yaw rate from wheel speed?

[forestdweller]

Hello. This is my first post in this forum so please forgive me if my question is extremely elementary.

In the February '09 issue of Racecar Engineering there's an article by Danny Nowlan about race car stability. In it he makes a reference to yaw rate being deduced from differential wheel speeds. I'm assuming he's referring to the difference in wheel speeds on the non-driven end of the car. Is that something that's commonly done with usable results? It seems like there's a lot of room for error from chassis slip angle, wheel slip from braking, etc.

-Chris

[dtype38]

Welcome to 10-10ths forestdweller. You'll find a range of users in this area, some of whom might think it's a straight forward question, but most (like me) will be interested in to see if anyone can post an understandable answer. Good question!

[gregpe]

I don't think is usefull...
What about slip angle and when you lock brakes...
Gyro is better solution ...

[silente]

Hi

I read the article you talk about.

What Gregpe say is true, the oncly way to get some real measurment of yaw speed is a gyro sensor.

But i've tried to measure yaw speed with the difference of front wheel speed. Saying i had not a gyro on the car to make a comparison, in my opinion the biggest problem is when the driver locks a wheel.

From a calculation and a comparison with another car with gyro, the result was not so far away, but if the driver locks, then you have really really bad situations in your data..

Slip and steering angle are usually very small, so in first approximation can be neglected..

[forestdweller]

Thank you for the replies. What I'm hearing is that it's something that might be worth setting up a math channel for as long as one watches out for wheel lockup or liftoff, deals with small chassis slip angles, and understands that it's an approximation at best. At least until I can get ahold of a proper yaw sensor.

-Chris

[gregpe]

You can use Subaru Yaw sensor work fine and is cheap.

[speedsense]

Yaw from wheel speed is possible, though it is similar in that it lacks accuracy like calculating Long G from wheel speed.
Simply put, the path of the tire on a chassis at the limit, does not follow the path direction of the wheel it's connected to. The definition of how much yaw and the actual number of the measured amount of yaw is effected by several effects when calculated by wheel speed.

1)The type of differential has an effect on the rear wheels and their difference in speed.
2) Tire pressure, tire growth and tire flexiblity effect each tires speed number.
3) The track surface acting on the tire, example inside tires on concrete, outside tires on asphalt, the grip difference.......
4) As mention before, wheel lock, wheel spin and thirdly raising an inside wheel into the air....
5) "Jitter" from the measurement of wheel speed. As with all wheel speed measurements, there's always alterations to the number (see #2), and most systems have some filtering and averaging of the signal to get rid of the "jitter" to make the signal readable. (Using GPS speed AND wheel speed together shows this clearly)

A highly skilled driver in reasonably well handling car will rotate a car WITHOUT creating a slide or wheel spin and will do so within the limits of the slip angle of the tires.

If you want an accurate number of car rotation, then actually measuring it with a yaw sensor is best and you can even add to that two GPS units, one in the front, one in the rear, operating at 20hz with a differiental beacon located within the racetrack, to give a fairly close reading of the "heading" of both ends of the car. Then you could add an optical sensors to read slip angle of each tire.
Only then would be it really close to an accurate rotation number, but I did just ring up about $10-12k euros worth of stuff in less than a paragraph....
maybe the math would be cheaper and you would be proud to not spend the money and you could agrue that your math is correct and the driver's wrong in believing that the car is over rotating....driver's are cheaper to replace than math equations...so is not buying sensors....
There's still no excuse for having accurate data to work with, we all know where assumptions lead to .....

If you believe a calculated Long G channel is accurate, then you would also believe that a formula ford could attain 2.-3 g's under braking. When actually measured with a g sensor, you would know that 1.6-1.8 is reality.

The data acquisition community deals with accuracy that comes from correct and measured numbers to make any math equation worth using. Any variation in the collected data throws the equal sign into a greater than/less than or equal to answer.
No thanks, I'll stick to equals to,... even it if costs more money to get there..