where you find a perfect aero and mechanically balanced car, let me know how you did it!

[Joe Bosworth]

Seryt writes: "I'm curious as to whether understeer/oversteer balance in modern F1 cars is totally dominated by downforce or whether the amount of load transfer at the front relative to the rear still plays a major role."

I will try to put two bobs worth of info out to put the answer into context.

The mechanical grip of an F1 car now generates about 1.4-1.5 G. That is the the grip that it would generate entirely without any under body or wing induced force. This is the G force available when an F1 car is negotiating a very slow hairpin type turn. Once upon a time with stickier tyres it would have been a little higher.

The G forces above this level come from aero induced downforce and increase at the square of the increase of car velocity up to max speed at which about 4+ G can be generated.

So we have a relationship of mechanical to aero contribution for a typical set up that looks like, (and this will change with set up from track to track):

MPH Mech G Aero G Total G Aero %
50 1.4 .2 1.6 16
100 1.4 1.1 2.5 43
150 1.4 2.4 3.5 63
200 1.4 4.3 5.7 75

(Pardon any rounding errors.)

So you can see that below 100MPH mechanical grip is more important and above the aero becomes predominant.

However one would be really dumb to ignore either as both contribute. Likely most would want to set the car up close to balanced on mechanical grip without aero and bias the aero so that it induces just a bit of understeer load on the front.

Regards

[Ben]

I would represent this as normal load seen by the tyre. If we assume normal load can simply be represented as:

(m*g) + (0.5*Cl*A*rho*V^2)

And representative values are:

m = 600Kg
Cl = 1, 2 or 3
A = 1.5 m^2
rho = 1.15 Kg/m^3

The speed at which the downforce contributes approximately 50% of the normal load is around 306km/h for a Cl of 1, 216km/h for a Cl of 2, and 180 km/h for a Cl of 3.

It’s an interesting exercise to calculate this equation for speeds from 0 to 300km/h and see the way the percentage of load provided by downforce varies with speed. The next thing would be to compare front and rear CL values and look at how the balance might change with speed.

Ben

[seryt]

Thanks for the replies. It looks like the aero loads are likely to be dominant in an F1 car beyond any reasonable speed. I was trying to do some sums looking at the relative significance of aero/mech balance and it seems that, with tyres having friction factor to vertical load sensitivity in the range of that in the Avon published data, a fairly small change in downforce balance is equivalent to a much larger change in roll stiffness ratio. For example, increasing the front:rear roll stiffness ratio by about 10% could be compensated for by only 1% change in the front:rear downforce ratio when the total downforce is about 25% of car weight. With more than this amount of total downforce the car would begin to oversteer even though there is a higher roll stiffness at the front. I'm not sure I've done the sums right because it seems quite large changes in roll stiffness ratio produce relatively modest changes in balance. This would suggest that adjusting roll stiffness ratio would not be expected to have a major impact on steady state cornering balance compared to other variables. Is it possible this is correct? Maybe there is something else in tyre behaviour that plays a bigger role than sensitivity of friction coef to vertical load?

[McGuire]

Originally posted by seryt


I'm not sure I've done the sums right because it seems quite large changes in roll stiffness ratio produce relatively modest changes in balance. This would suggest that adjusting roll stiffness ratio would not be expected to have a major impact on steady state cornering balance compared to other variables. Is it possible this is correct?

I wouldn't think so. What does the ARB adjuster do? Changes the balance from over to understeer and back via a small change in roll distribution.

Anyway, balance is not a steady-state condition.

[Fat Boy]

I think there are a lot of people in F1 that like to forget the mechanical component of the handling balance. If you listen to some of the designers/chief engineers it's aero, aero, aero. Certainly Gascoyne is one of the worst. Look at the performance of this cars. On a super smooth, high speed track, they can be fast. Do they have the ability to run with McLaren for the entire race, contend for the win, or contend for a championship? No, they don't. They're always just a little lacking. Why is this? The devil is in the details. It is not just a game of who can make the best aerodynamically peforming car. It's about who makes the best compromise that involves both aero/mechanical components.

Most F1 corners are something like 70 mph at the apex speed. The aero contributes huge to the braking and entry, but in the roll-through portion of the corner is maybe only 15-20% of the equation. Mechanical balance matters.

[Lukin]

Originally posted by McGuire

I wouldn't think so. What does the ARB adjuster do? Changes the balance from over to understeer and back via a small change in roll distribution.

Wouldn't it vary from car to car, depending on the amount of lateral load transfer? If your car transfers 25% of its mass then a small change in roll stiffness distribution would have a pretty big effect I would think. Not that you'd ever want that!

[McGuire]

Originally posted by Lukin


Wouldn't it vary from car to car, depending on the amount of lateral load transfer? If your car transfers 25% of its mass then a small change in roll stiffness distribution would have a pretty big effect I would think. Not that you'd ever want that!

Sure. All I am saying is that the ARB adjuster will indeed change the balance of the car. Its total rate will typically be some fraction of the total roll resistance and wheel rates. Yet in a halfway sorted car, one or two clicks on the ARB adjuster will result in a change of balance that is apparent to the driver. But that's just it: balance is not a steady-state condition, or a point on a sheet of paper or a word in the data stream.

And this is how two qualified, professional drivers can work on identical cars for a full day on the same track and arrive at two remarkably different spring/damper/ARB packages. Yet the two setups may have essentially identical cornering power, and surprisingly similar "balance," at least in the center of the corner. Where the two cars will differ is mainly in their transient responses: that is, how they enter, exit and take a set at any given point. And each driver may declare the other's setup undriveable. Car A does not communicate to Driver B in his language what it is about to do. Engineers hate this, but traditionally chassis setup is in large degree a matter of getting the driver a car he can work with. They are currently trying to stamp this out in F1 and other series right now. The engineers now appear to have the upper hand, working at turning the drivers into interchangeable robots, optimizing the car to a computational model ideal and telling the driver to shut up and drive it the way it needs to be driven. Will be interesting to see how that all works out.

[Boro]

Originally posted by Fat Boy
I think there are a lot of people in F1 that like to forget the mechanical component of the handling balance. If you listen to some of the designers/chief engineers it's aero, aero, aero.

I was at a Milliken lecture two days ago and he said that it's amusing how some F1 designers to this day still have the gross oversight on the effects of unwarranted rear bump steer in their suspension design.

especially when said bump steer causes the wheels to increase in toe out

could not believe my ears when he said that lol; i guess those guys are human afterall

[zac510]

Originally posted by Boro


I was at a Milliken lecture two days ago and he said that it's amusing how some F1 designers to this day still have the gross oversight on the effects of unwarranted rear bump steer in their suspension design.

especially when said bump steer causes the wheels to increase in toe out

could not believe my ears when he said that lol; i guess those guys are human afterall

It is probably the subject of a whole new thread, but now that two design teams are being employed at big budget teams it seems more of a case of quantity over quality now?

[angst]

Originally posted by McGuire


Sure. All I am saying is that the ARB adjuster will indeed change the balance of the car. Its total rate will typically be some fraction of the total roll resistance and wheel rates. Yet in a halfway sorted car, one or two clicks on the ARB adjuster will result in a change of balance that is apparent to the driver. But that's just it: balance is not a steady-state condition, or a point on a sheet of paper or a word in the data stream.

And this is how two qualified, professional drivers can work on identical cars for a full day on the same track and arrive at two remarkably different spring/damper/ARB packages. Yet the two setups may have essentially identical cornering power, and surprisingly similar "balance," at least in the center of the corner. Where the two cars will differ is mainly in their transient responses: that is, how they enter, exit and take a set at any given point. And each driver may declare the other's setup undriveable. Car A does not communicate to Driver B in his language what it is about to do. Engineers hate this, but traditionally chassis setup is in large degree a matter of getting the driver a car he can work with. They are currently trying to stamp this out in F1 and other series right now. The engineers now appear to have the upper hand, working at turning the drivers into interchangeable robots, optimizing the car to a computational model ideal and telling the driver to shut up and drive it the way it needs to be driven. Will be interesting to see how that all works out.

An interesting thought, and perhaps the top two constructors in this year's championship showed the two almost diametrically opposed strategies on that front. Renault seem to have designed the car and, as you suggest, told the driver to just drive the damned thing - where McLaren produced driver specific suspension upgrades through the year.

[Fat Boy]

Originally posted by McGuire
And this is how two qualified, professional drivers can work on identical cars for a full day on the same track and arrive at two remarkably different spring/damper/ARB packages. Yet the two setups may have essentially identical cornering power, and surprisingly similar "balance," at least in the center of the corner. Where the two cars will differ is mainly in their transient responses: that is, how they enter, exit and take a set at any given point. And each driver may declare the other's setup undriveable. Car A does not communicate to Driver B in his language what it is about to do. Engineers hate this, but traditionally chassis setup is in large degree a matter of getting the driver a car he can work with. They are currently trying to stamp this out in F1 and other series right now. The engineers now appear to have the upper hand, working at turning the drivers into interchangeable robots, optimizing the car to a computational model ideal and telling the driver to shut up and drive it the way it needs to be driven. Will be interesting to see how that all works out.

This is not just something that is relegated to F1. It's a very common argument. I was talking with an engineer a couple months ago that was telling me to go a certain direction. I told him that my drivers didn't like that type of car. His comment was, "They're paid, right?". What he was getting at was that it wasn't their decision what type of car they should or should not like.

Try as I might, I just can't swallow this. I really feel this is the approach by guys who write/use a certain simulator, make the ultimate car in that simulator, and then want to produce the exact same car on the track. The driver's job is to understand that he is driving the 'ultimate' car and to drive it in a fashion that is fast.

I treat the driver like just another car component. He is the same as a spring or a shock. He is a component that is highly variable. His interaction with everything else on the car is just part of a hugh spiderweb of feedback loops. If I ignore what he is telling me, then I might as well ignore what the tire wear is telling me. It's not using one of the parts of the equation to the fullest. Keep in mind, the driver is where it all happens. A driver in a 100% car that is only comfortable driving at 95% of the car's capabilities is only going to run 95%. A driver that is in a 98% car that is comfortable driving at 98% of the car capabilities is good for 96%. If I've got a 1% speed advantage on my competitor by _slowing down_ my car and making it easier to drive, I'm going to take it.

I guarantee you that there were times over the last couple years that the engineers at Jordan were standing around wondering how they could make their car go faster, and they just couldn't figure it out. I've got $100 that says if they would have just stuck a 1/2" gurney on the rear wing they could have picked up a second. The sim would say it's slower, but the driver would have thought you were a genius and would have been your best friend. All you had to do was watch that car drive around the track. It was the most nervous looking, terrible oversteering thing that they could have come up with. Why the hell that Gary Anderson never makes it out to the track to actually watch his car escapes me.

[Ross Stonefeld]

Originally posted by McGuire


They are currently trying to stamp this out in F1 and other series right now. The engineers now appear to have the upper hand, working at turning the drivers into interchangeable robots, optimizing the car to a computational model ideal and telling the driver to shut up and drive it the way it needs to be driven. Will be interesting to see how that all works out.

Somehow I cant see Ferrari falling into this trap. The drivers, and even engineers/designers, often talked about how driveable the car was, and when that car was on, it wasnt slow anywhere. Very rarely did you have both Ferraris having a bad day in the better part of the last decade.

Makes you wonder just how badly they got lost this year. Likewise Williams seem the ideal of design-by-blindfold.

[wegmann]

Originally posted by Ross Stonefeld
Makes you wonder just how badly they [Ferrari] got lost this year. Likewise Williams seem the ideal of design-by-blindfold.

I'm pretty convinced now that the tire rules were the most significant portion of Ferrari's undoing. No amount of engineering or driving could overcome the lack of modern on-the-track tire testing.

This F1 philosophy of telling the drivers how to drive may actually be the way forward until the rules change. Look at how otherwise good drivers performed at new teams, where they probably had to be "taught" how to drive a different car - Fisi, Montoya, Villeneuve all struggled immensely. Or perhaps it was these drivers' age and how much they are set in their ways? Each was easily beaten by a younger teammate, probably better programmed to drive these modern F1 machines.

[Boro]

Originally posted by Ross Stonefeld



Somehow I cant see Ferrari falling into this trap. The drivers, and even engineers/designers, often talked about how driveable the car was, and when that car was on, it wasnt slow anywhere. Very rarely did you have both Ferraris having a bad day in the better part of the last decade.

Makes you wonder just how badly they got lost this year. Likewise Williams seem the ideal of design-by-blindfold.

Mark Ortiz wrote a short blurb about why he thought Williams was doing bad even last year.

[Ross Stonefeld]

Originally posted by wegmann


I'm pretty convinced now that the tire rules were the most significant portion of Ferrari's undoing. No amount of engineering or driving could overcome the lack of modern on-the-track tire testing.

This F1 philosophy of telling the drivers how to drive may actually be the way forward until the rules change. Look at how otherwise good drivers performed at new teams, where they probably had to be "taught" how to drive a different car - Fisi, Montoya, Villeneuve all struggled immensely. Or perhaps it was these drivers' age and how much they are set in their ways? Each was easily beaten by a younger teammate, probably better programmed to drive these modern F1 machines.

JV seems pretty clued up about his driving though, and when they have done what he's asked for, he's performed. Why it takes so long always baffles me. Isnt that why you go testing, to try things out? Wouldn't it be better to 'waste' a day on a setup you know isnt going to work that the driver wants, just so you dont have a season of him complaining?

[desmo]

What if he's consistantly faster with the "wrong" set-up? That might throw a development program largely developed through modelling and simulation off the rails. I'd imagine there's a fair number of people working on such things that would bemoan the loss of control they probably feel they need to do their jobs as they'd like.

Organizations as large as a modern F1 team surely must have to deal with the sorts of internal politics and rivalries that beset any group of people that large.

[angst]

Originally posted by desmo
What if he's consistantly faster with the "wrong" set-up? That might throw a development program largely developed through modelling and simulation off the rails. I'd imagine there's a fair number of people working on such things that would bemoan the loss of control they probably feel they need to do their jobs as they'd like.

Organizations as large as a modern F1 team surely must have to deal with the sorts of internal politics and rivalries that beset any group of people that large.

But why, as someone remarked earlier, would how the driver requires certain parameters to be met to suit his needs be any different to how a tyre(for example) requires certain parameters to be met to suit it's needs?

[seryt]

Originally posted by McGuire


I wouldn't think so. What does the ARB adjuster do? Changes the balance from over to understeer and back via a small change in roll distribution.

Anyway, balance is not a steady-state condition.

Thanks McGuire - I was looking at the difference in slip angle to get balance and how the lateral load transfer ratio affects this. For the cases I was looking at, the changes in slip angle were quite small, hence my comment about relative insensitivity to roll stiffness distribution. This was a misleading way of looking at it - better to look at the differences in the associated forces which are indeed pretty large even for a smallish difference in slip angle. Agreed that things are seldom actually anywhere near steady state but its still interesting to try understand the steady state as a start to understanding transients.

[Ross Stonefeld]

Originally posted by desmo
What if he's consistantly faster with the "wrong" set-up? That might throw a development program largely developed through modelling and simulation off the rails. I'd imagine there's a fair number of people working on such things that would bemoan the loss of control they probably feel they need to do their jobs as they'd like.

Organizations as large as a modern F1 team surely must have to deal with the sorts of internal politics and rivalries that beset any group of people that large.

I can see this being a bigger problem at a smaller, midfield team actually. If you're Sauber with a brand new fancy wind tunnel, and not a lot of money for on-track testing; you have to put a lot of faith in your sim work.

[Fat Boy]

Originally posted by desmo
What if he's consistantly faster with the "wrong" set-up? That might throw a development program largely developed through modelling and simulation off the rails. I'd imagine there's a fair number of people working on such things that would bemoan the loss of control they probably feel they need to do their jobs as they'd like.

If the track setup is faster than the theoretically 'correct' setup, then you have to ask yourself a couple questions. Is my stopwatch wrong or is my computer model wrong? A racecar is such a difficult thing to model, that I don't know how anyone can ever truly think that the completely understand it. You might understand portions of it quite well, but even at that, there are going to be blind spots. You can't possibly know how every component is going to act in all situations. I don't care if you're talking about a guy writing notes on the back of a napkin or a guy spending all his time doing an ADAMS Sim. The difference is, the guy that is writing note on the napkin has pretty much accepted that he might be incorrect and is generally willing to reverse himself if he slows the car down. The guy running the ADAMS Sim can just blame the driver.

Read what Jaques V. says in his Bishop Interview:

MB: Going back to Sauber, what are your impressions of them as a team?

JV: "I'm happy now. But it took a while - and, at first, I wasn't happy. At first I was told, 'Well, that's the way Felipe drives. Just shut up and get in the car.' That wasn't necessarily my race engineer, by the way, but the team as a whole were a little like that at first.

"But, in the end, I got some things changed. But I really had to force the issue. It was a difficult way of working - but then, when they realised that what I'd insisted on worked better, the guys all became easier to work with. They opened up a bit. And some of the stuff I insisted on is now on Felipe's car, too."

I think getting in a car or kart and actually trying to compete is really good for an engineer. I've competed directly against some of my drivers in the past. Same engine & same kart. This guy was good, too. I measured everything on my kart to make sure it was the same as his. I had all the stuff on mine that he had on his. No matter what I did, I was 2 seconds off, and that was when I was happy with everything. Most of the time it was more. Was I not 'trying'? No, I was trying like hell. I just didn't have the goods. When engineers say, "He just isn't trying", it's a complete cop-out.

This is not to say that drivers are always right. Sometimes they over-charge corners or drive in a manner that doesn't suit the type of car they're driving. Smart drivers will take a little criticism and adjust what they're doing to suit the car. In general, though, the car has to be changed to suit the driver. At the very least, it has to be changed to make the driver comfortable. If that isn't fast, then you can use that setup as a starting point. Just throwing something simulated at the car and telling the driver it's perfect is forgetting the fact that you, just like the driver, are falable.

[Greg Locock]

Maybe it is different in race teams, but of the 8 hours a week I spend talking to the ride and handling guys, I'd say ~50% is discussing /how/ we can get results from the sim that are directly meaningful and useful to them. Admittedly we don't have relatively straightforward objectives like "get around the circuit as fast as possible", so I don't really have the option of developing the car in that way.

[desmo]

I've posted the following excerpt here before. A bit here from Peter Wright's excellent "Formula 1 Technology" on the topic I thought was both interesting and informative. There's more in the book:

"The main use of the more sophisticated lap simulators today is to guide and evaluate the results of R&D programs. F1 cars are so optimized that changes to almost any aspect of them can easily detract from performance. Any proposed development must be quantified in terms that can be input into the car model, and its effect on lap times must be evaluated on a full range of circuits. It is particularly useful in assessing control system algorithms (e.g., brake balance, differentials, drive-by-wire, and any proposed new gizmo from R&D).

"If we take a Stage 3 model (pretty comprehensive but essentially steady state -desmo) and run it under the same conditions as a real car, the output should match the data system of the car, give or take a few bumps on the track. If it does not, it is possible to make changes in the car model setup until it does. The necessary changes will indicate possible areas where the car settings do not corraspond correctly to those input into the model, and will act as pointers to the settings on the car that are not performing as expected. What is perhaps more useful, is when the output of the lap simulator broadly matches the data system of the car, but the driver is critical of the handling of the car. Matching the point in the outputs from the two systems from which the car does not respond the way the driver believes it should, enables the output of the model to be examined closely to determine the car parameters that cannot practically be measured by the data system (e.g., tire slip angle, rear wheel torque distribution, and wheel cambers).
Another application is to turn the lap simulator model into a virtual tire dynomometer. We have already seen that the most important input parameters are tire characteristics and that they are also the most uncertain. However if all the parameters are well validated, it is possible to compare the output of the lap simulator and the car data systems (very good data for wheel loads, slip angles, and such factors are needed) and to use parameter identification techniques to adjust the tire characteristics in the model, until all the aspects of the outputs match. Thus the actual performance characteristics of the tires on the track can be established.

"A good Stage 3 simulation will make an accurate prediction of the performance potential of a car, but will not assist the race car engineer in understanding what a highly skilled driver is complaining about when he describes what he has to do to extract that performance from the car, nor why a less-skilled driver cannot extract it. For that, the stability and control characteristics of the car must be modeled, including the transient response to any input or disturbance- Stages 4 and 5. For these, each subsystem that influences the transient responses must be modeled and, of course, validated by rig or track data. The development of accurate simulations of the dynamics of the car and its components is controlling the pace of development of stages 4 and 5.

"Stage 4 simulation models the response of the car to driver inputs. The lap simulation may use filters to model the effects, but the characteristics of these filters must be established. This can be done either by measuring the response characteristics of of the car or by modeling them. The latter approach requires accurate dynamic characteristics for the aerodynamics, tires, dampers and any control systems that may affect vehicle response, and the structural stiffness of the chassis, suspension, and steering. Models and validation rigs are now common for most of these systems, and a library of characteristics should be available, except for the aerodynamics and the tires.

"Predicting the non steady-state flow characteristics of a body that incorporates so much vorticity and seperated flow is extremely difficult. Dynamic model tunnel testing is almost impossible at the frequencies of interest, because the dynamic characteristics of the force balance and surface pressure sensor system preclude accurate measurements. CFD is at its weakest when it comes to modeling vorticity and seperation, and these may vary as the car moves.

"The development of Stage 4 simulation is invaluable in terms of understanding the characteristics of the car, but may offer little improvement of the lap time estimation. Modeling the response of the car to control system inputs is somewhat meaningless, if a major part of the control system (i.e., the driver) is omitted. Because circuit racing is now conducted on tracks with only 10-15 corners, the drivers learn them within a few laps and soon drive them open-loop. Control inputs include lead to compensate for phase lags in the response, and the magnitude of the input will be adjusted according to the stability and damping of the system. Only when the conditions are changing, or if the driver is at the limit searching for that last 0.5% of performance, does he drive closed-loop. The effect of a skilled driver on the transient response of the car when calculating lap times is to virtually eliminate transient effects. Even worse, the driver may actually drive in a non-optimum manner to overcome a handling shortcoming in the car, achieving a lower lap time than the simulation.

"Stage 5 simulates the effect of disturbances on the car, at two levels, and requires surface topography (e.g., bumps, camber and curbs). The effect of these on the horizontal motion of the car is an extension of Stage 4 but uses circuit feature inputs instead of driver inputs. It is sensitive to the driver's response, as discussed previously, and the fact the driver may modify his line to avoid or exploit circuit features, particularly curbs.

"Of greater importance is the effect of undulations in the track on grip of the tires. While the horizontal response characteristics are generally less than 7 Hz, the vertical osicillations that affect the forces on the tire contact patch are in the range of 3-30 Hz and higher. Knowledge of the resonant modes in the tires, suspension, chassis structure, engine, and transmission is necessary. If doubt exists that a rubber tire is sensitive to these higher frequencies, remember that Honda discovered that altering the firing order on its 500cc GP motorcycle to reduce the frequency and increase the amplitude of the 200+ Hz torque impulses dramatically increased traction.

"A tire has more than 50 modes, some in the hundreds of hertz range. A suspension has two fundamental modes, one at the sprung mass natural frequency (3-7 Hz) and the other at the unsprung mass frequency (12-15 Hz). The structure also has a large number of modes, but first torsion, and first bending, at 30+ Hz, are the most relevant. The dampers are the primary means of damping these modes but may cease to function above 25 Hz. Modeling and measuring these characteristics is a mammoth task, but it is being undertaken by most F1 teams that have 7 post excitation rigs. The road input consists of white noise, and most of the modes will be excited. Exavtly how the frequency and magnatude of the variation in vertical tire loads affects the mean coefficient of friction and the heat input of the tire is not yet fully understood. Tuning the chassis to a particular tire for a particular circuit is one of the last black arts of motor racing. When it becomes a science, Stage 5 can be completed."

I've bolded the bit that seems to suggest how such differences can perhaps be cooperatively resolved through good communication between drivers and engineers. The list of "car parameters that cannot practically be measured by the data system (e.g., tire slip angle, rear wheel torque distribution, and wheel cambers)" suggests that a driver's more or less anecdotal observations might still be critical to validating any sim work i.e. "art" and science must find their awkward interface for the development process to be effective.

[Greg Locock]

Bear in mind that we do measure vehicle slip angles, from which it is relatively easy to determine the tire slip angles (at least, it seems easy in my head, without doing it). Whether it is easy or not they are generated by the magic of Matlab. Basically we have a toolbox that takes our standard list of measured channels and spits out all the funny measurements that can be derived. It is more or less automatic. That's not practical in a circuit car, since our standard instrumentation weighs about 60 kg.

[McGuire]

Originally posted by Fat Boy
I've got $100 that says if they would have just stuck a 1/2" gurney on the rear wing they could have picked up a second.

Interesting. While watching on TV I have had that exact same thought, and more than once too. Not the $100 of course, but the wicker.

[Boro]

Originally posted by Fat Boy



I think getting in a car or kart and actually trying to compete is really good for an engineer. I've competed directly against some of my drivers in the past. Same engine & same kart. This guy was good, too. I measured everything on my kart to make sure it was the same as his. I had all the stuff on mine that he had on his. No matter what I did, I was 2 seconds off, and that was when I was happy with everything. Most of the time it was more. Was I not 'trying'? No, I was trying like hell. I just didn't have the goods. When engineers say, "He just isn't trying", it's a complete cop-out.

That's why I thank god everyday for FSAE.

and also why I race karts in the off-season as well.

I'd be damned if I put all my trust in the computer That thing is supposed to work FOR me, not the other way round

[Ben]

I've just come back from a bike test and the point made by Fat Boy is even more accepted in the two wheel world. If the rider doesn't like the bike no one cares what is theoretically the fastest.

The most interesting part of the week for me has been looking at the data with no preconceptions of what it "should be like" and trying to understand what the rider's comments actually mean.

I've done a bit of ADAMS simulation in my FSAE days and a lot of non vehicle dynamics ADAMS simulation in my two years in Aerospace so I feel I can comment. I think tools like ADAMS can certainly look at a wide variety of options and assist in baseline setup, but to treat any results as biblical fact is a serious mistake.

One of the key reasons being that flat track tyre testing is generally done on hard compounds and the flat track surface is nothing like a race track surface.

Ben

[dosco]

Originally posted by Fat Boy
This is not just something that is relegated to F1. It's a very common argument. I was talking with an engineer a couple months ago that was telling me to go a certain direction. I told him that my drivers didn't like that type of car. His comment was, "They're paid, right?". What he was getting at was that it wasn't their decision what type of car they should or should not like.

Try as I might, I just can't swallow this. I really feel this is the approach by guys who write/use a certain simulator, make the ultimate car in that simulator, and then want to produce the exact same car on the track. The driver's job is to understand that he is driving the 'ultimate' car and to drive it in a fashion that is fast.

To go off further on this tangent....

I can appreciate both sides of this argument. What about the driver that can't drive anything but his "favorite car"? Now, there are very few "driving geniuses" (guys that are fast in anything they drive) that come around....but it seems to me that the majority of F1 drivers are *only* fast in their "favorite car." If the car is slightly off, the driver is not competitive.

To your original statement, though, what that engineer said is a pretty common mentality just about anywhere in industry. It's not uncommon to see engineers (or ) that think they "know everything" close their minds to constructive feedback. Part of the human condition, IMO. The challenge is to rise above it.....

[-RM-]

But F1 is somewhat of a special case since your driver might have hammered override the better part of last race and the main question might be how many practice laps you dare to run.
How many practice laps is one tenth worth in F1?

[Ross Stonefeld]

testing... testing...

[Ross Stonefeld]

A snippet from this week's JPM interview...

...But we went through three different front geometries this year, just to adapt to my driving style, and at the same time we were changing the car, and it was going quicker and quicker.

[gbaker]

A good model will provide direction and some sense of magnitude, until it becomes tuned with empirical data. To suggest that it can nail every variable in a complex system is naive.

[Christiaan]

I wonder, you know how jets like the F16 are built aerodynamically unstable and then use computers to make them execute normally impossible moves? What if you seperated the two centres of downforce and through clever TC made the car swing through corners like the Mclaren with the second brake used to do

[dosco]

Originally posted by Christiaan
I wonder, you know how jets like the F16 are built aerodynamically unstable and then use computers to make them execute normally impossible moves? What if you seperated the two centres of downforce and through clever TC made the car swing through corners like the Mclaren with the second brake used to do

Well, the "unstable airplane" idea is kind of a misnomer. There are 2 kinds of stability when refering to aircraft - static and dynamic. The F-16 has "relaxed static stability" which means it's CP is close to the CG - very close to "neutral." The dynamic stability has more to do with the sizes of the wings, fuselage, empennage, etc etc and how the forces of each vary with angle of attack, etc - it's substantially more complex than the static stability issue.

Modern fighters are dynamically unstable - meaning that upsets will lead to undamped responses from the airplane (which behaves much like a spring-mass system). The computers are used to control the dynamic instability.

Lastly, I'm pretty sure such a setup on an F1 car would amount to a "computer controlled" or "active" system, and also would probably be a "variable geometry" system as well, and therefore would most likely be illegal.