Would You Like Tire Data?

[Brian Smith]

I have posted this same message in two other forums.

Hello everyone, my name is Brian Smith. I am working with a team of five other mechanical engineers on our senior design project. We are currently in the planning stages and are considdering different projects we might build. One of the projects is a tire testing machine. This machine will be attached to the back of a truck and driven in a straight line, maybe on an unused runway. The slip angle of the tire would be varied in lateral force test and the slip ratio would be varied in longitudinal force test (we may decide to do combined loading test). We may decide to record infrared tire temps. We will be able to vary the vertical loading, and camber between test.

These test will probably be used to generate slip angle vs lat force curves and slip ratio vs long force for various normal loadings and camber angles. If temperature is recorded that could also be a plot axis. I think that would be very usefull data to race teams.

I guess the ultimate goal of this would be to come up with the majic tire (Pacejka) coefficients for each tire we test. I also think wet testing would be interesting.

In order for this project to get off the ground, we need to know if people would be willing to donate tires or wheels. We would like to use a 13" tire 6"-10" wide. Idealy we would like to test two different brands of the same size. If you supply anything,tires or wheels, we will send you all of the data, we can put your name on the machine, whatever. You can even have input on what specific data you would like to have say, if you want data for a specific loading. If you supply tires or wheels, we can return them when finnished testing but no gaurantees on the tires being useable.

If you are interested, the university can write you a reciept for a tax credit.

Pictures
http://tntech.facebook.com/album.php?aid=2031037&l=6a572&id=52701181

Thanks
Brian

[Brian Smith]

I have posted this same message in two other forums.

Hello everyone, my name is Brian Smith. I am working with a team of five other mechanical engineers on our senior design project. We are currently in the planning stages and are considdering different projects we might build. One of the projects is a tire testing machine. This machine will be attached to the back of a truck and driven in a straight line, maybe on an unused runway. The slip angle of the tire would be varied in lateral force test and the slip ratio would be varied in longitudinal force test (we may decide to do combined loading test). We may decide to record infrared tire temps. We will be able to vary the vertical loading, and camber between test.

These test will probably be used to generate slip angle vs lat force curves and slip ratio vs long force for various normal loadings and camber angles. If temperature is recorded that could also be a plot axis. I think that would be very usefull data to race teams.

I guess the ultimate goal of this would be to come up with the majic tire (Pacejka) coefficients for each tire we test. I also think wet testing would be interesting.

In order for this project to get off the ground, we need to know if people would be willing to donate tires or wheels. We would like to use a 13" tire 6"-10" wide. Idealy we would like to test two different brands of the same size. If you supply anything,tires or wheels, we will send you all of the data, we can put your name on the machine, whatever. You can even have input on what specific data you would like to have say, if you want data for a specific loading. If you supply tires or wheels, we can return them when finnished testing but no gaurantees on the tires being useable.

If you are interested, the university can write you a reciept for a tax credit.

Pictures
http://tntech.facebook.com/album.php?aid=2031037&l=6a572&id=52701181

Thanks
Brian

[Erich Ohlde]

don't know if you know about it but someone already did this. While they didn't do this on an actual driving surface it was done on a professionally built test rig at CalSpan (?). Other than that I don't really know so I'll let someone who knows more about what was done fill you in

[John Valerio]

http://www.millikenresearch.com/fsaettc.html

info regarding the tire testing

[Brian Smith]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">regarding the tire testing </div></BLOCKQUOTE>

I was on our fsae team and am aware of the TTC. Just thought people might be interested anyways. This could be almost free data if you have a semi used tire.

[Alan]

I think this is a worthwhile endeavor even with TTC data. Doing the test on an actual road surface to me would be invaluable just to see how different it is to a belt test. My main concern would be control of the test conditions (i.e. load, slip, camber, etc...).

[BenB]

Hi Alan,

Sorry to see your boy Paul Tracy didn't do better in Portland last week (I was at the race). Although for driving with a broken back it was a testament to his ability.

I was wondering what the normal path is to getting tire data from tire manufacturers. I'm not a large manufacturer, but I am still trying my best to design a good performing car. Will they even work with me? I noticed emailing their general tech support email is not a good way to go about this (although I cant say I'm surprised)

Do you know a good place to start at Bridgestone? bbrubaker@gmail.com

[James Waltman]

Brian,
Maybe you could team up with BYU.
They had pretty much the exact same thing on their design boards at West last week:
http://dot.etec.wwu.edu/fsae/HostedPics/2007_FSAE_West/Misc_Before_Gallery/IMG_1671_BYU_Tire_Testing.JPG

[Brian Smith]

Video
http://www.youtube.com/v/_ZaC7LkyKFA

Pictures
http://tntech.facebook.com/album.php?aid=2031037&l=6a572&id=52701181

http://tntech.facebook.com/album.php?aid=2034163&l=11b61&id=52701181

more updates when I have time

[murpia]

Hi Brian,

Interesting project. Just a few technical questions...

1) Looks like you just brake the wheel, and don't drive it? So I guess you get slip angle data and slip ratio data for braking only, no traction data?
2) How do you measure and control slip ratio?
3) How do you measure slip angle, an optical sensor? How do you ensure the truck goes straight?
4) Looks like you apply vertical load using masses? Do you measure true vertical contact patch load as the tyre and the truck both bounce around?
5) Looks like there is no damping in the rig except the tyre carcass, is that right?
6) Are you testing combined slip angle and slip ratio?
7) Are you testing cambers?

Sorry if that's too many to answer...
Regards, Ian

[Brian Smith]

Good questions.

Keep in mind that this project needed to be simple to build and cheap. I believe we could have spent about 200$ provided by our department. Even though I am excited about gathering data, the goal for the class was to design, build and test the rig itself. Gathering useful data was part of getting a good grade, but we weren't required to generate a model for temperature, pressure, camber ect. But I guess that comes next after graduation...


1) Braking the wheel-Yes we only apply braking forces to the tire. Driving the wheel would have been more complex and expensive. We only get one side of the slip ratio/Longitudinal curve. I would like to add drive to the wheel later, maybe hydraulic motor. This could also make controlling slip ratio easier.

2) Measuring Slip Ratio-Wheel Speed sensors. Knowing the speed of the truck and test tire the slip ratio can be calculated (assuming no slip at the truck). It is controlled by a motorcycle master cylinder/brake caliper that you see in the video. Torque is reacted through the caliper mount into the torque reaction tube and force transducer. This torque tube creates a 4-bar linkage and keeps the tire from unloading during braking. The master cylinder is actuated by an R/C servo and it worked out nicely. The operator gradually applied brake pressure and we got a nice force vs. time curve in our data.

3)Slip angle- In the video there is a shot of our potentiometer. It is set up as a voltage reducer and this is read by the daq. The potentiometer is mounted to the cross piece in the universal joint so it is unaffected by bump travel and rotates during slip angle changes.
We dont ensure that the truck travels straight. The topic of over the road tire testing was researched before we built anything. We found a great previous example called the "Calspan 6-component rig". It is mentioned in RCVD. They were using a similar setup and determined that vehicle slip was negligible.

4) Vertical Loads- Yes we cast lead ingots to fit the machine. Each ingot weighs 25 lbs. We do not measure dynamic vertical loading. The airport we tested at was fairly flat in the test section although the ends of the runway had some bumps. If any bumps were hit, it would show in the data. I have considered logging vertical acceleration of the arm and using this as a filter during post processing of the data. In the future I would like to use an air cylinder and air tank to apply vertical loads (a zero rate spring, constant force).

5)Damping-Correct. the only damping is from the tire itself. Once again, research showed that previous rigs needed no damping. Considering that the vertical natural frequency of a tire is around 80Hz, that probably wont be excited by rotation in the 8.4 rev/s range.

6)Combined slip angle and slip ratio- No, as it is currently setup, the machine only test one condition at a time. I plan on changing this so combined slip angle and ratio can be measured. I have to modify the torque reaction tube so that it doesn't apply a moment about the vertical slip angle pivot axis.

7) Camber- It is easily adjusted on the left side truck frame mount. The right frame mount is a pivot and the left side accepts shims. We did not have time to test anything but 0deg camber.

Hope that answers everything

[murpia]

Thanks for that Brian. Now some follow-up questions...

1) How are you measuring & controlling the truck speed?
2) Are you closed-loop controlling the brake pressure to achieve a constant slip ratio, or allowing slip ratio to increase gradually by ramping brake pressure?
3) How are you post-processing the data? Filters, averaging, regressions?

What I am wondering is how 'steady-state' your test is. It seems to me so far that there will be variations in the vertical load (which is not measured) around the 'mean' of the masses used. Also the slip ratio seems to vary continually through the test, both if the truck speed changes and due to the way you control brake pressure. Seems to me at any instant in time you can't be certain where to plot each data point on the 3 axes of vertical force, slip ratio and vertical load?

I hope you don't think I'm criticising too much, I think it's great you've gone out and done this, I'm just interested to know how good the data is likely to be.

Regards, Ian

[Chris Allbee]

Just read this topic. Good stuff. I would reckon (yeah, i said it: reckon) that the quality of the data, if logged on a decent data ac., would be at least as useable to teams as data logged from a steel belt with sandpaper on it. If anything, I would love to get my hands on this data just to compare it to the TTC data. Even with fluctuations around a mean (a more realistic mode anyways) you could see any trends as compared to the TTC's info. 'sides, never hurts to have more data to try and make an informed decision with. So...who do i talk to to get it? Brian?

[Brian Smith]

All of the test were run at 30 mph using the truck's speedometer. We weren't about to make some fancy controller to make sure the truck went exactly 30. Speed "probably" has little effect on the forces the tire outputs, but I would like to find out. ..I'm trying to envision a 5 or 6 dimensional curve fit...how does pressure, camber, speed, load and temperature effect friction coefficient and cornering/slip stiffness?...

Brakes-Brake pressure was open loop and just gradually increased (by a human controlling the servo) until lockup. Worked well. We could see the test tire from inside the truck via wireless video. As you can see we also used a following car to record video.

Post Processing-I'm using the nondimentional method to create models. I came up with plots for max friction coefficient and cornering stiffness vs. vertical load. Vertical origin offset was zero for all loads due to no camber thrust/no camber, and tire symmetry. All the runs were normalized and estimated by a continuous function (typically use the magic tire formula).

Before any data analysis, results were averaged over .5 second intervals. Only decreasing slipangle data was analyzed.

Looking at increasing (fast SA change) and decreasing (slow change) SA data was interesting though. 2 different plots were visible. The faster data has pretty much a straight line from 0,0 to max slip angle and the corresponding steady state lateral force. The quasi steady state data has much more of a curve.

By the way, is it possible to visualize 4 or more continuous variables on a graph at once?

[Chris Allbee]

You could visualize 4 variables with a 3D contour plot with colors. Can't quite picture a plot capable of displaying more than 4 dimensions at once...

[Edward M. Kasprzak]

Brian,

This is a neat project you're doing, and it's definitely worthwhile in parallel with the FSAE Tire Test Consortium efforts at Calspan TIRF. You'll learn a lot about buildng a test rig and testing procedures, and you'll probably even get some useable tire data out of it.

Some of the obvious questions have been asked above, especially those pertaining to control of the operating variables. Steadyness of the test platform, tire temperatures, road surface consistency, etc. are all issues. Note that at TIRF we never eliminate all of these, although there are far less unknowns (the biggest issue at TIRF is how clean the surface is--it allows a lot of grip). You'd be smart to compare your data with the TTC data from Calspan. Don't expect perfect correlation, but if it's wildly different it's time to ask yourselves some questions to pinpoint why.

I'm glad to see that you've referenced the Calspan on-road test machines that predated (by 20 years) the TIRF test machine the TTC currently uses. Do you have any references beyond RCVD for it? And allow me to put my Milliken Research hat on for a minute and compliment you on using the Nondimensional method for modeling your data.

I'd like to have an e-mail conversation with you. Please drop me a note at my address below.

Again, nice project!

Regards,
Edward

[Dick Golembiewski]

I don't have the SAE paper handy right now, but Dave Holloway at the University of Maryland built a similar on-road tire testing rig for a NHTSA study he did about twenty years ago. I think its real purpose was to test FSAE tires, though... http://fsae.com/groupee_common/emoticons/icon_wink.gif

[ben]

Interesting topic.

At Formula Student Germany last week it was clear that lots of teams were using strain gauging to determine suspension loads.

Is anyone out there attemping to combine suspension strain gauging with a slip angle sensor to get force and moment data from a vehicle in operation?

Ben

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by ben:
Is anyone out there attemping to combine suspension strain gauging with a slip angle sensor to get force and moment data from a vehicle in operation? </div></BLOCKQUOTE>
I've been involved with an attempt to do this, but not on an FSAE car though.

It proved virtually impossible to separate 'grip' from 'noise' (to simplify the description somewhat) no matter how we attempted to filter or post-process the data. And this was for logging rates of 500Hz with very well validated anti-aliasing and good calibrations.

This was on track, however. So my recommendation to anyone wanting to attempt it would be to try using a (very) smooth skidpan and simple manouvers. Should be OK at FSAE g levels, but a previous attempt on a skidpan had made our driver so ill that a serious discussion about using a remotely operated vehicle ensued...

Regard, Ian