Race Technology

I have fitted a steering angle sensor and I am wondering how to calculate the angle of the steered wheels.

I have got a user defined variable which, using a formula of corner radius and g gives me the required steering angle and I want to compare this against actual steered angle to give me a handling channel.

I plan on putting the front wheels on turnplates to callibrate the sensor but, due to ackerman, the angle of the nearside and offside wheel will be different at any given steering wheel position.

Logic tells me to use the heavilly weighted (Outside) wheel for callibration - is this correct or do I use a mean value from both front wheels?

thanks.

I'd start with the outside angle, but you might find the average or inner gives a better result for your comparisons. Calculated steering angles based on corner radii etc are incredibly inaccurate really, as there is no such thing as 'corner radii' - it's changing ALL the time based on line, bumps, camber, wind, tyre condition, driver inputs etc etc. If you can get it to within +-1° you will be doing incredibly well.

We have found that the required steering angle is very accurately when you drive the car well within the vehicle limits - this is done on the grid formation lap prior to a race or a period behind the safety car.

We have the sensor for sterring input in volts and when you adjust the scale it overlaps the required steering angle nicley- knowing it in degrees would be more helpful for some theories we have though?

I have a string pot sensor for steering angle. I was going to use it to look at my driving smoothness and don't care so much about actual degrees. Though I was going to check it for how linear it is. I was just going to put the car on jacks and turn the wheel 90* or so and take a reading, then build a calibration curve from the logged volts. Close enough to see how smooth I'm not.

osborni wrote:I have a string pot sensor for steering angle. I was going to use it to look at my driving smoothness and don't care so much about actual degrees. Though I was going to check it for how linear it is. I was just going to put the car on jacks and turn the wheel 90* or so and take a reading, then build a calibration curve from the logged volts. Close enough to see how smooth I'm not.

What sensor did you purchase the RT one or outsourced?

celesco SP 25" stroke.

I haven't had time to debug it, but it reads perfectly when not mounted and I pull the string but gives funky readings when mounted up. Don't know what's up with it yet.

Keep us informed of your progress!
What was the cost of that unit?
Regards Michael

I go the thing hooked up. Figured out the hard way that I had wired it wrong. I did a calibration "drive" down my driveway and back up.

I'll post a screen shot of the results in a few days.

String pot fitted to steering column - works perfectly.

Analysis results of a string pot

I did the formula for "Turns". +1 = one full turn to the right, -1 = one full turn to the left, etc.

Hey, I just did the exact same Celesco SP2 install onto the steering column of my Lotus Exige, with the same goal of comparing steering angle and required steering derived from corner radii.

What i found was that at anything other than very low speeds, the apparent steering angle on track was way more - like more than 40 degrees more than what would be implied by the corner radius ...... but then I remembered slip angle

If you assume a slip angle of 2-3 degrees at the wheel, and a steering ratio of 15.8 (in my case), then the wheel input will be required steering + 2-3 * 15.8, or my observed 40 to 50 degrees off.

So, the data you collect won't be as neatly organized as you think, but it is possible to build a correction into the data to scale it somewhat: in my case, I fooled around with various ways and ended up using:

STEERING ANGLE = (-497.22 * x) + 1678.78 which is my Celesco SP2 input calibrated to output in Degrees

STEERING_NEEDED = 15.8*atan(sign(VAR_0004)*2.301/VAR_0052)*(360/(2*3.142))

LAT_G_ADJUSTED_STEERING = STEERING NEEDED +(VAR_0004/1.4)*60

where VAR_0004 = LAT_G and VAR_0052 = corner radius

but many similar approaches will work - they are all "fudges" to try and fit the data - the aim being to better scale the steering input and steering needed so you can spot oversteer / understeer and corrections etc. and you will certainly see these very easily

You can also work back the other way and define the difference between Steering Needed and Steering Angle Sensor, divide by Steering Ratio = estimated Slip Angle. Plotting this vs. LAT G is interesting - very linear on corner entry / exit but therafter becomes much more noisy toward the limit of grip, above say 3-4 degrees of slip angle, where steering input becomes increasingly divorced from steering response For example, this might show you where you are over-using the steering wheel and scrubbing as a result

I'm looking at the equation for steering needed. I'm assuming that 2.301 is your wheelbase (in meters), is that right?

Paul

Yes:

STEERING_NEEDED = STEERING RATIO * ATAN(WHEELBASE/CORNER RADIUS)

The SIGN(VAR_0004) factor ensures -ve for LH turns and +ve for RH turns

The (360/2 * 3.142) factor converts from Radians to Degrees

Thanks for the feedback Chris,

I have calibrated the streering sensor by using turnplates on the wheels to measure the angle the front wheels turn rather than degrees of steering wheel movement.

OK - angle at wheels should equal:

angle at steering wheel / steering ratio

Is this what you found?