Witchdoctor Motorsports         Chapter 192

So you want to take real-time tire temps...


In 2005 I bought a DL1 and began to log data. I bought it primarily for car development and driver coaching (comparative data) and it has served me well from the get-go. In 2007, after many consultations with Neil Roberts I added three external IR temp sensors. I wish I could have afforded more at the time, but now that Iím used to it, Iím fine. If you have an IRS setup then perhaps youíd want the ability to log 6 at a time, but from the cars Iíve put this on, we dial in the rear mighty quickly. It is the front that takes time.


Above: The DL1 the day I got it. With only a GPS antenna and the internal voodoo magic, it will give you a lot of data.

The IR sensors themselves are quite small and very light. I sourced them from Ellen at Veracity Data. Ellen has forgotten more about DA systems than most of us will ever know. She got me the solidstateracing.com IR sensor that has, of course, been discontinued by now. Don't worry, they have replaced it. They ran me about 160$ each back then and here is the datasheet. With DL1's coming down in price (Ellen can help you, and so can Dave at SPS as he deals in new and used equipment) you could duplicate my whole system for about 1k now. The sensors have three wires; a ground (which you can gang together), a +5v power (again, ganged together from the +5v out on the DL1), and then the sensor out. So, you need five wires and Iím using WeatherPak connectors for this. Actually, I use WeatherPak connectors for almost everything since they are reliable, sturdy and allow quick disconnects. Iíd love to be able to afford Deutsch DTM, but hey, Iím a club racer. My DL1 has WeatherPaks for power and sensors, my ChaseCam has WeatherPak for power, MSD boxes, etc etc. This allows me to leave the main harness in the main car (Scratcher), yet I can quickly remove the DL1 and ChaseCam and dunk them in another vehicle with a cigarette lighter plug in already WeatherPakíed with its own harness. Understandably, this makes jumping the units around extremely simple. I have a spare DL1 antenna so the antenna in Scratcher stays put. WeatherPak kits have come down in price a lot over the years and I encourage folks to take a good look at them for wiring solutions. Even my AMB transponder is on a WeatherPak so it can come out and go in a TT car or the CrapCan car or whatever. I put a paint marker of 1, 2 and 3 on the tire temp sensors to help me keep them straight where 1 = Inboard. The last thing I wanted to do was waste time re-tuning do to a simple installation error. 


Above: One of the three IR sensors that I have. This one is mounted to the underside of the hood clamshell on Scratcher on a small aluminum shallow-angle bracket I glassed in. 

Below: All three, ready to go take some temps! 

Once youíve got them wired in, you tell the software what is on those channels and setup a quick math table so the display will show the corresponding temperature for the voltage output. Easypeasy. If I can figure out how to make the chips talk to the salsa then you can too. I display Fahrenheit because, well, we won the wars and you can take your metric system and Ö yeah. (The flip side of that equation is that some prefer C and of course you can display it that way too)


Above: The 5-pin Weatherpak connector so the hood can come off and on easily.

So with my three IR sensors plugged into the analog channels 1-3 on the DL1, I was ready to learn. I would put the three sensors pointing at the inboard, middle, and outboard of the target tiresí tread surface. In the rear this has always been simple as the sensors can be easily taped (blue tape, no residue) to an .5Ēx..5Ē angle of aluminum and simply mounted to focus on the rears. Sure, the tires go up and down a bit, but Iíve never had any issues getting rear tire temps taken on the many cars Iíve used this on. In the front, this has been trickier and I have built contraptions (Iím not a Blaine-level fabricator, so Iíll refrain from calling them fixtures) that attached to a spindle where a dust shield normally bolts, and I have also used hose clamps on the tie rod assemblies. Both of these move with the tire and were adequate solutions. They Ďlostí sensor contact at extreme turning angles, but Iíve found on track that the wheel really isnít turned that much so the data haul was fine. The best solution Iíve found (by far) is to mount the sensors directly above the tire. On a circle-track car I tuned there was a fender brace that I attached my aluminum angle to, and on Scratcher Iíve glassed in some real thin aluminum mounts to the hood clamshell that allow the sensors to be tywrapped with the ubersmall tywraps to them. Remember, you only need to capture the temps in the steering range that is used on track. If you need to, take in car video and see how much you turn the wheel. Typically, it aint much, so the Ďin zoneí placement of the sensors isnít too radical. Iíve even had to mount them a ways off of vertical (directly above tire) but since not much steering angle is used, they captured usable data easily. If you figure you never turn the wheel more than 180 at the track and youíve got a 12:1 rack and no ackerman, youíve got about a 15degree sweep from centerÖand that should be pretty easy to cover. 

*Caution: I'm not an engineer. I don't design car parts. This is merely what I've done to make my cars (and some around me) faster. It has worked for me. It may not work for you. It may anger the racing gods and they will smite you with a plague of blown head gaskets and cross threaded wheel studs. Proceed with caution, skepticism and squinty eyes.

So, now that Iíve got the ability to snag tire temperatures a test outline was in order. First off I put the sensors on the rear and made a few laps. Came in, looked at the results and made an air pressure adjustment and went back out. In short order I had temps as good as they were going to get for a stick axle car. For an IRS car it may take longer due to possible interaction between camber and toe. Rear toe has a huge effect on handling and if youíve got to add some camber and it changes toe, then you need to reset toe also. Still, it isnít too bad. Additionally, with a stick axle (even one that has some negative camber in it) youíll quickly find that you wont get even temps across the face and it becomes a compromise to get them as balanced as they are going to get.

Costas Tuning Note #1: On a stick axle rear car, once youíve found the right air pressure it takes a substantial change to get you Ďout of the zoneí. A simple damper adjustment (either compression or rebound) which will make the car different enough to drop lap times and make the driver happier will not make a difference in the temp variance. Yes, the tires may be a touch hotter or cooler but remember it is the relationship across the surface that we care about. Even a minor spring change does not seem to affect it much. It seems to take a lot of spring or a lot of swaybar before youíre finding the need to change air pressure. This applies to radial slicks and hard sidewall R-comps (HooHoos, 710s). I suspect that a soft sidewall R-comp (Ra1, R888) may need air pressure change sooner as well as true streets (of course).

So after Iíve got the rear where I feel it is good, it is time to go to the front. This will take longer (sometimes a LOT longerÖ.sigh) and remember that the only thing we care about is Ďin the turní readings. Preferably, in the Ďmeatí of the turn where you are at 90% or more of maximum lateral G. Remember what I said about the rear not needing a reset after a shock change? Yeah, forget that when you work on the pointy end of the car. Costas Tuning Note #2: On the front of the car, any handling change that is noticeable to the driver typically shows up in the temps. This includes a two-click change of dampers (either adjustment), or a one-hole change on the sway bar arm length. Think about it: ANY change you make that has the car cornering a touch flatter or at a slightly greater lean angle will result in the outboard edge of the tire working harder or not working as hard. These will be displayed in the temps captured and you can take appropriate action. Obviously, making the driver happy is a huge priority, but so far what Iíve found is that when I make the tires happy, the drivers have always been happy.

Front suspension handling interactions: As we all know, camber and castor go hand in hand. If we have a tire sitting at 2 degrees of negative camber on a spindle mounted with 7 degrees of castor, when we turn that tire 10 degrees to the left (assuming it is the RF) we have a lot more than 2 degrees of negative camber. I say this because if the temps like 3 degrees of negative camber in the most critical corner on the track, there are several ways to get there. We can have a static 2.5 degrees of negative camber and run a little bit of castor. Or, we can run 1 degree of static negative camber and run a much higher number on castor. As we turn the wheel 10 degrees both tires will have 3 degrees of negative camber at that point and the lateral grip will be the same. How important is straight line feel and how critical is braking performance? As you straighten that tire up in a static position and have less negative camber, the braking performance goes up. Noticeably. As you go down the straight the car has less self-centering leverage, but the temps stay more even. Obviously, we donít want the car difficult to control going straight, but what Iíve found is that I want to run as little castor as possible while still maintaining decent straight line stability. Iím sure youíve all read Neilís book (thinkfastbook.com and if you've not read it, you need to spend the 25$ and get it...trust me) and have the part about "everything depends on everything else" memorized by now. This is another good example. So initially, less static camber and more castor seems to be a great solution for good braking and still getting the required camber in the tire for the turns. But no. To further muddy the water: Less castor means less weight jacking. For fun, next time your car is on scales and youíve disconnected the swaybars and got the weights exactly where you want them, turn the steering wheel 90 degrees and look at the readings. Just for fun. This shocked the hell out of me when I put my Lola on scales and did it almost 17 years ago...and only later did I realize it could be an effective low-speed tuning tool. Once I got my head around the basics, more reading and chatting told me that KPI is in the mix, coupled with caster and the steering wheel angle to give us the steering camber. More mud in the waters! Tight corners with more steering input are more susceptible to the effects of steer camber than the big sweepers. Yet another tuning tool to throw in the box labeled "everything depends on everything else" (thanks Neil and Matt M.). So now I've got a much better understanding of what I can change to produce more desired temps depending on low-speed or high-speed corners as well as traditional methods that will more evenly effect both.  

Again, that last paragraph has worked for serious sidewall tires. Iíll likely be doing some testing on R888/Ra1s in the very near future and I am sure Iíll have to settle for a lot bigger temperature spread going down the straight in order to get even temps in the corners, but Iím interested to see how far Iíll have to go to achieve this.

Once Iíve got the front working well, I do a quick check on the rears again and almost every time theyíve not needed any change or just a very slight change.

Back to the front and some other things to ponder. As easy as the back ends seem to be, the more wicked and interaction-condensed the front is. Would you give up a very slight lateral grip loss in order to get a much better braking number due to much better braking control? In spec-car racing would the ability to go another car length deeper with confidence into the braking zone be a worthwhile tradeoff for a tiny loss in ultimate lateral grip? Toe is typically used to enhance turn-in. If you could eliminate toe (and the horsepower sucking scrub down each and every straight) and still have competitive turn-in, would you do it? How much is toe costing you? If a turn is banked a bit and you are getting more outboard suspension compression at 10 degrees of steering angle, how much camber is that adding? What if you are running on a banked oval? Sure, your tire temps look great on the flat corners but go all to crap when you are on a banked corner. The suspension is compressed another inch and so what changed? Was the temp change surface induced*? Or something else? Do you have some ackerman geometry messing with your camber curves at extended compression?

*I've run on some tracks that have repaved some corners and with old asphalt on some corners and new asphalt on others (and once with a concrete strip arcing through the corner)...this produces measurable differences, same as a driver can feel the difference in grip.

Lastly (and maybe most critically), this data is somewhat predicated on knowing the steering angle and knowing the driver is not overdriving the car. If I were doing this with a driver I wasnít familiar with Iíd want to ensure that I knew steering angle to qualify the data with. On some of my outlaps Iíve really worked/scrubbed/pushed the front of the car hard to help get the tires up to temp in short order and the plots for those corners look really odd. Iíd hate to chase that in a tuning situation. Tire temp sensors can rapidly develop a chassis setup, but require that the driver already be fairly qualified and not making rookie mistakes.

Let's look at some screenshots and I'll point out a few things I find interesting.

Above: Where the little circles are is about time 291.6 and my RF outboard temp is 223 (blue line), and at time 297.71 the temp is 166. That's a 57 degrees change in 6 seconds. Think about that.

Above: A little more data to play with. The car is in a left hand corner and the speed trace shows a max of about 126 on the approach. At that point the outboard sensor (on the right front tire...yellow trace) is a bit cooler than the middle which is a bit cooler than the inboard. After the turn-in and about a second after max lateral grip is achieved the outboard heats up quickly (just after the circles) to match the other two (red and black) for the majority of the left hand complex. No smoothing is turned on the software, the track is partially cambered in spots and while not overly bumpy, it is far from smooth. Passing traffic coming into the corner was a factor also (see about 5:20 in the video link below).

Want to see it all together with Video? I got TrackVision off the web and a chasecam from Dave at SPS and put the data with the video. The tach channel is scrambled intentionally. The tire temp data display is in the middle of the tach and labeled in/middle/out and the sensors are on the right front tire. The MPH display is direct from the GPS so it should be pretty accurate. The g-meter is direct from the DL1 also. 

VIDEO WITH DATA OVERLAY: http://vimeo.com/30574720

Again...*Caution: I'm not an engineer. I don't design car parts. This is merely what I've done to make my cars (and some around me) faster. It has worked for me. It may not work for you. It may anger the racing gods and they will smite you with a plague of blown head gaskets and cross threaded wheel studs. Proceed with caution, skepticism and squinty eyes.