brucegre

Whoa - that'll teach me not to go away for holidays and skip reading the forum for a few days!

There's a great book called Formula 1 Technology by Peter Wright (SAE International) where he explains and demonstrates a g-g-V diagram, which is basically longitudinal G-force, lateral G-force and speed. It's three dimensional and looks like a parabolic funnel with one vertical side cut off pretty flat (acceleration). It's car specific because of downforce, which is why speed is plotted, and even for our bricks we know that the downforce changes the grip.
This makes the theoretical side of using data very difficult because the modelling is so hard to do.

However, on the practical side, I'm all for using the data to analyse what's going on in the car as I change things and try to go faster - with one caveat. You have to be experienced enough, and consistent enough, to use the data. If you can't feel that the car is sliding out from under you mid corner, or that it's pushing like crazy on turn in, or whatever, the data will just drive you crazy. If you maximize g-force all the way through the corner, but you turn in ten feet early, the data won't help you.

Random stuff:
back in this post someone talked about getting to WOT before unwinding the wheel, continuing the turn under full acceleration. Well, if I could stand on it before I unwind the wheel, then obviously I wasn't going fast enough in the first place - if I'm at the limit, then there is no more grip in my tires, and something has to give when I accelerate. Usually the hammer end in my case

Just my .02

ColorChange

Bruce: I'm a huge fan of the F1 Technology book and have quoted it here often. I definitely agree with your comment about full acceration early in a turn. How do you like the Diasio?

RJay

Didn't say that, said that this measurement seems to me to be a result, like a segment time and therefore, knowing all the values that went into producing it seems far more valuable to me.
Coincidentally, the software I work on has recently been incorporated in a telemetry system in use at the highest levels of motorsports so I have recently had a glimpse into to how this stuff is used by the very best in the world (can you tell I really can't say much about it?). What I can tell you is that these guys primary interest is adding ever more sensors to the car. There are dozens of engineers looking at the data in real time both in testing and on race day. Its pure speculation on my part, but I suspect on race day these guys are relaying information to the driver for in-car adjustments to suspension, brake-bias etc., i.e the component tweaking that results in lower lap times (or higher g-sums if you prefer). In the end, these guys are looking for the "unfair advantage", from what I can tell, at the top levels, its far more about maximizing the car, than the driver, perhaps, because at this level, driver skill is a given.

FixedWing

I and a number of friends have recently bought, or will soon buy, the Race-Tech DL1 to use at the Nürburgring. So I find this a very interesting discussion.

Reading through the comments, I think those who suggest that the data can be analysed in isolation have missed the boat. Rather, I agree with Larry’s comment that the data is most useful when being used to compare one driver to another.

The ideal situation, in my opinion, would be same car, same corner and same conditions. Then I think that the way to approach it is to look at section times to see how well various drivers are doing on different parts of the track. Only then is it useful to look at the detailed specifics to see how one person might be handling a corner better than another.

Most of us have a pretty intuitive feel for the limits of traction on a tyre. This is a relatively easy skill to master. So I doubt g-loading alone will be useful. The g-sum formula makes the mistake of summing up all of the data where what you really need to be doing is analysing specifics. This is exactly the same problem as with lap times which only look at the macro picture. Rather, I think the key is going to be in looking at braking points, turn-in points, apexes, etc. and the various sacrifices a driver makes in an individual corner.

Someone else mentioned that it is necessary to compare with drivers who have decreasingly better lap times. I don’t think this is true. By using lap segments it is possible to compare just short segments of tracks. I think we can all learn from each other as some of us will always do better on certain parts of the track than on others.

One of the things we are doing is to set up a database of data from different drivers for the Nürburgring. That way we will all have the data we need to analyse. The same might be a good idea for yourselves for the tracks in North America.

Stephen

forklift

This is a good example of comparing data of two different drivers in one corner:

http://molinier.thomas.free.fr/msdrivingstyle.wmv

RJay

Fabulous! But wheres the g-sum trace?

carreracup21

I think the trace of Schumacher's speed tells the story. He hit his max speed for the corner right at entry and kept it there very level all the way through vs. the other car which dropped speed shortly after entry and could never regain it. Assuming a constant radius turn, a corner speed trace and g trace will look the same. If you are trying to be as fast as possible, the entry appears to really separate the men from the boys, and in that regard you have to get on the gas early to maintain speed otherwise you are playing catch up in a very high drag situation with lots of lateral g trying to decel the car. It was also interesting how twitchy Schumacher was on steering input playing the balancing act on the edge of traction

brucegre

But that's the point - it's not a constant radius corner. It never is, unless you want to give up exit speed. What you see from Shumi is consummate skill in judging his corner entry speed and in holding it. We're all taught to do this, but not many of us ever get that good.

Colorchange, the Diasio is a hoot to drive. In the dry, anyways. It was my first real race car, I had to relearn the meaning of immediate input, but I love to drive it. The Cup car is way faster, but the Diasio is just as much fun, and it looks like a 962 to boot!

carreracup21

Yeah, I hear a lot about the best line through a corner and what it's supposed to be, slow in fast out that sort of thing. I'm beginning to think some of that is wrong in some instances. I think the line Schumi took through that corner was pretty much a constant radius and that's why his speed was pretty consistent throughout the corner. A long fast turn like that one at 160 mph, you are not going to try and give up any speed on entry in hopes of making it up on exit. The fastest line for that type of corner will be the maximum constant radius allowable for the corner using up all allowable track width. If you think I'm wrong it's ok, because I would have thought so too before I read this article. I think what you are talking about works for slower shorter corners though. You shift the corner to favor a later apex and opening radius on exit. You have to have a lot of reserve power to overcome the g's though and a momentum type car with low hp would probably do better just maximizing speed through the whole corner rather than giving up speed early.

Larry Herman

Seems to me that what Michael did in Becketts was to back out of the throttle a little as he bent into the turn, and then fed it back on again as he passed the apex. Pretty much what a Indy car driver would do on those big high speed sweepers. Johnny Herbert drove more of a classic late apex; slow the car more, turn it harder and get on the power sooner, harder. What Michael is doing here makes sense (to me) because of the tremendous acceleration that those cars have, plus the fact that the aerodynamics are such that the higher you keep your speed, that faster you can go through a corner. There is no question, too, that Michael was more at the limit of the grip at the front of the car than Johnny was. Real-time DAS with video...that's the ticket!

RJay

Totally agree. I think thats the big lesson of this video. Both drivers employ SIFO, but while JH is later to lift and earlier back to full throttle, he's slower overall. I would be facinating to know just how different the lines were between the two of them. I suspect it is more measured in inches than feet. Strikes me Herbert did (to a lesser extent) what I know I've done far too many times, pushed a little too hard on the way in, overslowed and paid the price. It actually appears to me that MS was feeding throttle back in just before the apex (admittedly impossible to tell) which suggests that he might have been able to carry a tiny bit more speed in, so again in essense SIFO.

In the context of the discussion on g-sum, while its clear MS's is higher, whats interesting and instructive here is that we can see why it was higher. I'll certainly be paying a lot of attention to my throttle traces next year! Hmmm.... I wonder if I should spring for that steering position sensor. Can't wait to add that video module!!!