Measuring Slip Angles

[Mike Macie]

I always thought if i knew the slip angles that i would be able to do a better job of optimizing LLT based on the tire data. Specifically for skidpad. If you recorded steering input while doing a constant radius test you should be able to get a rough idea of the steady state slip angle for each tire. Due to other factors and from testing I'm sure the car will perform better by adjusting the LLTD. But from an initial design standpoint i thought it might be close or better approach then other rules of thumb about how your LLTD should be this much more then your roll stiffness distribution. Anyone ever try to measure slip angles? Thoughts?

[Zac]

You can get a rough idea of slip angles from a really rudimentary system (like a Vbox and steering wheel sensor). But at high levels of lateral acceleration compliance effects are really significant.

[PBnG]

Corrsys Datron makes some really nice optical slip angle sensors. OEMs can occasionally be caught using less elegant slip measurement methods.

[JamesWolak]

Was just talking about this an hour ago with my co worker.

Found this when i was looking for a price on some.

http://www.corrsys-datron.com/Support/PDF_Downloads/CDS_OptimumGSlipAngles.pdf

[exFSAE]

Slip angle sensors are the only good way to do it, particularly with the amount of compliance typical in these cars. $25k a piece. There is at least one team with access to them though.

Trying to correlate it to tire data is probably not worth your time. The TTC tire data isn't bad, but it's also not great. And if you're down to splitting hairs with slip angle sensors... probably not worth it.

TLLTD front bias is just there to ensure when you go out the first test day you'll have initial U/S under no longitudinal acceleration. It takes all of minutes (or seconds if its cockpit adjustable) to change that and the balance to better suit the driver and course.

Much cheaper than setting up slip sensors.

[Mike Cook]

There are a lot of ways to measure slip:

1) Nice optical Daytron Slip sensor (can be had for less than 10k$ if your really nice to daytron)

2) V box (uses two gps antennas spaced apart to figure out heading and path which it uses to calculate slip angle. Also calculates pitch or roll angle depending on the orientation of the antennas)

3) drift box (uses a single gps sensors and a yaw sensor (integrated to find yaw angle - not that accurate I don't think, probably for drifters(duh) not much resolution)

4) A 5th wheel. Basically a tire hooked to the back of your car with rotational pots on it so it can measure path angle vs chassis angle (essentially slip) --you could probably make something like this and get decent results (cheap)

I use slip sensors a lot. I even used them on FSAE cars. Not that useful really, it basically tells you what you already know (at this level at least). I guess one question it can answer, is, are our cars operating at the slip angles that tire data says our tires will make peak grip?

Understeer sucks. Make sure your LLTD is a lot more at the rear than your weight distribution. I like to use springs for adjustment, just because I can get them for free and thats easier than making sway bars. What I do is set the car up with a base spring, then I take all of my other springs I have available and swap them out, one set at a time. I adjust the spring perches for each set as to keep my corner weights and ride heights. I write down my perch heights and then when I'm in the field, I don't need and scales at all to change springs. Makes for a simple quick way to tune the car.

Adjustable sway bar is nice too.

[Rickertsen2]

There are at least two SAE papers out there that describe methods for estimating slip angle based on more cost effective sensors.

[exFSAE]

You can get vehicle sideslip angle from above methods, but not tire slip angle...

[rjwoods77]

Wasn't UWA doing something involving carbon paper under the wheels before the design event. I think I read about that in a racecar engineering. Dont know if it was for this purpose but the article seemed to think is was very clever.

[Mike Cook]

Originally posted by exFSAE:
You can get vehicle sideslip angle from above methods, but not tire slip angle...

All it takes to get from vehicle side slip to tire slip is math and a yaw sensor.

[Garlic]

Originally posted by rjwoods77:
Wasn't UWA doing something involving carbon paper under the wheels before the design event. I think I read about that in a racecar engineering. Dont know if it was for this purpose but the article seemed to think is was very clever.

Could not possibly have anything to do with measuring slip angles.


Originally posted by Mike Cook:
All it takes to get from vehicle side slip to tire slip is math and a yaw sensor.

What do you need the yaw sensor for? It's just straight math, right?

I think the question of exFSAE is compliance. the joints of racecars don't usually have much compliance, however the parts design of most FSAE cars unfortunately does.

A kinematics and compliance test would help out with this (and a good idea anyway) but you'd still have to know the exact loading situation.

Nothing beats an actual measurement.

[VinceL]

@rjwoods77

UWA was looking at the shape of the tire contact patch under different loads, cambers, pressures, with the carbon paper.


To go between wheel slip angles and body slip angle you need to know the radius of the turn, which you can get with lat accel and speed. Or you can use yaw rate and speed. Then you plug those into an equation like eq 5.3 in RCVD. But, that equation assumes that the yaw center is at the c.g. (Yaw center probably isn't the best term to use here. What I mean is the center of rotation of the body slip angle of the vehicle - yaw center is just easy to say.) Is there a way to find the yaw center of the vehicle?

[Mike Cook]

Slip angle is really just the angle formed by long velocity and lat velocity. So all you need to know to get side slip at each axle, is to know side slip at any point a long the vehicle. Because the car is yawing, the further you are away from the cg, the more lateral velocity there is. So if you take your slip sensor, and then figure out how far away it is from your axle, and use a yaw sensor to figure out how much lateral velocity to add or subtract, you will know your axle side slip. From there, you really just need to know your steering geometry (and I suppose you need a steering pot too) and your static toes.

EDIT: Also, in the article posted above, I think they do a comparison between having one slip sensor and using a yaw sensor to figure out the axle slip angles, vs. using two slip sensors. As I recall, using a single slip sensor produced results within 20% of actual. I think you could probably do better than this too if you tried.

[exFSAE]

There is so much compliance in the wheels and suspension of a FSAE car (even the 'good' teams), the only way you're going to be anywhere close on getting actual slip and camber angles will be to measure them AT the wheels, with an optical sensor. Trying to calculate it from anything measured at the chassis is not going to be accurate.

You'd be surprised how much compliance there is in some "real" racing cars as well. Not just Nascar!

[Garlic]

Originally posted by exFSAE:
You'd be surprised how much compliance there is in some "real" racing cars as well. Not just Nascar!

Do you work at a K&C rig or something? Just wondering how you'd know this. I don't think there are many people who have any REAL knowledge of what kind of compliance is in a modern racecar.

I only know the cars I've dealt with, but it's very, very little. I don't see any successful modern racing car having significant compliance.

I do agree that most FSAE cars APPEAR to have a lot of compliance. But again, I haven't measured it (though some cars it is clearly visible on track!)

[exFSAE]

I didn't say they were successful http://fsae.com/groupee_common/emoticons/icon_smile.gif

But yea, on the FSAE topic you can see the compliance on some cars on the track.

Or if you want to see your own just put a camera watching some of your suspension and wheel assembly.

[VinceL]

@ Mike

Yeah, you're right. The equation isn't assuming anything about the yaw center. So then where is the 20% of error coming from? It can't be all from the yaw sensor. Is there an assumption with the method you described?

[Zac]

Originally posted by Garlic:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by exFSAE:
You'd be surprised how much compliance there is in some "real" racing cars as well. Not just Nascar!

Do you work at a K&C rig or something? Just wondering how you'd know this. I don't think there are many people who have any REAL knowledge of what kind of compliance is in a modern racecar.

I only know the cars I've dealt with, but it's very, very little. I don't see any successful modern racing car having significant compliance.

I do agree that most FSAE cars APPEAR to have a lot of compliance. But again, I haven't measured it (though some cars it is clearly visible on track!) </div></BLOCKQUOTE>

I work with a K&C rig. We've tested a wide range of vehicles, from Short Buses to F1 and IRL cars. In the case of FSAE, the compliances in these cars (even the podium finishers) are significant. A lot of these cars will have pretty big toe and camber changes under longitudinal and lateral load.

Yes, you can get a rough idea of your slip angles from a Vbox, drift-box, or inertia pack. Optical slipe-angle sensors are expensive after all. However, assuming your suspension is comprised of rigid members can and will get you into trouble if you aren't careful.

A big part of being a good engineer is knowing the assumptions and limitations inherent in your analysis.

[BillCobb]

disagree with the suggestion that you can't use a constant radius test to get good front & rear slip angle information (related to tire and chassis setups). That is, if the tests are done right. By that I mean:

The radius must be large enough to generate max lateral g at max power.

You must have the transducers. A slip angle sensor is a must. SWA is highly recommended. Height sensors for roll is good, too. Yaw velocity and speed are a must.

That being said, if you have SWA, then you have a rotary sensor, use a second one to measure Steering ratio versus steer angle using grease plates or air bearings. If you have slightly imperfect rotary joints, rack gears and small steer arms, you're gonna need to measure the overall ratio function.

Then you need an appropriate proceedure. I prefer constant speed increment groups because tires will adjust their temps in each group. You will miss the diff breakout, though, so run the constant ax method, too.

The whole point of testing (IMHO) is to confirm or validate the modeling and lab work you've already done. Make sure someone on your team wears glasses. Then you will know how an Optometrist gets you fixed up right: "Better or worse?" for each condition you test for.

You good Math for processing the data (data channel versus g and speed is a good start). I like spline function myself.

Then, the sideslip vs. g derivative is the rear cornering compliance (Tire + compliance) and the SWA vs. g derivative is the understeer when you normalize the SWA with you ratio function (FYI: that would be ratio function vs. SWA) You need to integrate the SR function out to the SWA in order to get the road wheel reference steer angle correctly). If you have "lumpy" and or non-linear steering effects, this is a big hitter for error fixation.

Subtract the 2 and you have the front cornering compliance (Tires + chassis).

The speed at zero sideslip (Tangent Speed) is a good gauge of how good the car really is: You want higher tangent speeds

You can do some things to explore your car: Remove compliance elements, reinforce the framework, tighten up the steering gear, better wheel bearings, lower cg, better spherical joints, stiffer uprights, larger spindles, different materials (metals). You should get the pitcher.

Have the Math guide the testing: "better or worse" ?

You will also be giving your driver some experience with the car, and the realization that for every minute of racing, there are 30 minutes for testing. Its not all track time during the race, its the prep.

The use of Tires + chassis slips is the reality of how the car is used. You really can't separate the two, but you can minimize the damage done by a bad actor (like front camber lateral force compliance in a strut).

Meanwhile, at limit handling trim, about the only thing the tires care about are the vertical loads. Slip angle gain is just about flattened out, aligning moment has gone south, maybe some camber gain left.


A driveable car is one where the response is friendly when you put the binders on, so you shouldn't go for max "Q" unless braking zones are shallow.

I've looked at thousands of K&C tests, also. IRL, F1, IndyCar and NASCAR. The amounts of compliance are small and often non-linear is overlooked. (Loose or slipping parts is a more common difficulty, and the powertrain position and bending can be difficult to believe much less fix). But, its there. I'd speculate that compliance effect magnitudes are in qualifying order.

Yes, you can tell who is Tight or Loose during a race if they will drop down to the yellow line. Butt out means they are tight.

Clever use of parts, materials and Math (you need to run the test in Math, too, BTW) give the Crew Chief a large arsenal of improvements to all vehicles, from Heavy trucks to Golf carts). (Need I mention that I ran a constant radius test on my Bass Boat)? Only problem is deciding what the wheelbase of a boat is)....

As a matter of fact, I will be enjoying retirement. Ran into Claude the other day, exchanged some good humor ("Turn the apple around").

[Buckingham]

There are a lot of ways to measure slip:

1) Nice optical Daytron Slip sensor (can be had for less than 10k$ if your really nice to daytron)

2) V box (uses two gps antennas spaced apart to figure out heading and path which it uses to calculate slip angle. Also calculates pitch or roll angle depending on the orientation of the antennas)

3) drift box (uses a single gps sensors and a yaw sensor (integrated to find yaw angle - not that accurate I don't think, probably for drifters(duh) not much resolution)

4) A 5th wheel. Basically a tire hooked to the back of your car with rotational pots on it so it can measure path angle vs chassis angle (essentially slip) --you could probably make something like this and get decent results (cheap)

You can also measure vehicle sideslip in STEADY STATE using the dual-axis accelerometer built into most data aq systems. Anytime the vehicle is yawed, the longitudinal accelerometer will pick up a component of the chassis lateral acceleration vector.

During a steady-state constant radius corner, net longitudinal acceleration is zero, so any longitudinal acceleration recorded (after filtering our the noise) is actually just the lateral accel compontent due to chassis yaw. The two accelerometer channels and some simple trig can give you a good approximation of your chassis sideslip, assuming you mounted the accelerometers square on the chassis.

[screwdriver]

Good idea but the component is rather small and will probably be swallowed by the noise.