This year for FSAE (a collegiate race car competition) we are hoping to measure the dynamic camber angles on all four corners of our car. This requires us to measure the angle of the uprights within 0.1* while the car is driving.
Is there a reasonably easy way to go about this? I would assume an accelerometer wouldn't work since the readings would be thrown off by the driving accelerations of the vehicle, and I can't think of a way a rotary pot could be hooked up (accurately or trustworthy at least).
I am not clear as to what you want to measure: the angle of a specific hinge? One problem I see with measuring the hinges is that you can't tell whether you're on a horizontal surface or on a slope.
An accelerometer doesn't work indeed to measure the position as "upright" towards the horizon, it gives the "upright" with respect to the direction of the downward force on the car.
If you're going through a canted curve, your angle sensors and force sensors may tell your car is perfectly upright, while it's at a possibly really steep angle with respect to the horizon. So you will have to define what you want to measure, really. Is it the "upright" with respect to the road, the horizon, or to where the downward force (the combination of gravity, road slope/cant, centripetal force, etc) of the car is pointing to?
Well if your talking race cars and education NASA has a cool patent on 3D position measument of a object via single fiber optic code.
With that technology you would just need to run a fiber cable to the hub and after lots rocket science stuff you will always be able to know where the wheel is in relation to the car within 1mm.
We want to measure the angle of the upright relative to the road surface. For the most part the tracks we compete on are horizontal and flat so a measurement relative to horizontal would be sufficient.
In the past our team has attempted to calculate the camber angle based off the linear pots on the shocks and body roll; however, these measurements were disregarded as it was difficult to account for the dynamic compliance in the components (and well... students did it).
Slumpert:
Well if your talking race cars and education NASA has a cool patent on 3D position measument of a object via single fiber optic code.
With that technology you would just need to run a fiber cable to the hub and after lots rocket science stuff you will always be able to know where the wheel is in relation to the car within 1mm.
For considerable less, I was thinking you could use a TOF sensor to measure the distance change between the wheel and the car body.
If only fiber optic measurements were within the scope of our project!
What are your thoughts on the location of the TOF sensor and what it would be calculating distance changes off of? We want the "true" camber angle which would have to be measured off the upright.
Could you mount a high precision rotary encoder at one the inboard suspension joints on each corner of the car so that it reports changes when the angle of the joint changes. Then by comparing the angles at all 4 corners you may be able to reproduce the suspension movement by reference to the at-rest values.
For this I reckon you would need something a lot faster than a 16MHz Arduino.
Chrisdof:
We want to measure the angle of the upright relative to the road surface. For the most part the tracks we compete on are horizontal and flat so a measurement relative to horizontal would be sufficient.
OK, so we hold that thought!.
To be clear, we are not talking about one part of the car relative to another part of the car, including the wheel.
We are talking about a part of the car - essentially, the body structure - relative to the road surface, at speed. And you wish to assume the road surface is completely flat.
So, this is non-contact measurement.
Its accuracy is strictly limited by whether the road surface is indeed, flat. A resolution of 0.1° sounds rather improbable in this context.
At speed, sonar would not be very accurate. The only technology I could suggest would be two laser time-of-flight sensors looking at the road at different positions to the side.
Paul__B:
We are talking about a part of the car - essentially, the body structure - relative to the road surface, at speed. And you wish to assume the road surface is completely flat.
I suspect it would be reasonable to assume the surface is flat in a transverse sense - i.e. across the road. However it may or may not be level.
My interpretation of the requirement is that the OP wants to get real-time feedback about suspension angles - ultimately to know if the tyres remain flat on the surface for maximum grip.
I think the suggestion I made in Reply #5 offers some prospect of doing that - but the measurements would be relative to an at-rest calibration.
A pendulous resolver. A weight hanging off a position encoder. You may have to dampen the weights swinging, such as putting the weight in olive oil. The entire instrument can be made quite small.
Idahowalker:
A pendulous resolver. A weight hanging off a position encoder. You may have to dampen the weights swinging, such as putting the weight in olive oil. The entire instrument can be made quite small.
I can't see that being practical in a racing car with rock-hard suspension and huge G forces.
And if it was in an oil bath it would not react fast enough.
bluejets:
I doubt however that a race car would be dead level for any period of time.
Paul__B:
Its accuracy is strictly limited by whether the road surface is indeed, flat. A resolution of 0.1° sounds rather improbable in this context.
This is very true. We test on on airstrip, nonetheless I'm sure its not within 0.1* of being flat in any direction. The real need for high accuracy is that the angle of the upright shouldn't change by more than 5* under the highest loads. 0.1* might be ambitious, but it becomes hard to talk about a 1* change if it's much worse.
Robin2:
I suspect it would be reasonable to assume the surface is flat in a transverse sense - i.e. across the road. However it may or may not be level.
My interpretation of the requirement is that the OP wants to get real-time feedback about suspension angles - ultimately to know if the tyres remain flat on the surface for maximum grip.
This is what we are after. For Senior Design we need to validate our calculations and simulations with a precise measurement of our camber angles.
Robin2:
Could you mount a high precision rotary encoder at one the inboard suspension joints on each corner of the car so that it reports changes when the angle of the joint changes. Then by comparing the angles at all 4 corners you may be able to reproduce the suspension movement by reference to the at-rest values.
For this I reckon you would need something a lot faster than a 16MHz Arduino.
...R
This might just work. The mounting will be challenging as there isn't a physical axis to connect an encoder to. I'll look into it!
I like the TOF sensor idea as well. How much does surface texture affect their readings? For the best accuracy we would want them placed as close to the ground as possible, right?
Given that the upright is connected to the wheel and the lower track control arm , if you know the geometry , then the angle of the upright is directly calculable from the compression of the springs on the front wheels and the steering angle . In fact in designing the suspension geometry you have to be able to calculate that lot to design the parts .
That is readily measured with devices used in er the racing car industry.. ( potentiometer or LDVR)
If you don’t know it , take the spring off and measure the change in angle as you raise and lower the suspension . Job done.
Chrisdof:
I like the TOF sensor idea as well. How much does surface texture affect their readings? For the best accuracy we would want them placed as close to the ground as possible, right?
Correct, limited range. Surface texture - hard to say. Get a couple and play with them.
Digital camber gauges measure the angle of the wheel relative to a vertical plane. I have a Longacre digital gauge that measures up to 0.1 degree, but that of course is with the car absolutely motionless and presumably on a horizontal surface. If you could somehow duplicate the circuitry in one of the commercial digital gauges and then attach the gauge to the hub holding the wheel you might be able to get a reading, but under race conditions I am not sure how accurate it would be. Some filtering in the software could give you an acceptable reading.
Robin2:
Could you mount a high precision rotary encoder at one the inboard suspension joints on each corner of the car so that it reports changes when the angle of the joint changes. Then by comparing the angles at all 4 corners you may be able to reproduce the suspension movement by reference to the at-rest values.
For this I reckon you would need something a lot faster than a 16MHz Arduino.
...R
I was thinking something similar but wondered how to account for vehicle motion around the roll center.
You may be able to combine the readings from all 4, treat the chassis as a rigid body and determine a number of parameters including the angle of the uprights and the roll.
