I am thinking about buying a DL1/Dash3 combo
How do you detect understeer/oversteer conditions, is it possible or do you need an expensive gyro connected to the system
Understeer/Oversteer
If you install a steering wheel sensor (string pot wrapped around the shaft or a linear displacement sensor on the rack) then you can get close to and under/oversteer look without a yaw sensor.
What you know based on the GPS data:
Corner radius
Lateral G
What you can calculate based on corner radius:
The amount of required steering to make the measured corner radius.
What you measure from the car:
Actual applied steering
Wheel base
The equation is:
Required steering angle = (wheelbase / (Speed^2 / lateral G))*57.3
This is all laid out in the book: "The Competition Car Data Logging Manual"
The actual formula that is in the analysis package looks like:
((CAR_01/((VAR_0015*1.46667^2)/VAR_0004))*57.3)/20
Car_01 = 8.06667
VAR_0015 = speed
VAR_0004 = lat accel
I do this and it's pretty good. Data is pretty clear where the car over and under steers.
any big gaps between the calculated and actual steering angle is where there is under/oversteer. Red is actual, black is calculated.
I tweaked the wheel base and added in the constant at the end (57.3) to convert the steering angle into steering wheel turns. So 1 = one full turn of the steering wheel. I did this as it's MUCH simpler to calibrate as it's easy to measure turns to lock, but difficult to measure steering angle if you don't have turn plates. I did this a few years ago and can't remember the actual math I did to get the two number.
BTW - string pots are around $200 or so. You will want a ~25" long one.
What you know based on the GPS data:
Corner radius
Lateral G
What you can calculate based on corner radius:
The amount of required steering to make the measured corner radius.
What you measure from the car:
Actual applied steering
Wheel base
The equation is:
Required steering angle = (wheelbase / (Speed^2 / lateral G))*57.3
This is all laid out in the book: "The Competition Car Data Logging Manual"
The actual formula that is in the analysis package looks like:
((CAR_01/((VAR_0015*1.46667^2)/VAR_0004))*57.3)/20
Car_01 = 8.06667
VAR_0015 = speed
VAR_0004 = lat accel
I do this and it's pretty good. Data is pretty clear where the car over and under steers.
any big gaps between the calculated and actual steering angle is where there is under/oversteer. Red is actual, black is calculated.
I tweaked the wheel base and added in the constant at the end (57.3) to convert the steering angle into steering wheel turns. So 1 = one full turn of the steering wheel. I did this as it's MUCH simpler to calibrate as it's easy to measure turns to lock, but difficult to measure steering angle if you don't have turn plates. I did this a few years ago and can't remember the actual math I did to get the two number.
BTW - string pots are around $200 or so. You will want a ~25" long one.
Last edited by osborni on Fri Oct 15, 2010 12:39 pm, edited 3 times in total.
BMW 2000 M Coupe
Page 67
edit: the wheel base is in feet - to be consistent with the other units. The number is not massaged as I had implied.
It's the 57 and the 20 at the end of the formula I tweeked to get it to scale properly and overlay the measured steering turns.
edit: the wheel base is in feet - to be consistent with the other units. The number is not massaged as I had implied.
It's the 57 and the 20 at the end of the formula I tweeked to get it to scale properly and overlay the measured steering turns.
Last edited by osborni on Fri Oct 15, 2010 6:32 pm, edited 1 time in total.
BMW 2000 M Coupe
I have just found this method described on the website
anyone got that formula
Steering angle: Steering angle is a bit more involved to fit, it requires a special sensor normally a string potentiometer. This is attached to the steering column and produces a trace of how the steering wheel is used. Viewing this trace can help you identify turn in points and any oversteer or understeer moments. For those that want to get even more involved with the data logging system a neutral steering path can be calculated and overlaid on the actual steering trace highlighting very quickly any understeer or oversteer tendencies in the car, this can be very helpful when it comes to set up as its often difficult for an amateur driver to sense whether under/oversteer behavior is caused by the car or the driver.
anyone got that formula
But... after it's all said and done, what do you do with the information? If the driver can feel it under and oversteering, the logged data doesn't seem that useful. OTOH, if the driver can't feel it, how does telling him that it's happening improve things? Is it something that the crew chief uses in order to adjust balance? However, if the driver doesn't even notice it in the first place, it makes me wonder if the adjustments will help him (the driver) at all.
I'm not asking this to be a pain; I genuinely don't know how the data can be used and am hoping someone can share.
I'm not asking this to be a pain; I genuinely don't know how the data can be used and am hoping someone can share.
Not a problem.midlana wrote:But... after it's all said and done, what do you do with the information? If the driver can feel it under and oversteering, the logged data doesn't seem that useful. OTOH, if the driver can't feel it, how does telling him that it's happening improve things? Is it something that the crew chief uses in order to adjust balance? However, if the driver doesn't even notice it in the first place, it makes me wonder if the adjustments will help him (the driver) at all.
I'm not asking this to be a pain; I genuinely don't know how the data can be used and am hoping someone can share.
Right now, this something I look at after I've looked at the basics (speed, Gee, slip, sector/lap times, etc). It helps me understand why I was or wasn't faster on a particular section. Granted the sector I showed was on worn out tires, so the usefulness of this particular data set is limited as a training tool.
What i really look at it for is to see where I'm getting on/off the brake or gas and when I'm turning in or tracking out. Too soon, too late, not blended well, etc.
At least on my car with a dedicated track suspension under it, it's pretty easy to feel the attitude of the car.
There are some driver skills that can be looked at. The steering turns is essentially a "speed" measurement, so differentiating once gets you acceleration. If you move the wheel too fast, you are not being smooth.
What's missing from all of this is what the road is doing. Hitting a curb will induce oversteer and it will look the same in the data as if you went 0.2mph too fast. So you really need to look closely at this stuff to really see what's going on.
All that said, it's a secondary measurement. If I'm noticeably faster in a sector, it's usually because there was nothing dramatic going on with the steering wheel (as well as other non-dramas) - which is a lesson in and by itself.
BMW 2000 M Coupe
-
- Posts: 65
- Joined: Fri Jul 06, 2007 10:26 pm
- Location: Saint Louis, Missouri, USA
Return to “General support questions”
Who is online
Users browsing this forum: No registered users and 30 guests