Jump to content


Photo

Saving slides by locking front brakes


  • Please log in to reply
25 replies to this topic

#1 100cc

100cc
  • Member

  • 3,178 posts
  • Joined: December 00

Posted 29 May 2012 - 14:47

There's a lengthy discussion on the iRacing forums about what happens in this video in the F1 car (watch from about 8s to 35):



The video is slowed down at that point - the actual speed can be seen in the video 100 to 200kph+.

What happens is when the car (exiting a corner for example), starts to lose its rear end at a yaw angle of anywhere from 10° to ~70°, you can save it instantly by stabbing the brakes. Stabbing the brakes locks up the fronts, and brings the rears to between 0%-8% forward slip (from ~30% or much higher when there is excessive wheelspin).

This is the best way to correct a slide in the sim and is very easy to do regardless of brake bias or what speed the slide occurs at. The unfortunate consequence is that if someone is behind you they will smash into the back of you.

The phenomenon has been explained with the front tyres losing their grip when locked so they snap back around - perhaps slightly excessively compared to what would really happen.

There is a very knowledgeable member that has patiently tried to explain to me why the locked fronts have less braking force than the rear tyres at that point but his explanations have left me more confused. He said that the sliding rears will always have more force than the locked up fronts.

Data that I've found shows that locked tyres produce a braking force in the region of 75% of the absolute maximum threshold braking of a tyre (at ~8% slip) or so depending on the tyres (at 0° slip angle) makes me feel that the fronts should have more braking force than the rears which are only being slowed down by the slip angle.

The telemetry of the incident shows that the car dives slightly when the brakes are applied - so I assume that means weight transfers forward.

If someone could explain to me what they feel is going on there it would be greatly appreciated.

Edited by 100cc, 30 May 2012 - 08:11.


Advertisement

#2 Lukin

Lukin
  • Member

  • 1,983 posts
  • Joined: January 03

Posted 29 May 2012 - 22:42

My thoughts..

- I thought you meant a real F1 car!
- I wouldnt call a tyre at 8% SR locked. With some tyres its about ideal. A fully locked tyre has a slip ratio of 100%. And no grip. And a big mark. And possibly a modified air venting system.
- Is the front locking actually correcting the slide or just a by product of what's going on? If the rears have lost grip and your trying to slow the car at the same rate, then the front won't be able to deal with it and will lock too. If the braking G is reduced it would reduce the front being loaded and again reduce the front braking ability compared to 'normal' braking.
- Is the car accelerating (driving) or is the rate of braking just reducing? Reduction of braking makes sense, you get very little grip with partially or fully locked tyres.
- If the front locking does correct the slide, I imagine it would be due to the Fz from the tyre being reduced which reduces the net yaw moment on the car. I heard or read somewhere that NAscar teams looked at different brake pads across the front axle to generate different Fx and help the rotation of the car while on the brakes.

#3 Greg Locock

Greg Locock
  • Member

  • 6,353 posts
  • Joined: March 03

Posted 29 May 2012 - 22:46

Depends entirely on the tire model and braking model and vehicle dynamics model. Basically it is wrong. I suggets you investigate it with car sim and find out exactly how crazy your model has to be before it behaves like that. I suppose 100% rear brake bias might do it.

#4 munks

munks
  • Member

  • 428 posts
  • Joined: January 03

Posted 30 May 2012 - 01:46

Depends entirely on the tire model and braking model and vehicle dynamics model. Basically it is wrong. I suggets you investigate it with car sim and find out exactly how crazy your model has to be before it behaves like that. I suppose 100% rear brake bias might do it.


I think you mean 100% front, right?

In any case, I think perhaps the argument being made is that engine power is being applied to the rear tires, and so the brakes and engine are fighting each other at the rear, which effectively makes for more front brake bias. Plausible?







#5 Greg Locock

Greg Locock
  • Member

  • 6,353 posts
  • Joined: March 03

Posted 30 May 2012 - 03:31

I was thinking that if you can get the rears to break traction they would slide sideways and kill the overyaw condition, assuming terminal plow. I don't care, it basically sounds all wrong.

#6 100cc

100cc
  • Member

  • 3,178 posts
  • Joined: December 00

Posted 30 May 2012 - 07:40

Thanks for the replies guys. I was wondering if anyone would give a crap about racing sim physics :)

Depends entirely on the tire model and braking model and vehicle dynamics model. Basically it is wrong. I suggets you investigate it with car sim and find out exactly how crazy your model has to be before it behaves like that. I suppose 100% front brake bias might do it.


In any case, I think perhaps the argument being made is that engine power is being applied to the rear tires, and so the brakes and engine are fighting each other at the rear, which effectively makes for more front brake bias. Plausible?



This is the conclusion I've come to. What seems to be wrong (or at least one thing that is wrong) is that the front tyres always lock very easily while the rears don't without the driver having to be an expert in adjusting brake bias super fast and perfectly ;) - as long as you have any thorttle on, the rears don't lock - they just have the power removed to low levels.

It's impossible to lock the rears even with maximum rear brake bias if you have just a tiny bit of throttle, but the fronts still lock perfectly.

The below picture shows what is happening:
Posted Image

Edited by 100cc, 30 May 2012 - 08:29.


#7 100cc

100cc
  • Member

  • 3,178 posts
  • Joined: December 00

Posted 30 May 2012 - 07:57

- I wouldnt call a tyre at 8% SR locked. With some tyres its about ideal. A fully locked tyre has a slip ratio of 100%. And no grip.

Neither would I ;).. you misread slightly:

Stabbing the brakes locks up the fronts, and brings the rears to between 0%-8% forward slip.


I don't agree that a fully locked tyre has no grip.. it has more braking traction than a tyre braking at (for example) 2% longitudinal slip, and depending on the tyre and surface could have about 70% the braking traction of a tyre braking at the perfect longitudinal slip %. :)

Edited by 100cc, 30 May 2012 - 07:58.


#8 Kelpiecross

Kelpiecross
  • Member

  • 1,730 posts
  • Joined: October 10

Posted 30 May 2012 - 12:30

Neither would I ;).. you misread slightly:


I don't agree that a fully locked tyre has no grip.. it has more braking traction than a tyre braking at (for example) 2% longitudinal slip, and depending on the tyre and surface could have about 70% the braking traction of a tyre braking at the perfect longitudinal slip %. :)



Being somewhat old-fashioned and not very "computery" and having never played a racing game or used a racing simulator - is it possible that the simulator program just can't cope with unusual actions like jamming on the brakes to correct oversteer? Maybe it just confuses the program.

If it is an accurate reflection of what would happen in real-life maybe it is related to the seemingly routine recovery from huge oversteer angles by racing front wheel drive sedans by applying full power and full opposite lock - the combination of the two possibly reducing the front grip to almost zero.

#9 100cc

100cc
  • Member

  • 3,178 posts
  • Joined: December 00

Posted 30 May 2012 - 14:19

If it is an accurate reflection of what would happen in real-life maybe it is related to the seemingly routine recovery from huge oversteer angles by racing front wheel drive sedans by applying full power and full opposite lock - the combination of the two possibly reducing the front grip to almost zero.

Yes that it pretty much what happens. That's a nice way to explain it.

Since I posted the thread it has been pretty explained why it happens. There's the picture I posted which explains why the car stops rotating when the fronts are locked - which is accurate to real life.

This is very easy to do in the sim because the fronts lock instantly while the rears keep rotating.

One of the reason the front snaps around in the sim but not in real life is because they have far less lateral grip on tyres at high slip angles (such as locked up) than they should - this was admitted by a developer. So effectively like you said full steering lock and full power on a front wheel drive. Lack of front wheel lateral grip makes the front swing back around.

The confusion I had with the locked tyres grip debate was a simply a case miscommunication - we were talking about two different situations.

#10 Kalmake

Kalmake
  • Member

  • 4,492 posts
  • Joined: November 07

Posted 30 May 2012 - 15:52

And it's not just the tyres. The loss of downforce in yaw tends to be way too little in sims. In rFactor where you could see these parameters they were wrong in almost every mod.

An F1 car going fast and very sideways does have some tendency to straighten up (like an arrow). But again it depends on the sim if this is modeled at all.

#11 Wolf

Wolf
  • Member

  • 7,883 posts
  • Joined: June 00

Posted 31 May 2012 - 21:48

I may be looking at this the wrong way and from my (perhaps wrong) experience, but on rare occasions I had my front wheels locked it made my rear end 'skittish' (which would to my untrained™ eye suggest that fronts retained enough grip to resist the influence of road irregularities, whereas it was the rear that lost it to the point it was suspectable to bump-steer). I'm a bit confused as it (a) might signify understeering condition, which would support original theory, but (b) original theory doesn't feel right, and pictire 100cc posted seems a bit odd (shouldn't lateral forces be pointed the other way around, for example?)...

What I think might also bear relevance is the diff behaviour (I would venture a guess that the yawing moment of torque split is not to be ignored when considering the yawing moment of the whole vehicle). Somebody correct me if I'm wrong but I think it was said that it is (or was in the early days when Salisbury adjustable diffs were introduced) to have much more locking on power than on coast (in case of GPL era, I think it was 30/80 degree ramp angles). I'm unsure how diff would act with throttle and brake simultaneously applied, but my uneducated™ guess is that coat would override the power which would, if my previous assumptions were right contribute to tendency to oversteer.

I know it might be of little consequence for the discussion, but wisdom of the ancient ones (Sir Stirling Moss in particular) and my own experience do suggest the excess speed may be scrubbed and oversteering tendency may be countered by artificially introducing understeer (as well as being accepted method of emergency slow down in case of missed braking point or brake failure*), but by trailing off the throttle and giving a rapid and excessive turn-in on the steering wheel.

* as well as correct way of setting four wheel drift for S-bends, IIRC Jenk's account of epic Mille Miglia

#12 Slowinfastout

Slowinfastout
  • Member

  • 9,681 posts
  • Joined: October 07

Posted 01 June 2012 - 19:02

I picked up this bad habit in sims as well, but it's only through trial and error (spinning a lot in the virtual world is basically a viable tool to learn the limits :D )... I don't think that behavior translates to the real-world, at least not to that extent. My theory is that the physics in racing simulations tend to get iffy when going over the limit, like when spinning or even just throwing in inputs that don't make sense for the situation you're in. For instance, steering heavily full-lock back and forth at full speed on a long straight also feels surprisingly safe, speed and inertia seems to kill those crazy inputs more than they would in real life.

I believe the reason we're able to save slides in the fashion described is that by locking the wheels the simulation basically reverts back to something that is more predictable, both for you as a driver and for the code going nuts in the background.

Edited by Slowinfastout, 01 June 2012 - 19:04.


#13 munks

munks
  • Member

  • 428 posts
  • Joined: January 03

Posted 02 June 2012 - 01:45

My theory is that the physics in racing simulations tend to get iffy when going over the limit, like when spinning or even just throwing in inputs that don't make sense for the situation you're in.


Even if the physics are partly validated, there's rarely any good data for tires beyond the limit. Even if you wanted and measured it, one would have to understand how the temperature and wear was affecting your measurement. Because you'd get a different result if you measured constantly or stopped and measured again.

I'm actually still not convinced that the described behavior is definitely right or wrong. While F1 drivers have been known to use throttle and brakes at the same time, nobody has ever done it to the extent shown here.

#14 Tenmantaylor

Tenmantaylor
  • Member

  • 18,041 posts
  • Joined: July 01

Posted 02 June 2012 - 03:37

4 reasons locked rears have more grip than locked fronts (when all tyres locked)

-rear tyres have bigger contact patch
-rear axle has more weight on it (f1 cars about 55% rear weight dist.?)
-rear has more downforce (usually kept similar to weight dist.)
-fronts usually have higher brake bias to take advantage of weight transfer that occurs under heavy (non-locked) braking.

Another major factor is that combined these reasons mean that fronts will usually lock easier than rears meaning the fronts will slide first (unless brake bias is too far to the rear).

This is my layman understanding of the situation.

Edited by Tenmantaylor, 02 June 2012 - 03:42.


#15 Ross Stonefeld

Ross Stonefeld
  • Member

  • 70,106 posts
  • Joined: August 99

Posted 02 June 2012 - 14:37

Isn't this purely a balance issue? In extreme oversteer/spins the car is rotating more and more around the front of the car. The most extreme example of this is someone doing a post-victory 'doughnut' with a little bit of brake so the car rotates with the front wheels more or less at the anchor point.

By 'removing grip' at the front wheels the center of rotation goes back more towards the middle of the car.

You see this sometimes with onboards in the wet, Rubens Barrichello in particular. When the rear is nervous they crank in a ton of steering angle on corner entry so the front wheels slide a little too, so the car is effectively in a four wheel drift. Which isn't ideal but it's better than spinning. We just normally associate four wheel drifts(I'm using the visual term rather than the mechanical one) with close to neutral steering angle rather than arms folded over.

#16 Bloggsworth

Bloggsworth
  • Member

  • 9,394 posts
  • Joined: April 07

Posted 03 June 2012 - 14:14

Please explain how you can lock the front wheels without locking rears which already have reduced grip (they are sliding after all), also explain why a driver could more quickly and precisely lock the brakes than apply a touch of opposote lock. Every F1 car I have seen lose grip at the back continues to rotate if the brakes are jammed on - The cars are weight-biased towards the rear, when the brakes are locked physics takes over.

Demonstrate that effect on Mclaren's or Red Bull's simulator and there might be a point to this thread.

#17 Tenmantaylor

Tenmantaylor
  • Member

  • 18,041 posts
  • Joined: July 01

Posted 03 June 2012 - 15:42

Could be quite easily proven either way depending on the brake bias. Changing 1-2% fore or aft too much is usually enough to change from fronts locking first to rears.

Ps bloggs, not sure about mclaren but I believe red bulls simulator is effectively rfactor pro.

Also, this technique will invariably cost you alot of time and isn't something you'd want to do often. Its more of a last resort to prevent a spin. In real life it would heavily flat spot the tyres and make them almost undrivable.

Edited by Tenmantaylor, 03 June 2012 - 15:48.


#18 Ross Stonefeld

Ross Stonefeld
  • Member

  • 70,106 posts
  • Joined: August 99

Posted 03 June 2012 - 15:44

It is, but with their own model.

#19 munks

munks
  • Member

  • 428 posts
  • Joined: January 03

Posted 04 June 2012 - 02:47

Please explain how you can lock the front wheels without locking rears which already have reduced grip (they are sliding after all)


You apparently haven't read anything in this thread so I'll help you out. It's already been suggested that power is being applied to the rear (counteracting some portion of the rear braking effort). And they're sliding due to power oversteer, in other words the rear tires are spinning too fast, something that a bit of brakes might cure.

As for the F1 simulators like McLaren's or Red Bull's, or CarSim which Greg suggested above, I'm not sure why anyone thinks they are the arbiters of truth. They are all *models*, just like iRacing, just like rFactor, just like anything else. More validated? Possibly and probably, but typically only validated for normal operating conditions, not sliding 45 degrees sideways and hitting nearly full brakes and full throttle at the same time. And when tires get into the range where they are essentially unmeasurable, all bets are off. The only way to actually test the described behavior is with the real car, in my opinion. And good luck with that.

Advertisement

#20 Ross Stonefeld

Ross Stonefeld
  • Member

  • 70,106 posts
  • Joined: August 99

Posted 07 June 2012 - 14:22

Looks like it's a problem with the game

http://www.virtualr....ses-controversy

#21 munks

munks
  • Member

  • 428 posts
  • Joined: January 03

Posted 07 June 2012 - 15:16

Looks like it's a problem with the game

http://www.virtualr....ses-controversy


Where does the link give evidence of your description?

#22 Ross Stonefeld

Ross Stonefeld
  • Member

  • 70,106 posts
  • Joined: August 99

Posted 07 June 2012 - 15:23

Would you be satisfied with 'controversy caused by an exploit in the online racing simulation’s physics simulation.' or that they call it the “2 Foot Magic Save Hax" ?






#23 100cc

100cc
  • Member

  • 3,178 posts
  • Joined: December 00

Posted 07 June 2012 - 23:09

As I touched on on my previous post the thread on iRacing has explained why it works in the sim but not in real life. A few sim developers were part of the discussion.

Bullet points are:
  • Tyres on iRacing (and other sims) lose a lot more lateral grip than they should in reality when at high slip angles such as locked up. This is the main reason the car snaps back around. In reality the car should remain pointed where it is with the fronts locked up - it shouldn't instantly reverse the spin.
  • There is some strange torque transfer going on when throttle is applied even if all 4 wheels are locked. (a car spinning with all 4 wheels locked magically starts spinning the other way as soon as throttle is applied)
  • The brakes lock the fronts very easily but not the rears if any throttle is applied (tested with max/min brake bias and the save is just as easy). The rears are brought to an ideal slip angle (a few % from ~30% spinning) and don't lock up.

iRacing is a great racing sim and once they get these issues solved - which seems to be primarily the excessive loss of lateral grip at high slip, it will be a lot better.


explain why a driver could more quickly and precisely lock the brakes than apply a touch of opposite lock.

This is the main reason the "hax" is so annoying. It works a lot better than countersteering. In fact countersteering seems the fastest way to spin on iRacing :|. Drifting is almost impossible in most cars on the sim.


Also, this technique will invariably cost you alot of time and isn't something you'd want to do often. Its more of a last resort to prevent a spin. In real life it would heavily flat spot the tyres and make them almost undrivable.

You'd be surprised how often the fast drivers use the technique and how little time it costs them. A 1 tenth lockup wouldn't do any more damage to tyres than a spin would.

Edited by 100cc, 07 June 2012 - 23:10.


#24 munks

munks
  • Member

  • 428 posts
  • Joined: January 03

Posted 08 June 2012 - 03:50

Would you be satisfied with 'controversy caused by an exploit in the online racing simulation’s physics simulation.' or that they call it the “2 Foot Magic Save Hax" ?


No.

Edited by munks, 08 June 2012 - 03:52.


#25 Wirra

Wirra
  • Member

  • 1,326 posts
  • Joined: December 08

Posted 08 June 2012 - 05:13

I remember reading in 'Grand Prix international' at the time about Mario Andretti using this technique during the running of the Las Vegas GP around 1981-82 and how it prevented an incident with would have involved race winner Alan Jones. It attributed Andretti’s technique to his oval racing experience.

#26 Ulledulledof

Ulledulledof
  • New Member

  • 7 posts
  • Joined: July 10

Posted 08 June 2012 - 08:13

I think I saw Coulthard doing this in magny cours 2008 or so on first lap final chicane. I can't find a video evidence though.

About why do people do it in simracing. I guess sims are generally a bit on slow side. Cars don't react fast as real ones do. Players have more time to react and do all sorts of goofy things.