Why G-Sum?
#16
Originally Posted by ColorChange
TD:
The goal of the driver (non race conditions) is to complete a lap as fast as possible. If both cars drive the ideal line, the one who goes faster will have the lower lap time. This, by definition, means he had a higher g-sum around the track. In real terms, what does this roughly mean? It means brake as late and as hard as possible, turn with maximum lat g's, and accelerate as hard and as early as possible. G-sum plotted through the turns shows this perfectly, specifically, how well the driver trades braking g's, for lat g's, for accel g's.
The goal of the driver (non race conditions) is to complete a lap as fast as possible. If both cars drive the ideal line, the one who goes faster will have the lower lap time. This, by definition, means he had a higher g-sum around the track. In real terms, what does this roughly mean? It means brake as late and as hard as possible, turn with maximum lat g's, and accelerate as hard and as early as possible. G-sum plotted through the turns shows this perfectly, specifically, how well the driver trades braking g's, for lat g's, for accel g's.
#17
Originally Posted by TD in DC
JCP911S,
I bet you posted this before you read my last post . . .
I am trying hard to steer this thread away from the traditional debate about the usefulness of data. If you read my post, you will see that I personally don't disagree with what you are saying. I am just curious about the question I posted.
TD
I bet you posted this before you read my last post . . .
I am trying hard to steer this thread away from the traditional debate about the usefulness of data. If you read my post, you will see that I personally don't disagree with what you are saying. I am just curious about the question I posted.
TD
As far as "friction circle" it is a useful model to help you visualize what the car is doing in a turn... but in the real world all it tells you is to manage transitions as smoothly as possible...
To use a football analogy, the "friction circle" is like telling a reciever "Accellerate as quickly as possible, make your cut cleanly, jump and the exact right moment and look the ball into your hands and catch it..." great advice... all you have to do it do it. and you are Jerry Rice.
#18
Originally Posted by JCP911S
Guess in a nutshell, G-Sum doesn't mean sh*t IMHO... maybe I'm old fashioned, but you have to systematically develope your driving technique and learn how to drive the car closer and closer to the limit... this means seat time and becomming both familiar and comfortable with how the car behaves... and how to control it in all situations.
As far as "friction circle" it is a useful model to help you visualize what the car is doing in a turn... but in the real world all it tells you is to manage transitions as smoothly as possible...
To use a football analogy, the "friction circle" is like telling a reciever "Accellerate as quickly as possible, make your cut cleanly, jump and the exact right moment and look the ball into your hands and catch it..." great advice... all you have to do it do it. and you are Jerry Rice.
As far as "friction circle" it is a useful model to help you visualize what the car is doing in a turn... but in the real world all it tells you is to manage transitions as smoothly as possible...
To use a football analogy, the "friction circle" is like telling a reciever "Accellerate as quickly as possible, make your cut cleanly, jump and the exact right moment and look the ball into your hands and catch it..." great advice... all you have to do it do it. and you are Jerry Rice.
I hear you, I hear you, believe it or not I do focus on seat time and driving . . .
#19
TD, very good! Character AND a brain? In an attorney now? No way!
Guess what happens if you correctly manage your slip ratio under braking (usually 20-30 for optimal) and slip angle (usually 5-7%) during turning? Guess what measure you can use to show you’ve done it? You guessed it, you have maxed long g and lat g respectively. Now when you want to see how well the driver transitioned these forces, you look at g-sum. It’s that simple.
Admittedly you could argue that the track changes minutely, blah, blah, blah, but these minor effects are dwarfed by driver variability/error, for everyone I’ve seen including pro’s (but not F1/CART/IRL level guys because I haven’t seen the data). Boy would I love to see their data but I am highly confident that g-sum would tell tons, if only because it’s so hard to take a F1 car from -4.5 long g’s, to 3.0 lat g’s, to 1.5 long g’s in a typical turn.
Guess what happens if you correctly manage your slip ratio under braking (usually 20-30 for optimal) and slip angle (usually 5-7%) during turning? Guess what measure you can use to show you’ve done it? You guessed it, you have maxed long g and lat g respectively. Now when you want to see how well the driver transitioned these forces, you look at g-sum. It’s that simple.
Admittedly you could argue that the track changes minutely, blah, blah, blah, but these minor effects are dwarfed by driver variability/error, for everyone I’ve seen including pro’s (but not F1/CART/IRL level guys because I haven’t seen the data). Boy would I love to see their data but I am highly confident that g-sum would tell tons, if only because it’s so hard to take a F1 car from -4.5 long g’s, to 3.0 lat g’s, to 1.5 long g’s in a typical turn.
#20
Originally Posted by ColorChange
G-Sum analysis is one of the best things you can do to evaluate driver performance. I do not state this as an opinion; I state it as a fact
#21
Originally Posted by JCP911S
As far as "friction circle" it is a useful model to help you visualize what the car is doing in a turn... but in the real world all it tells you is to manage transitions as smoothly as possible...
#22
Originally Posted by Geo
You may be trying to state it as fact, but numerous times people who are professionals in the business and others who use DA regularly have chimed in and disagreed. It is not fact.
My guess is you have to be pretty far up the driving foodchain to take DAS information and turn it into actual speed on the track.
So somebody tells me, a weekend tire squeeler, that I could brake 10 feet deeper and carry 3 more MPH into Deadman's Turn, I'd say "gee, thanks, right...all day long...uhuhh... phfffft!"
#23
Originally Posted by ColorChange
Admittedly you could argue that the track changes minutely, blah, blah, blah, but these minor effects are dwarfed by driver variability/error, for everyone I’ve seen including pro’s (but not F1/CART/IRL level guys because I haven’t seen the data).
The secret was something to the effect that during a track walk the teammate had noticed there was imperfection in the paving of that particular corner and if one moved off line the outside wheels would hook that impefection, the car would stay planted and be carried around the corner. In effect a bit of curbing to rely on. This enabled him to gain more grip than others who were taking the ideal line through the corner and it turned out to give him the edge to take the pole and go on to win the race...or so the story goes. Sadly, as Dennis related, the track was repaved before the next season, so he was never able to actually take advantage of that knowledge.
I wouldn't underestimate the abilities of professionals and skilled amatuers to seek out and find just those variables that you are suggesting we might as well ignore. Since hearing this lecture, in my own experience I've found a few of these place on several of the tracks I've visited where there is a bit more camber or a bump or groove that can be used to ones advantage. I frankly think that in the end, it is just this sort of thing that you are interested in, for its the way you really can achieve an advantage over those your competing against.
#24
Originally Posted by TD in DC
JCP911S,
I bet you posted this before you read my last post . . .
I am trying hard to steer this thread away from the traditional debate about the usefulness of data. If you read my post, you will see that I personally don't disagree with what you are saying. I am just curious about the question I posted.
TD
I bet you posted this before you read my last post . . .
I am trying hard to steer this thread away from the traditional debate about the usefulness of data. If you read my post, you will see that I personally don't disagree with what you are saying. I am just curious about the question I posted.
TD
Last: Segment times are very helpful and any decent DA system allows infinite dissecting of the lap,corners, straightaways. This allows you to dissect specific problems areas in driving, tire pressure/temps and suspension tuning.
Works for me at least.
#26
G sum gives a relative value from lap to lap.
If you are serious about analyzing your driving from data - use RPM, wheel speed, steering angle, brake pressure and throttle angle. Those things tell you what you are doing for various segments. Most people with Motecs are using G's for little more than track mapping.
In addition to the previously stated problems with G's, they also don't tell you things like, 'how smoothly did you release the brake or gas?'
Oh, and I have studied physics - actually dynamics and classical mechanics (in response to a previous question about someone's aptitude)
If you are serious about analyzing your driving from data - use RPM, wheel speed, steering angle, brake pressure and throttle angle. Those things tell you what you are doing for various segments. Most people with Motecs are using G's for little more than track mapping.
In addition to the previously stated problems with G's, they also don't tell you things like, 'how smoothly did you release the brake or gas?'
Oh, and I have studied physics - actually dynamics and classical mechanics (in response to a previous question about someone's aptitude)
#29
Originally Posted by PatK-Mpls
Mark-
What are you using? (I assume rpm, b press, steer angle, throttle pos and wheel spd)
What are you using? (I assume rpm, b press, steer angle, throttle pos and wheel spd)
RPM, Speed, engine temps/pressures. Front and rear brake pressures, throttle, steering, 4 shock pots and 2 axis G.
The interesting math channels include calculated brake bias (this is actual clamping force using cylinder sizes), ride height and downforce (fron change in ride height) and center of pressure (from downforce).
I am also starting to play around with front vs. rear roll. I am wondering if I can see chassis flex/oscillation.
A lot is just interesting stuff such as seeing the difference in downforce when there is a tailwind. Actaully seeing how light the car gets over humps. The shock stuff is fascinating and I can see some patterns when the car is not handling the way I want, but I have yet to be able to make changes based on shock data - seat of the pants works far better for me. I am not the greatest suspension tuner, but I can tell the difference between fairly subtle changes and can recall entry/mid/exit characteristics. Have not made it to the level where I can really tune one corner of the car that is not right.
I have used the services of a pro (as in he was paid for many years to do that job) data engineer to help me understand how to use this stuff. I still need to hire him for a couple of days to help tune the car and coach the driver.
The most valuable thing I have ever done with data is overlay my traces with a faster driver in my car. I have never gotten any value from g data, corner radius, inverse corner radius or the deriviative of g's for smoothness.
Feel free to fire away with any other questions.
#30
Very interesting. Particularly the downforce math channel calc. Are you able to make informed judgements about adjusting splitter/wing vs suspension based on the downforce feedback?
Have you ever found a good way to really measure slip angle (or even relative slip for that matter)? I was going to buy a yaw sensor and right before doing so was told they are a waste of money.
And, since you offered for me to fire away...
How difficult is it to tune you bump/rebound by analyzing the shock pot data. There is a good chance that I'll be running 2 way Ohlins next year and I am a bit nervous about setup. This last year I got myself messed up more than once putzing around with my shocks at races right up until qualifying. When I spoke with an engineer from Sachs he told me that if I did not have shock pots and a qualified engineer - I was merely groping in the dark. (Incidentally - last year car was 05 gt3 cup and next year car will be 997 cup)
Have you ever found a good way to really measure slip angle (or even relative slip for that matter)? I was going to buy a yaw sensor and right before doing so was told they are a waste of money.
And, since you offered for me to fire away...
How difficult is it to tune you bump/rebound by analyzing the shock pot data. There is a good chance that I'll be running 2 way Ohlins next year and I am a bit nervous about setup. This last year I got myself messed up more than once putzing around with my shocks at races right up until qualifying. When I spoke with an engineer from Sachs he told me that if I did not have shock pots and a qualified engineer - I was merely groping in the dark. (Incidentally - last year car was 05 gt3 cup and next year car will be 997 cup)