G-Sum - questions for the proponents
12-10-2005, 10:44 AM
#16
Race Director
Chris, nobody is questioning the value of data acquisition. What is in question is the value of g-sum.
Now, before I go on, I have to admit to never having played with it personally. HOWEVER, this is the profession of one of my best friends. He has written the software for vehicle data acquisition systems and for shock dynos and has clients from amatuers to the Ferrari F1 team. When I discussed this issue with him his response was that NOBODY uses g plots. Everybody uses segment times as that is what is important.
So, I cannot (and will not) argue the finer points of this. I'll only say my info comes from someone who makes his living working with it and that is does not support g plots as a valuable tool.
Now, before I go on, I have to admit to never having played with it personally. HOWEVER, this is the profession of one of my best friends. He has written the software for vehicle data acquisition systems and for shock dynos and has clients from amatuers to the Ferrari F1 team. When I discussed this issue with him his response was that NOBODY uses g plots. Everybody uses segment times as that is what is important.
So, I cannot (and will not) argue the finer points of this. I'll only say my info comes from someone who makes his living working with it and that is does not support g plots as a valuable tool.
12-10-2005, 10:52 AM
#17
Addict
Rennlist Member
Rennlist Member
Indeed...
NOW we are getting somewhere. As Colin and Larry point out, without the heaps of flying invective fostered by Tim's rather ego-driven manner
, we can all start to see the driving forrest beyond the DAS trees. Of course, like any tool, it is useful. But like any task, there are different way of getting the nut off the bolt. Some better than others. Some requiring less/more work up front, and more/less on the backside. Use any tool at your own risk, or to your own reward. Just don't say that something flat "does not work" or "is the only way" if that is only in your own case!
I too find Acq interesting, even if I am too dense to understand most of it. I even understand the grand theory behind;
The rub is in the accomplishment of this, or the attempt to. The truth is that we all over-brake. The other truth is that the theory is just that, and that this grand paradigm is likely "impossible" to achieve with such perfection. A worthy goal, though. Just don't toss too many cars away trying to achieve it!
Yes, Larry nailed you, Chris. Ride this horse, Boy!
NOW we are getting somewhere. As Colin and Larry point out, without the heaps of flying invective fostered by Tim's rather ego-driven manner
, we can all start to see the driving forrest beyond the DAS trees. Of course, like any tool, it is useful. But like any task, there are different way of getting the nut off the bolt. Some better than others. Some requiring less/more work up front, and more/less on the backside. Use any tool at your own risk, or to your own reward. Just don't say that something flat "does not work" or "is the only way" if that is only in your own case!I too find Acq interesting, even if I am too dense to understand most of it. I even understand the grand theory behind;
It should quickly rise to the peak decel value (provided there is significant braking for this corner) and stay at or near this value until almost the track out.
Yes, Larry nailed you, Chris. Ride this horse, Boy!
12-10-2005, 10:58 AM
#18
Racer
Join Date: Dec 2003
Posts: 497
Likes: 0
Received 0 Likes
on
0 Posts
Combined G and Gsum appear to be the same thing:
sqr((latG^2)+(lonG^2))
Example of lower Gsum being faster:
90 degree corner. Normal line gives the largest possible radius. If you drive in the middle of the track, the radius is smaller. If you drive both lines at the same speed, the smaller radius will yield higher G.
I don't think this will too helpful to the experienced driver on line choice, but it could suggest a better line in a complex series of corners, such as at Mid-Ohio.
I think discussing what we think Tim said is going to be pointless and is going to only detract from the usefullness of this thread. But the posts boiled down to:
1. Look at Gsum or combined G trace to find which corners you are not maximizing.
2. Work on these corners first since there is the most to gain there.
I installed a David (combined video and DAS) in a car recently. I don't see this having much value for me, but I agree that for newer drivers, or perhaps drivers who are not inclined to study graphs, it would be useful. For me, the problem with video is that it cannot stop the action and zoom in on a one second segment. Generally an experienced driver is working in the range of tenths and hundreths of seconds and video isn't going to show that.
You have to use your head when interpreting all this. Here is a real world example:
Infineon turns 3 and 3a:
3 turns left with a severe uphill starting at the apex.
3a turns right again, with the crest of the hill right at the apex, and the exit falling away.
With the cup car, I could rapidly build up to a Gsum of 2 (which is not possible in 'normal' corners). but then the Gsum would fall rapidly almost to 0 as I transitioned from turning left to turning right (at some point during the transition you have no lateral G). Gsum would then build up again to almost 2 (since I was still climbing), but then drop way down as the car went light over the crest.
The Gsum trace looked terrible, but I was doing it at least reasonably well. With any data on any subject, you must consider the larger context.
sqr((latG^2)+(lonG^2))
Example of lower Gsum being faster:
90 degree corner. Normal line gives the largest possible radius. If you drive in the middle of the track, the radius is smaller. If you drive both lines at the same speed, the smaller radius will yield higher G.
I don't think this will too helpful to the experienced driver on line choice, but it could suggest a better line in a complex series of corners, such as at Mid-Ohio.
I think discussing what we think Tim said is going to be pointless and is going to only detract from the usefullness of this thread. But the posts boiled down to:
1. Look at Gsum or combined G trace to find which corners you are not maximizing.
2. Work on these corners first since there is the most to gain there.
I installed a David (combined video and DAS) in a car recently. I don't see this having much value for me, but I agree that for newer drivers, or perhaps drivers who are not inclined to study graphs, it would be useful. For me, the problem with video is that it cannot stop the action and zoom in on a one second segment. Generally an experienced driver is working in the range of tenths and hundreths of seconds and video isn't going to show that.
You have to use your head when interpreting all this. Here is a real world example:
Infineon turns 3 and 3a:
3 turns left with a severe uphill starting at the apex.
3a turns right again, with the crest of the hill right at the apex, and the exit falling away.
With the cup car, I could rapidly build up to a Gsum of 2 (which is not possible in 'normal' corners). but then the Gsum would fall rapidly almost to 0 as I transitioned from turning left to turning right (at some point during the transition you have no lateral G). Gsum would then build up again to almost 2 (since I was still climbing), but then drop way down as the car went light over the crest.
The Gsum trace looked terrible, but I was doing it at least reasonably well. With any data on any subject, you must consider the larger context.
12-10-2005, 11:12 AM
#19
Addict
Rennlist Member
Rennlist Member
Hmmm....
I knew I would get lost quickly. However, this might have been the most sage thing to come out of all of this.
I knew I would get lost quickly. However, this might have been the most sage thing to come out of all of this.
Originally Posted by ChrisC
The Gsum trace looked terrible, but I was doing it at least reasonably well. With any data on any subject, you must consider the larger context.
12-10-2005, 11:27 AM
#20
Racer
Join Date: Dec 2003
Posts: 497
Likes: 0
Received 0 Likes
on
0 Posts
Geo,
I am basically a professional Porsche mechanic. I make a living repairing broken Porches and modifying ones that are not broken. Therefore I have a very large collection of tools. Almost none of these tools use fractional SAE measurement. I also do not use a 1 inch drive impact gun. Does this mean that these tools are useless for everyone?
I am sure John has a bunch of oddly shaped hammers and shot bags in his tool box. He might also have an English wheel. I am sure he as a bunch of sanding blocks, paint sprayers, and orbital polishers. None of these would help me at all. But I am not going to call them useless.
If you give me a wrecked car I know exactly what to do. Load it up in the trailer and bring it to the body shop. That is the approach that is successful for me. Do I think this is the only approach?
Here is a challenge. I give you some of my data for a track that you are not familiar with. Now you look at the data and tell me how to go faster without looking at Gsum. I don't think you(or anyone else) can do it as well as they could if they did use Gsum.
If I am looking at someone's data and they lift in the middle of the straight, I probably don't need to look at Gsum data to tell them how to go faster. Maybe this example illustrates the problem you all have with Tim's approach.
John says:
"The rub is in the accomplishment of this, or the attempt to. The truth is that we all over-brake. The other truth is that the theory is just that, and that this grand paradigm is likely "impossible" to achieve with such perfection. A worthy goal, though. Just don't toss too many cars away trying to achieve it!"
All true. But it seems as if statements like this are made to persaude someone to not adopt this approach. Here is now I teach it:
Now that you can drive reasonably well here is what you need to do to go really fast. I explain friction circle etc. However, working towards this is neither easy nor safe, so I suggest that you work up to it by using these simple techniques...
It seems like you guys are just saying 'Don't go there."
Chris Cervelli
Premier Motorsports
I am basically a professional Porsche mechanic. I make a living repairing broken Porches and modifying ones that are not broken. Therefore I have a very large collection of tools. Almost none of these tools use fractional SAE measurement. I also do not use a 1 inch drive impact gun. Does this mean that these tools are useless for everyone?
I am sure John has a bunch of oddly shaped hammers and shot bags in his tool box. He might also have an English wheel. I am sure he as a bunch of sanding blocks, paint sprayers, and orbital polishers. None of these would help me at all. But I am not going to call them useless.
If you give me a wrecked car I know exactly what to do. Load it up in the trailer and bring it to the body shop. That is the approach that is successful for me. Do I think this is the only approach?
Here is a challenge. I give you some of my data for a track that you are not familiar with. Now you look at the data and tell me how to go faster without looking at Gsum. I don't think you(or anyone else) can do it as well as they could if they did use Gsum.
If I am looking at someone's data and they lift in the middle of the straight, I probably don't need to look at Gsum data to tell them how to go faster. Maybe this example illustrates the problem you all have with Tim's approach.
John says:
"The rub is in the accomplishment of this, or the attempt to. The truth is that we all over-brake. The other truth is that the theory is just that, and that this grand paradigm is likely "impossible" to achieve with such perfection. A worthy goal, though. Just don't toss too many cars away trying to achieve it!"
All true. But it seems as if statements like this are made to persaude someone to not adopt this approach. Here is now I teach it:
Now that you can drive reasonably well here is what you need to do to go really fast. I explain friction circle etc. However, working towards this is neither easy nor safe, so I suggest that you work up to it by using these simple techniques...
It seems like you guys are just saying 'Don't go there."
Chris Cervelli
Premier Motorsports
12-10-2005, 12:30 PM
#21
Addict
Rennlist Member
Rennlist Member
Not at all, Chris;
I know that the key to being truly succesful is to exploit every tool at your disposal, and if that tool does not exist, MAKE IT! Anyone who is resisting what Tim was trying to do (including me) is only doing so as a result of his insistence that you could essentially teach yourself to be fast using Acq alone. He was very reluctant to back-pedal at all and admit that talent and instruction were also of equal value, or anything else for that matter. So also were the good things he was trying to accomplish lost in a sea of knee-jerk hip shots at people that dared question his thinking, motivation, or technique. Bluntly put, he might have had some good points, and got them through, if he were not such a butthole.
I would never tell someone to simply "not go there." I am a strong proponent of free will. I am also a strong proponent of giving people the information they need to make good decisions for themselves based on full and complete understanding. In other words, what you said is what I say. "You can do this, and here is how, but you must understand that there is real risk involved. Let's start here..."
We're on the same page. Yours just has a bunch of funny scribbling on it!
I know that the key to being truly succesful is to exploit every tool at your disposal, and if that tool does not exist, MAKE IT! Anyone who is resisting what Tim was trying to do (including me) is only doing so as a result of his insistence that you could essentially teach yourself to be fast using Acq alone. He was very reluctant to back-pedal at all and admit that talent and instruction were also of equal value, or anything else for that matter. So also were the good things he was trying to accomplish lost in a sea of knee-jerk hip shots at people that dared question his thinking, motivation, or technique. Bluntly put, he might have had some good points, and got them through, if he were not such a butthole.
I would never tell someone to simply "not go there." I am a strong proponent of free will. I am also a strong proponent of giving people the information they need to make good decisions for themselves based on full and complete understanding. In other words, what you said is what I say. "You can do this, and here is how, but you must understand that there is real risk involved. Let's start here..."
We're on the same page. Yours just has a bunch of funny scribbling on it!
12-10-2005, 12:32 PM
#22
Lifetime Rennlist Member
Thread Starter
Chris,
I hope you would agree that the ability to make judgements based on combined G depends on you knowing what the performance envelope of the car is at that point in time and that point on the track. (Notwithstanding that if you take a corner at 0.5 g's, we can easily judge it to be well under the limit.)
That is where I think the real issue is with trying to use g data - either combined or lat. I assume you are familiar with Mid Ohio. What is the g limit(s) for the keyhole? What is the limit for T11? Even T1 - what is the limit there?
My assertion is that none of us know and it depends on too many things. I am not even talking about changes in grip from lap to lap but consider...
At 140 mph, my car does almost 1 g if I lift (no brakes). It will do very high combined G's at those speeds if I am mostly braking with little cornering. It also corners very hard at those speeds. I can generate the same combined g by hard braking and light cornering OR by hard cornering and light braking. Which is faster?
Or consider a car like yours which should be able to generate higher braking g's than cornering. Max combined g's will come with heavy braking and light cornering. But again, is that the fast way around the corner? Who could tell from g data?
I agree that g data will tell you if you are too slow if you are way too slow. I fail to see hos g data can really tell if you are too slow when you are close to the limit. That is a big reason that the data pros I have worked with say that they pay virtually no attention at all to g data.
I hope you would agree that the ability to make judgements based on combined G depends on you knowing what the performance envelope of the car is at that point in time and that point on the track. (Notwithstanding that if you take a corner at 0.5 g's, we can easily judge it to be well under the limit.)
That is where I think the real issue is with trying to use g data - either combined or lat. I assume you are familiar with Mid Ohio. What is the g limit(s) for the keyhole? What is the limit for T11? Even T1 - what is the limit there?
My assertion is that none of us know and it depends on too many things. I am not even talking about changes in grip from lap to lap but consider...
At 140 mph, my car does almost 1 g if I lift (no brakes). It will do very high combined G's at those speeds if I am mostly braking with little cornering. It also corners very hard at those speeds. I can generate the same combined g by hard braking and light cornering OR by hard cornering and light braking. Which is faster?
Or consider a car like yours which should be able to generate higher braking g's than cornering. Max combined g's will come with heavy braking and light cornering. But again, is that the fast way around the corner? Who could tell from g data?
I agree that g data will tell you if you are too slow if you are way too slow. I fail to see hos g data can really tell if you are too slow when you are close to the limit. That is a big reason that the data pros I have worked with say that they pay virtually no attention at all to g data.
12-10-2005, 12:40 PM
#23
Chris,
Thanks for the explanation. I understand exactly what you mean by needing to view the data in context.
Little disclaimer first: I am a beginning/intermediate driver. I have a DL1 and I acquire data and video, but I don't sweat the data much at all at this point in the learning curve because I must improve my skills to implement what I learn from data before I focus too intensely on what the data is actually telling me. I mean, you can learn from the data that you are not Michael Schumacher, but if you don't have the skills to fix the problems then you are just wasting your time obsessing over the data until you develop the driving skills necessary to "do" what the data is telling you to do. At this point, I find video to be extremely useful and data to be interesting.
If I understand the data correctly, it would be a mistake to say that you always want to maximize or minimize G-sum.
If you are trying to explore different lines, then it would be wise to drive through using the same technique/speed on the various different lines, and then pick the line that gives you the lowest G-sum.
Once you have located the correct line (and you can drive this line consistently), then the goal should be to increase speed through the corner by maximizing G-Sum. For example, if you are not at the edge of the friction circle -- max G-sum possible for the car -- during the braking portion of the corner, you did not enter the turn as fast as you could because you still had traction left to brake while turning. On the way out of the turn, if you are not at the edge of the friction circle, you either should accelerate more or, if your car will not permit, alter your line on the way out to shorten the corner.
It would seem to be fairly easy to fix "big" problems where drivers are way off the potential. I think it would be much harder to use G-Sum to do fine tuning in light of the variables of traction and potential G-Sum even on the same corner on the same day.
Indeed, I have always been troubled more by how you determine your maximum potential G-sum -- which is a necessary value for analysis -- than by the G-sum calculation itself. I mean, like Chris points out, you can pull wildly different max Gs in various places on the track (e.g., 2 Gs in the corner Chris discusses). As such, assuming a max potential G-sum of 1.2 for our cars seems a little silly. THAT is the part of the analysis that I can't get over. The G-sum may show that you were only pulling .9, which might appear to be bad unless you knew that there was water on the track at that point and you, in reality, could not possibly have driven better.
Am I thinking about this the right way?
TD
P.S. I agree with Larry that Tony and Greg's software is incredible and very useful for drivers using video and data, including newbies like me.
Thanks for the explanation. I understand exactly what you mean by needing to view the data in context.
Little disclaimer first: I am a beginning/intermediate driver. I have a DL1 and I acquire data and video, but I don't sweat the data much at all at this point in the learning curve because I must improve my skills to implement what I learn from data before I focus too intensely on what the data is actually telling me. I mean, you can learn from the data that you are not Michael Schumacher, but if you don't have the skills to fix the problems then you are just wasting your time obsessing over the data until you develop the driving skills necessary to "do" what the data is telling you to do. At this point, I find video to be extremely useful and data to be interesting.
If I understand the data correctly, it would be a mistake to say that you always want to maximize or minimize G-sum.
If you are trying to explore different lines, then it would be wise to drive through using the same technique/speed on the various different lines, and then pick the line that gives you the lowest G-sum.
Once you have located the correct line (and you can drive this line consistently), then the goal should be to increase speed through the corner by maximizing G-Sum. For example, if you are not at the edge of the friction circle -- max G-sum possible for the car -- during the braking portion of the corner, you did not enter the turn as fast as you could because you still had traction left to brake while turning. On the way out of the turn, if you are not at the edge of the friction circle, you either should accelerate more or, if your car will not permit, alter your line on the way out to shorten the corner.
It would seem to be fairly easy to fix "big" problems where drivers are way off the potential. I think it would be much harder to use G-Sum to do fine tuning in light of the variables of traction and potential G-Sum even on the same corner on the same day.
Indeed, I have always been troubled more by how you determine your maximum potential G-sum -- which is a necessary value for analysis -- than by the G-sum calculation itself. I mean, like Chris points out, you can pull wildly different max Gs in various places on the track (e.g., 2 Gs in the corner Chris discusses). As such, assuming a max potential G-sum of 1.2 for our cars seems a little silly. THAT is the part of the analysis that I can't get over. The G-sum may show that you were only pulling .9, which might appear to be bad unless you knew that there was water on the track at that point and you, in reality, could not possibly have driven better.
Am I thinking about this the right way?
TD
P.S. I agree with Larry that Tony and Greg's software is incredible and very useful for drivers using video and data, including newbies like me.
Last edited by TD in DC; 12-10-2005 at 12:59 PM.
12-10-2005, 12:49 PM
#24
Originally Posted by SundayDriver
Chris,
I hope you would agree that the ability to make judgements based on combined G depends on you knowing what the performance envelope of the car is at that point in time and that point on the track. (Notwithstanding that if you take a corner at 0.5 g's, we can easily judge it to be well under the limit.)
That is where I think the real issue is with trying to use g data - either combined or lat. I assume you are familiar with Mid Ohio. What is the g limit(s) for the keyhole? What is the limit for T11? Even T1 - what is the limit there?
My assertion is that none of us know and it depends on too many things. I am not even talking about changes in grip from lap to lap but consider...
At 140 mph, my car does almost 1 g if I lift (no brakes). It will do very high combined G's at those speeds if I am mostly braking with little cornering. It also corners very hard at those speeds. I can generate the same combined g by hard braking and light cornering OR by hard cornering and light braking. Which is faster?
Or consider a car like yours which should be able to generate higher braking g's than cornering. Max combined g's will come with heavy braking and light cornering. But again, is that the fast way around the corner? Who could tell from g data?
I agree that g data will tell you if you are too slow if you are way too slow. I fail to see hos g data can really tell if you are too slow when you are close to the limit. That is a big reason that the data pros I have worked with say that they pay virtually no attention at all to g data.
I hope you would agree that the ability to make judgements based on combined G depends on you knowing what the performance envelope of the car is at that point in time and that point on the track. (Notwithstanding that if you take a corner at 0.5 g's, we can easily judge it to be well under the limit.)
That is where I think the real issue is with trying to use g data - either combined or lat. I assume you are familiar with Mid Ohio. What is the g limit(s) for the keyhole? What is the limit for T11? Even T1 - what is the limit there?
My assertion is that none of us know and it depends on too many things. I am not even talking about changes in grip from lap to lap but consider...
At 140 mph, my car does almost 1 g if I lift (no brakes). It will do very high combined G's at those speeds if I am mostly braking with little cornering. It also corners very hard at those speeds. I can generate the same combined g by hard braking and light cornering OR by hard cornering and light braking. Which is faster?
Or consider a car like yours which should be able to generate higher braking g's than cornering. Max combined g's will come with heavy braking and light cornering. But again, is that the fast way around the corner? Who could tell from g data?
I agree that g data will tell you if you are too slow if you are way too slow. I fail to see hos g data can really tell if you are too slow when you are close to the limit. That is a big reason that the data pros I have worked with say that they pay virtually no attention at all to g data.
Mark, I was writing my post while you were writing yours. This is EXACTLY the problem I have with trying to understand how to use G-Sum.
12-10-2005, 01:13 PM
#25
Racer
Join Date: Dec 2003
Posts: 497
Likes: 0
Received 0 Likes
on
0 Posts
Sunday says:
"I hope you would agree that the ability to make judgements based on combined G depends on you knowing what the performance envelope of the car is at that point in time and that point on the track. (Notwithstanding that if you take a corner at 0.5 g's, we can easily judge it to be well under the limit.)"
For sure it would be useful to know this. For sure it is impossible to know this. Way too many variables to consider. But, we both know that the performance envelope is not going to vary too much unless there is some unusual condition such as an upgrade or poor track surface.
The absolute number is not so useful. The shape of the line is what is useful. If you can decel at 1.6 and corner at 1.5 in a particular corner, it is reasonable to assume that during the transition you should be around 1.5. I can think of conditions where this would not be true, but it is generally true.
I have no experience with downforce cars and I admit that the picture is significantly more complicated. However, I think that the process is still valid: do some laps at a good clip and examine the data for Gsum dips. If you can't account for them coming from track or car conditions, then go back out and adjust your driving.
I have not suggested using Gsum and nothing else, and I don't think Tim has either.
Sunday says:
"I can generate the same combined g by hard braking and light cornering OR by hard cornering and light braking. Which is faster?
True. The Gsum alone will not tell you. Check the split time. Occaisionally I find it useful to display Lat G and Lon G as well as Gsum. This is normally if I am getting really picky about the entry of a corner that I am already doing well.
Let's consider what the experienced driver likely needs to do to go faster. Normally it is this:
Brake later and harder
More decel at moment of turn in
Shorter delay from brake release to throttle application.
These three show up very obviously on the Gsum graph. They are not obvious any other way. You say "but the Gsum graph does not tell you how hard you can brake or corner". I say right, but it is reasonable to assume that if I can brake at 1.6 in one part of the braking zone then I should be able to hold 1.6 for the whole braking zone. If the data shows that I did not do this, then I need to either figure out why (like downforce decreasing with speed) or go out and drive better. For every dip in the Gsum graph, I ask why. If I can't blame the car or the track, I go back out and try to drive better.
I am not looking for the high number, I am looking for the dips. The dips show where I gave up performance that I'd already proven was possible.
John, your page is the one with all the scribbling.
Chris Cervelli
Premier Motorsports
"I hope you would agree that the ability to make judgements based on combined G depends on you knowing what the performance envelope of the car is at that point in time and that point on the track. (Notwithstanding that if you take a corner at 0.5 g's, we can easily judge it to be well under the limit.)"
For sure it would be useful to know this. For sure it is impossible to know this. Way too many variables to consider. But, we both know that the performance envelope is not going to vary too much unless there is some unusual condition such as an upgrade or poor track surface.
The absolute number is not so useful. The shape of the line is what is useful. If you can decel at 1.6 and corner at 1.5 in a particular corner, it is reasonable to assume that during the transition you should be around 1.5. I can think of conditions where this would not be true, but it is generally true.
I have no experience with downforce cars and I admit that the picture is significantly more complicated. However, I think that the process is still valid: do some laps at a good clip and examine the data for Gsum dips. If you can't account for them coming from track or car conditions, then go back out and adjust your driving.
I have not suggested using Gsum and nothing else, and I don't think Tim has either.
Sunday says:
"I can generate the same combined g by hard braking and light cornering OR by hard cornering and light braking. Which is faster?
True. The Gsum alone will not tell you. Check the split time. Occaisionally I find it useful to display Lat G and Lon G as well as Gsum. This is normally if I am getting really picky about the entry of a corner that I am already doing well.
Let's consider what the experienced driver likely needs to do to go faster. Normally it is this:
Brake later and harder
More decel at moment of turn in
Shorter delay from brake release to throttle application.
These three show up very obviously on the Gsum graph. They are not obvious any other way. You say "but the Gsum graph does not tell you how hard you can brake or corner". I say right, but it is reasonable to assume that if I can brake at 1.6 in one part of the braking zone then I should be able to hold 1.6 for the whole braking zone. If the data shows that I did not do this, then I need to either figure out why (like downforce decreasing with speed) or go out and drive better. For every dip in the Gsum graph, I ask why. If I can't blame the car or the track, I go back out and try to drive better.
I am not looking for the high number, I am looking for the dips. The dips show where I gave up performance that I'd already proven was possible.
John, your page is the one with all the scribbling.
Chris Cervelli
Premier Motorsports
12-10-2005, 01:20 PM
#26
Originally Posted by Premier Motorsp
I am not looking for the high number, I am looking for the dips. The dips show where I gave up performance that I'd already proven was possible.
That is the single best answer to my question about G-Sum that I have heard. I think Tim was using this in his explanations, but he made it sound like he was also focusing on the values themselves, which seem less helpful to me. Thank you.
TD
12-10-2005, 01:26 PM
#27
Addict
Rennlist Member
Rennlist Member
Originally Posted by Premier Motorsp
John, your page is the one with all the scribbling.
Since we all participate here we know how to read English; MY scribbling, for what that is worth. It's the XY plots that get confusing to ME. For what THEY are worth!
So what you are looking at is not totals, but deviations. Or "Delta" in geek-speak? I get it. You then ask, "why the change?" Now, as always, it is up to the human interface to answer that. I don't imagine the data can tell you why, how, or even IF the anomaly is workable.
12-10-2005, 01:30 PM
#28
Addict
Rennlist Member
Rennlist Member
Hmmm...
It suddenly occurs to me... that it is interesting that in the study of crash data, the forward velocity is of no real importance. It is the Delta, i.e. the change in velocity from linear to lateral, that is key. I don't know if there is ANY relationship whatever to what we are disccussing with Acq, but the counterintuitive nature of both is interesting to me.
It's not the numbers themselves, but the change in them...
It suddenly occurs to me... that it is interesting that in the study of crash data, the forward velocity is of no real importance. It is the Delta, i.e. the change in velocity from linear to lateral, that is key. I don't know if there is ANY relationship whatever to what we are disccussing with Acq, but the counterintuitive nature of both is interesting to me.
It's not the numbers themselves, but the change in them...
12-10-2005, 02:12 PM
#29
Rennlist
Basic Site Sponsor
Basic Site Sponsor
Originally Posted by Premier Motorsp
The absolute number is not so useful. The shape of the line is what is useful. If you can decel at 1.6 and corner at 1.5 in a particular corner, it is reasonable to assume that during the transition you should be around 1.5. I can think of conditions where this would not be true, but it is generally true.
do some laps at a good clip and examine the data for Gsum dips. If you can't account for them coming from track or car conditions, then go back out and adjust your driving.
For every dip in the Gsum graph, I ask why. If I can't blame the car or the track, I go back out and try to drive better.
I am not looking for the high number, I am looking for the dips. The dips show where I gave up performance that I'd already proven was possible.
do some laps at a good clip and examine the data for Gsum dips. If you can't account for them coming from track or car conditions, then go back out and adjust your driving.
For every dip in the Gsum graph, I ask why. If I can't blame the car or the track, I go back out and try to drive better.
I am not looking for the high number, I am looking for the dips. The dips show where I gave up performance that I'd already proven was possible.
__________________
Larry Herman
2016 Ford Transit Connect Titanium LWB
2018 Tesla Model 3 - Electricity can be fun!
Retired Club Racer & National PCA Instructor
Past Flames:
1994 RS America Club Racer
2004 GT3 Track Car
1984 911 Carrera Club Racer
1974 914/4 2.0 Track Car
CLICK HERE to see some of my ancient racing videos.
Larry Herman
2016 Ford Transit Connect Titanium LWB
2018 Tesla Model 3 - Electricity can be fun!
Retired Club Racer & National PCA Instructor
Past Flames:
1994 RS America Club Racer
2004 GT3 Track Car
1984 911 Carrera Club Racer
1974 914/4 2.0 Track Car
CLICK HERE to see some of my ancient racing videos.
12-10-2005, 06:17 PM
#30
Banned
Join Date: Apr 2004
Location: KY
Posts: 397
Likes: 0
Received 0 Likes
on
0 Posts
Chris, very clear answer, glad we got there. I'm just turning into a data nut, now that I have enough seat time that it will be useful.
I agree, it's the dips, and the fact that they cause you to ask "why?". I was just reading Ross Bentley again, and one of his warnings is to not become a g-junkie. There are corners where we can pull higher lat g's, but it actually makes the corner slower (tighter radius when a larger radius is faster). But in the the context of a knowledgable driver, really trying to understand what is happening in a particular section of the track, data is invaluable.
I agree, it's the dips, and the fact that they cause you to ask "why?". I was just reading Ross Bentley again, and one of his warnings is to not become a g-junkie. There are corners where we can pull higher lat g's, but it actually makes the corner slower (tighter radius when a larger radius is faster). But in the the context of a knowledgable driver, really trying to understand what is happening in a particular section of the track, data is invaluable.