Data Acquistion Systems - use and interpretation
02-25-2004, 11:20 AM
#1
Rennlist Member
Thread Starter
Data Acquistion Systems - use and interpretation
ColorChange has been quite vocal about the use of Data Acquisition Systems (for example, a DL90). I'm trying to figure out how the results should be interpreted and what we should be looking for to tell us whether or not we are doing a good job through a given turn.
Here's a graph that he posted in another thread:
Ok, some questions now:
Is the throttle position really measuring where the throttle is? Or, is it a measurement of RPMs? The slight identation in braking between 1900 and 2000 is what I think is probably a blip of the throttle in downshifting. However, we don't see the throttle position move.
The steering angle - is that also a sensor on the steering wheel? This obviously isn't from the DL90 system - could these types of sensors be attached to that? Which system is this that we are seeing?
It seems that we are seeing some high lateral loads even though the throttle is flat to the floor. Is that what we want to see?
Here's a graph that he posted in another thread:
Ok, some questions now:
Is the throttle position really measuring where the throttle is? Or, is it a measurement of RPMs? The slight identation in braking between 1900 and 2000 is what I think is probably a blip of the throttle in downshifting. However, we don't see the throttle position move.
The steering angle - is that also a sensor on the steering wheel? This obviously isn't from the DL90 system - could these types of sensors be attached to that? Which system is this that we are seeing?
It seems that we are seeing some high lateral loads even though the throttle is flat to the floor. Is that what we want to see?
02-25-2004, 12:17 PM
#2
Three Wheelin'
Join Date: Aug 2003
Posts: 1,686
Likes: 0
Received 0 Likes
on
0 Posts
Brian:
The blip you see in the g sum from 1900-2000 is me lightening up on the braking force not a double clutch downshift in this case (good guess though). If I would have blipped, you would have seen it on the throttle position. It looks to me that I locked the wheels up or came real close to it (1.0+ long g's), and that is why I eased up on the brake pedal. If I show the same turn in lap 11, it has no drop in braking but I didn't get above 1 g.
Furthermore, I should have downshifted but I was really concentrating on the fc and not keeping my car in the upper rev ranges, that is why you only see .4 long g's in acceleration because I am slugging it through the track out. I could have kicked it up to .5 or .6 had I downshifted as I should have.
The throttle position is the position of the pedal, not rpm. Same with the steering angle. I had to put a belt and gear in on the spline of the steering shaft with a rotary encoder.
Yes, you want to see as high of loads as possible. This is the g sum at the bottom. Initially, it shows only the braking g's. Then I blend in lat g's while the g sum stays high (this is trail braking), then I squeeze the throttle to the max I can to keep the g sum high as possible through track out. This is exactly what you want to do, except that the g sum should be at 1.0 and I only have it at .9 most of the time.
The blip you see in the g sum from 1900-2000 is me lightening up on the braking force not a double clutch downshift in this case (good guess though). If I would have blipped, you would have seen it on the throttle position. It looks to me that I locked the wheels up or came real close to it (1.0+ long g's), and that is why I eased up on the brake pedal. If I show the same turn in lap 11, it has no drop in braking but I didn't get above 1 g.
Furthermore, I should have downshifted but I was really concentrating on the fc and not keeping my car in the upper rev ranges, that is why you only see .4 long g's in acceleration because I am slugging it through the track out. I could have kicked it up to .5 or .6 had I downshifted as I should have.
The throttle position is the position of the pedal, not rpm. Same with the steering angle. I had to put a belt and gear in on the spline of the steering shaft with a rotary encoder.
Yes, you want to see as high of loads as possible. This is the g sum at the bottom. Initially, it shows only the braking g's. Then I blend in lat g's while the g sum stays high (this is trail braking), then I squeeze the throttle to the max I can to keep the g sum high as possible through track out. This is exactly what you want to do, except that the g sum should be at 1.0 and I only have it at .9 most of the time.
02-25-2004, 12:22 PM
#3
Rennlist Member
Thread Starter
Technically, shouldn't the g-sum really be the square root of the sum of the squares? Or, is that what you mean by g-sum?
Also, should it really be at 1 the whole time? For instance, if you are in a banked turn, you are going to see a lot better than 1. Also, in most cars (you got some powerhouses, that's for sure), the cars aren't going to be able to accelerate at 1G. I would guess in most momentum cars, we'll see the g-sum be close to 1 until the acceleration comes on. Then, we should see a steady decrease in g-sum.
Also, should it really be at 1 the whole time? For instance, if you are in a banked turn, you are going to see a lot better than 1. Also, in most cars (you got some powerhouses, that's for sure), the cars aren't going to be able to accelerate at 1G. I would guess in most momentum cars, we'll see the g-sum be close to 1 until the acceleration comes on. Then, we should see a steady decrease in g-sum.
02-25-2004, 12:29 PM
#4
Three Wheelin'
Join Date: Aug 2003
Posts: 1,686
Likes: 0
Received 0 Likes
on
0 Posts
Brian, Yes you are correct! The g sum is the square root of the sum of the squares.
Yes, you are right on your other points as well. What you will learn from your car is the limit at each particular turn, as defined by the conditions of the turn (camber, banking, weighting, fluids, ...). Most turns at Gingerman are flat so 1.0 is my limit at most places. The only thing you should see, as you do in my car, is that after I am done turning, only then does the g sum drop to my acceleration only potential of .4g (at too low an rpm).
You must have some technical training to pick this up so quickly. Are you an engineer also?
Yes, you are right on your other points as well. What you will learn from your car is the limit at each particular turn, as defined by the conditions of the turn (camber, banking, weighting, fluids, ...). Most turns at Gingerman are flat so 1.0 is my limit at most places. The only thing you should see, as you do in my car, is that after I am done turning, only then does the g sum drop to my acceleration only potential of .4g (at too low an rpm).
You must have some technical training to pick this up so quickly. Are you an engineer also?
02-25-2004, 12:55 PM
#5
Rennlist Member
Thread Starter
Computer science major with a bunch of physics and math knowledge.
What did you use for measuring throttle angle? Also, which DAS is this?
I'll have to take a look at the Skip Barber book tonight as it has similar types of graphs for a professional driver at Sebring. It'd be interesting to see his g-forces throughout a turn.
What did you use for measuring throttle angle? Also, which DAS is this?
I'll have to take a look at the Skip Barber book tonight as it has similar types of graphs for a professional driver at Sebring. It'd be interesting to see his g-forces throughout a turn.
02-25-2004, 01:04 PM
#6
Three Wheelin'
Join Date: Aug 2003
Posts: 1,686
Likes: 0
Received 0 Likes
on
0 Posts
I use another rotary potentiometer on the throttle crank coming from the throttle cable. I new you had to be technical. I am pretty confident you will see exactly what I am talking about in “Going Faster”.
The system I am using is an AimSport Chrono3 XG. They have a new one that looks nice, the MXL.
http://www.aim-sportline.com/
Now, I am really waiting for them to come out with a link directly to my 996tt ECU so that I can use all of the existing sensors on the ECU and PSM system so that I don't have to wire my car. I will know more in a few months.
If you want to install one, these are the only guys to consider, trust me!
Dave Redszus, Ph.D.
http://www.PrecisionAutoresearch.com
The system I am using is an AimSport Chrono3 XG. They have a new one that looks nice, the MXL.
http://www.aim-sportline.com/
Now, I am really waiting for them to come out with a link directly to my 996tt ECU so that I can use all of the existing sensors on the ECU and PSM system so that I don't have to wire my car. I will know more in a few months.
If you want to install one, these are the only guys to consider, trust me!
Dave Redszus, Ph.D.
http://www.PrecisionAutoresearch.com
02-25-2004, 01:29 PM
#7
Rennlist Member
Thread Starter
Yes, a link up to the OBD2 or the EDU would be great.
Anyhow, I think there's more to it than just being on the friction circle. For example, let's say that your line is bad or you aren't using the whole track. You could still be on the edge of the fricition circle and yet be slower than someone who isn't on the edge of the circle but driving a better line.
Also, let's say you are trying to be on the edge. As you come up to your turn-in point, you'll have the front of the car heavily loaded (as you are braking at threshold). Is this the ideal turn-in balance distribution for the car? No. Maximum traction for the car is when the car is near its steady state (i.e., 0 g-forces on the car).
Someone who brakes hard into the turn will get 2 tires working really hard for them while someone who eases off the brakes sooner will get 4 tires of traction.
I'm not trying to say that trail braking doesn't work as I think Mark Donahue proved fairly conclusively that it does. Perhaps it's just the magnitude and duration of the trail braking. We might be thinking a good driver is on the brakes for a relatively long time in a turn while if we got into a car with a professional, we could end up saying "that's all the trail braking you are doing? It doesn't seem like much!"
Anyhow, I think there's more to it than just being on the friction circle. For example, let's say that your line is bad or you aren't using the whole track. You could still be on the edge of the fricition circle and yet be slower than someone who isn't on the edge of the circle but driving a better line.
Also, let's say you are trying to be on the edge. As you come up to your turn-in point, you'll have the front of the car heavily loaded (as you are braking at threshold). Is this the ideal turn-in balance distribution for the car? No. Maximum traction for the car is when the car is near its steady state (i.e., 0 g-forces on the car).
Someone who brakes hard into the turn will get 2 tires working really hard for them while someone who eases off the brakes sooner will get 4 tires of traction.
I'm not trying to say that trail braking doesn't work as I think Mark Donahue proved fairly conclusively that it does. Perhaps it's just the magnitude and duration of the trail braking. We might be thinking a good driver is on the brakes for a relatively long time in a turn while if we got into a car with a professional, we could end up saying "that's all the trail braking you are doing? It doesn't seem like much!"
Trending Topics
02-25-2004, 02:52 PM
#8
Three Wheelin'
Join Date: Aug 2003
Posts: 1,686
Likes: 0
Received 0 Likes
on
0 Posts
Brian OBD2 is way to slow (max 1 Hz, usually 5 sec per update). We typically want to be at 10 Hz. Suspension and others need to be much faster. The ECU/PSM system is great if we can get at the data. Aimsport already has interrogation logging units for the P race cars ECU’s, I just need the protocols for my ECU.
You are absolutely right in that you can't see the line you took or if it was the optimal one. There are two ways to approach this. 1, try different lines and see which is the fastest. This is best done with incorporation video. 2, incorporating gps so that you can overlay lines along with the DAS data and look for the best. These improvements are coming in future models.
You are correct in that you have maximum traction when the car is neutral. But, to achieve this on the race track would be very slow. You would have to threshold brake to your entry speed, release brake pressure and wait for the car to settle, turn in where while you are building lat g's and wait for the car to settle again possible. You are leaving speed and braking capability on the table. Same thing as waiting until you leave the turn before you hit the gas. You don't do that right? Trail braking is the same thing, you take as much braking as you can into the turn and squeeze as much throttle coming out of the turn as you can. All this puts you on the friction circle and makes you fast (as long as you have chosen the right line.
You are absolutely right in that you can't see the line you took or if it was the optimal one. There are two ways to approach this. 1, try different lines and see which is the fastest. This is best done with incorporation video. 2, incorporating gps so that you can overlay lines along with the DAS data and look for the best. These improvements are coming in future models.
You are correct in that you have maximum traction when the car is neutral. But, to achieve this on the race track would be very slow. You would have to threshold brake to your entry speed, release brake pressure and wait for the car to settle, turn in where while you are building lat g's and wait for the car to settle again possible. You are leaving speed and braking capability on the table. Same thing as waiting until you leave the turn before you hit the gas. You don't do that right? Trail braking is the same thing, you take as much braking as you can into the turn and squeeze as much throttle coming out of the turn as you can. All this puts you on the friction circle and makes you fast (as long as you have chosen the right line.
02-25-2004, 03:24 PM
#9
Rennlist Member
HURRAY! We finally have a data acquisition thread going. I've been closely following the discussions on Pelican regarding the DL-90:
http://forums.pelicanparts.com/showt...hreadid=146954
So how do you decide on a DA system? As a DE driver, I want the ability to compare segment times, so I can work on each corner. I don't care about temps and pressures, just learning to drive. And I want to spend under $1000. Is the DL-90 the only system that fits that?
http://www.race-technology.com/WebPa.../DL90Home.html
http://www.d2engineering.com/public_...a%20logger.htm
http://www.cb-racing.com/logger-index.htm
http://forums.pelicanparts.com/showt...hreadid=146954
So how do you decide on a DA system? As a DE driver, I want the ability to compare segment times, so I can work on each corner. I don't care about temps and pressures, just learning to drive. And I want to spend under $1000. Is the DL-90 the only system that fits that?
http://www.race-technology.com/WebPa.../DL90Home.html
http://www.d2engineering.com/public_...a%20logger.htm
http://www.cb-racing.com/logger-index.htm
02-25-2004, 03:29 PM
#10
Three Wheelin'
Join Date: Aug 2003
Posts: 1,686
Likes: 0
Received 0 Likes
on
0 Posts
Boy, under a grand is going to be tough. I wanted the system for driver analysis. What I think you need at a minimum is speed (you can calulate long g from speed changes), rpm, lat g, steering angle, and throttle position. Other things are nice but these are the must haves. The software to easily analyze is also critical.
The D90 looks pretty good but really needs steering and throttle to see what you are doing.
The D90 looks pretty good but really needs steering and throttle to see what you are doing.
02-25-2004, 04:33 PM
#11
Rennlist Member
Since the DL-90 has 2 analogue imputs, I assume you could monitor steering and throttle. Its price is about $800 versus $900 for the Mychron 3XG, which I learned also has track mapping features.
After calling AIM, it seems that the analysis features are pretty similar to the DL-90. You can do track mapping (Gold or Gold XG), segment times of multiple laps, and theoretical best lap. It sounds like AIM would be more difficult to install, since you MUST have the wheel speed sensor (unlike the DL-90, which uses GPS and a longitudinal accelerometer instead). I wonder how their accuracies compare.
After calling AIM, it seems that the analysis features are pretty similar to the DL-90. You can do track mapping (Gold or Gold XG), segment times of multiple laps, and theoretical best lap. It sounds like AIM would be more difficult to install, since you MUST have the wheel speed sensor (unlike the DL-90, which uses GPS and a longitudinal accelerometer instead). I wonder how their accuracies compare.
02-25-2004, 04:55 PM
#12
Three Wheelin'
Join Date: Aug 2003
Posts: 1,686
Likes: 0
Received 0 Likes
on
0 Posts
Eric:
Don't know. I can say the Aim software is pretty darn nice. I would have to believe the Aim is much more accurate as even WAS gps is only good to about 3 feet, and it probably does not have WAS for that price.
You can download the software and play with it if you are interested.
Don't know. I can say the Aim software is pretty darn nice. I would have to believe the Aim is much more accurate as even WAS gps is only good to about 3 feet, and it probably does not have WAS for that price.
You can download the software and play with it if you are interested.
02-25-2004, 05:17 PM
#13
Rennlist Member
Join Date: Feb 2003
Posts: 2,819
Likes: 0
Received 0 Likes
on
0 Posts
I don't want to sound like an idiot but....
If I want to be able to analyze brake, throttle, steering, track placement, segment time, g-load, speed, and tach what would I need in a DA system? What is the cost roughly and how hard is the install? Also is there a way to keep track of tire/brake temps?
If I want to be able to analyze brake, throttle, steering, track placement, segment time, g-load, speed, and tach what would I need in a DA system? What is the cost roughly and how hard is the install? Also is there a way to keep track of tire/brake temps?
02-25-2004, 05:29 PM
#14
Rennlist Member
Thread Starter
I can tell you what I know about the DL90. I should preface with "I don't own one, but I know people who do"
The DL90 will measure G forces on two axes via internal accelerometers. It also hooks up to a GPS system (provided with it) that can give you speed measurements and basic positional measurements. The positional measurements aren't accurate enough that you could see differences in the line from lap to lap, but the velocity measurements are hugely accurate (it uses doppler shifts from the GPS satellites). It also has a wire that you can hook up to the spark plugs to get RPM measurements.
Most of this should be easy to hook up by yourself. The software is fairly simple to use and you can download the data to excel for further post-processing. The folks at http://www.apexvisualizations.com can overlay the DL90 data on top of a video camera feed. I'm also working on software that can do the same thing (I'm using a Mac, so I can't use their software)
The few things missing from the DL90: steering wheel angle, throttle position, brake pressure. The DL90 does have a way to plug in external inputs, but you have to figure out how to build the sensors on your own.
The DL90 will measure G forces on two axes via internal accelerometers. It also hooks up to a GPS system (provided with it) that can give you speed measurements and basic positional measurements. The positional measurements aren't accurate enough that you could see differences in the line from lap to lap, but the velocity measurements are hugely accurate (it uses doppler shifts from the GPS satellites). It also has a wire that you can hook up to the spark plugs to get RPM measurements.
Most of this should be easy to hook up by yourself. The software is fairly simple to use and you can download the data to excel for further post-processing. The folks at http://www.apexvisualizations.com can overlay the DL90 data on top of a video camera feed. I'm also working on software that can do the same thing (I'm using a Mac, so I can't use their software)
The few things missing from the DL90: steering wheel angle, throttle position, brake pressure. The DL90 does have a way to plug in external inputs, but you have to figure out how to build the sensors on your own.
02-25-2004, 05:51 PM
#15
Rennlist Member
Originally posted by Brian P
The DL90 does have a way to plug in external inputs, but you have to figure out how to build the sensors on your own.
The DL90 does have a way to plug in external inputs, but you have to figure out how to build the sensors on your own.
I guess I'm leaning toward the DL-90 since I want a simple install and don't really care much about steering and throttle position. Don't wanna overload my simple mind.