NASCAR’s Laser Inspection Process: Confidence and Accuracy

There Will Be Rules…

The last time I was at the R&D Center, I was given a copy of the official NASCAR rule book for 1948.

It was about a quarter of a page.

Now there’s a thick booklet filled with detailed specifications, part numbers, measurements, tolerances and AutoCad drawings. Weekly bulletins augment and update the rule book.

It’s not really NASCAR’s fault. It’s the teams’ jobs to push the boundaries. It falls to NASCAR job to ensure the “level playing field”. So they make rules. The problem is that for the last two weeks — when the focus should have been on the Chase — people have been talking about The Rules.

How Did We Get to Lasers at the Racetrack?

nascar_chassis_700x422Just a disclaimer: I like technology everywhere but on the track during practice and races. I don’t want to see drivers adjusting fuel/air ratios from their seats or a slew of engineers driving the car from the pit box. But I think it’s the kiss of death to stay stuck back in the 1980s when it comes to the car. Motorsports faces a huge challenge in the future as people are less and less interested not only in racing, but in cars at all. Just my opinion.

The COT Chassis

In my view of NASCAR history, growing concerns over safety and the solutions advanced technology could provide really pushed NASCAR to embrace high-tech. The Car of Tomorrow was far from ideal, but the underlying motivation for it — the tube chassis that makes the current car the safest one we’ve ever had — is a major advance for NASCAR.

NASCAR literally sent an AutoCad (a 3D drawing program) file of the chassis to the teams and said “This is what it has to look like”.

They needed technology to monitor the new requirements. They utilized coordinate measuring machines (Romer and Faro Arms) a to certify every chassis a team builds. Once certified, NASCAR affixes RFID tags at strategic points along the chassis. They scan the RFID tags at the track to verify that the chassis is certified and hasn’t been changed since certification.

The Body

After the COT chassis was implemented, NASCAR started getting concerned over the amount of money spent in wind tunnels to twist and warp the body, so they mandated a fixed body shape.

The first way to measure the body shape was with metal templates (“the grid”) similar to the ones shown in the video below (which was filmed at Hendrick Motorsports).

Forced to stick with a body shape that was tested at a pretty large number of points, teams started looking at what they could do in the regions that weren’t being touched by the grid. They were looking at smaller parts of the body. Teams started using 3D laser scanners to measure the body to the level of thousandths of inches. If a car runs particularly well, they look and see if there’s some small difference they can exploit for even more speed.

NASCAR also has laser scanners and will use them to measure bodies; however NASCAR doesn’t use a laser scanner to measure the body at the racetrack. They still use the templates. Why?

  • The templates do the job relatively quickly. There’s a process in place and it works.
  • The minor things a laser scanner might catch aren’t worth the time and expense.
  • It’s a technique teams can easily replicate at their shops, so they have confidence in the measurements.
  • It’s easy for a crew chief (or anyone else) to see where the problem is if a template doesn’t fit.
  • It gives you a clear yes/no.
  • If a question arises, they can take a car back to the R&D Center and do a more detailed scan.

The Underbody

The Laser Inspection System (LIS) was introduced at Daytona in February 2013 to replace two separate stops on the inspection line that measured suspension-related parameters such as caster, camber, toe, rear-axle location and wheel base (and more!) You can see what the station looks like here.

Prior to the LIS, NASCAR inspectors needed ten or twelve different instruments to make all those measurements. Measurements were made at different times and different places and the car was moved between measurements. As the rear suspension became one of the few places where teams could experiment without penalty, efforts to increase car skew accelerated. Teams got more and more “creative”. And NASCAR needed a better way to inspect the underside of the car.

Enter the LIS.

The LIS system is a platform the car is pushed onto. The system uses the RFID tags on the chassis to center the car, then makes a series of measurements using lasers (explained here) that takes about two-and-a-half to three minutes. The results go to a computer monitor.

bspeed_laserinspection_displayresults

The range of allowed value is in green. The blue lines show where the car was measured to be. There’s a button on the left of each measurements. If it’s red, you didn’t pass. The crew chief pretty immediately knows what’s not in compliance — and by how much.

The Rear Toe Problem

The week before Chicago, NASCAR announced they would step up penalties for cars failing the LIS. The intent was to try to make sure that a team didn’t advance in the chase because they won a race with an illegal car. They needed to cheating trigger enough of a penalty that it would dissuade the crew chiefs from trying to push the limits.

Most of the recent LIS violations were in the rear toe because that was one place teams could get a little more skew into the car. Additional skew helps the car turn and provides some side force advantage. Toe is the slanting of the wheels in or out, as shown below. The diagram is shown looking at the car from the top.

How Toe Works

I’ve shown symmetric toe, because usually when you talk about cars, you talk about cars that turn right just as much as they turn left. In oval-track racing, though, you’re only turning left, so you’re going to toe the wheels in opposite directions.

Much like the yaw controversy, where teams started out small and kept at it until the cars were so off center they couldn’t get onto the inspection platforms, NASCAR finally had to make a new rule. They tested a rules package at the All-Star race this year that mandated a neutral rear toe. That worked well enough to keep the rule for the rest of the season.

And that’s where most of the pre-chase violations were being found.

Why Use LIS?

The Laser Inspection System offers many advantages, but also has some issues that haven’t really been addressed. Here are the positives:

Accuracy: The LIS can make measurements to the one-thousandth of an inch in some cases. We can argue about whether you really need that level of accuracy, but even if you don’t the LIS is going to be more precise than a hand-held gauge.

Consistency: The LIS makes all the suspension measurements at one time. Because the suspension of a car is complex, moving the car between measurements could result in something being missed or mis-measured. This is a one-step process that eliminates a lot of variables.

Ease of Use: The only thing teams have to do prior to LIS is place white wheel covers over all four wheels. Those covers give the lasers something to bounce off.

Objectivity: People make mistakes. We rarely do things exactly the same way twice, even when we’re trying to. One person may read an analog gauge slightly differently (or even at a different angle).  Two people may place a gauge slightly differently to make the same measurement. LIS eliminates a lot of those variables.

What About Drawbacks to LIS?

Drivers and teams have been more vocal about their doubts regarding the reliability and accuracy of the LIS. I think this goes to three pretty significant reasons.

Access: NASCAR’s system is proprietary. They developed it for this one specific application in collaboration with a manufacturer of such instruments. You can’t just go out and buy one. Unlike the laser scanners used to measure the car’s body, none of the teams have a similar device in their shops.

None of the teams (as far as I know) is trying to implement a this type of measurement system. There’s not much of point. Suspension geometry isn’t built into a car the way it’s built into the chassis. A chassis is a rigid unit that can’t be changed at the track (and if it were, it would have to go back to the R&D Center to be re-certified)

In contrast, suspension parts are mostly bolt-on. They can be adjusted, or switched in and out, fairly easily. The teams don’t have to check their measurements quickly the way NASCAR must do during the inspection process. It doesn’t make sense for the teams to commission their own Laser Inspection Systems. It would be expensive and pointless.

What they have to do is understand how the measurements they make in their shop relate to the LIS measurements, as pointed out by Doug Duchardt, VP of Development for Hendrick Motorsports:

“What we work on is making sure we understand the correlations of how we measure our cars in the shop to what happens at the race track. If we are measuring this, do they measure the same?”

Once you verify that you are indeed making the same measurements (or that you’re always off by some amount), then you don’t need the LIS because you’ve got a reference (a NASCAR Rosetta Stone of sorts) that translates between your measurements and theirs.

Complexity: The more technology you introduce, the more places there are for something to go wrong. A system requires every part of the system to work correctly in order for the entire thing to work correctly. The more parts, the more places for problems. All you need is one thing to be wrong: a sensor just slightly out of place, a piece that gets a tiny ding no one notices, a screw that isn’t all the way in.

Mobile laser scanning systems are not new; however, this LIS system is ambitious. It’s a large system that must be taken down and put up every week. We have had weeks where there were an abnormally large number of cars failing the inspection. When it’s one or two cars, you assume the cars are the problem. When it’s six or seven, you start to wonder whether there’s a problem with the system.

But even so, we hear many weeks a team voice surprise that they didn’t pass inspection because all their measurements suggest they are perfectly within tolerances. I think this is exacerbated because the teams can’t see how to the LIS works. It’s one thing when the inspector calls you over and shows you a caster gauge and you can see that, yes, you’re out of compliance.

A big fancy machine like this is a black box. They say there’s something wrong and there’s really nothing you can do except go back to your garage bay and try to fix it.

Consistency: This is probably the big one. After his car failed LIS last week, Martin Truex vented a little. Nate Ryan of NBC Sports quoted him as saying:

“We could probably go across that thing (the LIS) four times a day and get four different readings, so it’s a little frustrating, especially from our side of it — when people think that you’ve got a win and your car is illegal.

“One side is fine, and the way it read this time was the left rear was good, and the right rear was off. Usually, if the right rear is off, the left rear is off. So there’s a lot of weird stuff going on there with that machine, and it’s a little bit frustrating, for sure. At the end of the day, we need to make sure that the stuff doesn’t happen, and we’ll just have to be a little bit more conservative coming forward.”

You can hear his frustration: we’re going to have to not push the limits because there’s a chance the machine is going to say we’re off even when we’re not.

There’s two parts to consistency: the system and the car.

Can a Mobile System like the LIS be as Accurate as a Lab-Based System??

Let’s take the LIS first. There’s no way a mobile laser measuring system is ever going to be as accurate as a fixed system. Even if you sunk all kinds of money into it, you’d have variables that cannot be eliminated.

  • You’re dealing with different installation sites each week, which means different considerations when it comes to vibrations, leveling the system, etc.
  • Measurements made outside can be affected by high temperature or humidity
  • Every time the machine is set up is an opportunity for something to change.
  • The system can be perturbed or damaged during transport, loading and unloading

I’m sure the company and NASCAR developed a procedure for installing and calibrating the LIS system to insure that it is as accurate as possible; however, there are always going to be issues related to temperature, vibrations and humidity that cannot be solved. This is going to limit how precise the system can be.

How a Race Car is Like Jello.

Honestly, I don’t think the system’s accuracy is the issue. The big problem (IMO) is that a car is not a rigid object

The individual parts of a car are connected by springs and shocks and other things that give a little.  When they measured heights, you’d see the crew bounce the car a couple times and wait for it to settle. A car that didn’t pass the sticks could be given a few moments to let the shocks and springs relax, and then it would pass.

Let’s consider for a moment measuring Jello using a laser system. You could have the absolutely most accurate laser system in the history of the world; however, if there’s even a little vibration and the Jello moves even a little bit, you’re going to have some uncertainty in the your results.

There is no way around this. It’s not like measuring the chassis, which is rigidly welded together. It’s not like measuring the body, which is similarly rigid.

The unknown is how much uncertainty in measurement there is due to the inherent nature of the car.

How NASCAR Could Raise Confidence in the LIS

I suspect NASCAR and/or the LIS manufacturer has already done this (or a similar) experiment. But let me tell you what I would do to address Truex, Jr.’s comment. And it’s important to do so. He clearly doesn’t think his team was treated fairly and the rules are primarily focused on making sure everyone operates under the same constraints.

It’s an easy experiment, although it would be tedious and annoying if you’re the one who has to do it. If I were running the NASCAR R&D Center, I would get an intern to put a car on the inspection platform, do the measurements, and move it off the platform.

Then I’d ask him (or her) to jostle it around like would happen during the time the car is in the garage: jack one side up, bounce the back and front, roll it around. Just pick one or a few things to do. Then roll it back up on the platform and re-measure.

I’d have him or her repeat this a couple hundred times, jostling the car in random combinations each time. Then I’d ask for a plot of how many times each value popped up. It would look something like this:

bspeed_laserinspection_gaussiandistribution

Most of the measurements would fall in some cluster around the center (the value that is measured the most times). As you move away from the mean, you’ll find a smaller number of measurements.

How do I know this is what we’d get? Because that’s what random processes do. (and if we got a distribution that was different and implied there was something not random, that would be EVEN MORE interesting because it would tell me there was something else going on!) I would then use the distribution to find the standard deviation.

The standard deviation is how far from the mean you have to go before 68% of all the measurements fall between the mean plus and minus the standard deviation. The most intense red box is two standard deviations wide: one standard deviation to the left, one to the right.

If you go two standard deviations, 95% of all the measurements fall within that box. I’ve even shown (in yellow) three standard deviations, which is 98% of all measurements.

What I don’t know are the values on the horizontal axis. In other words, I don’t know whether one standard deviation would be fifty thousandths of an inch or five thousandths of an inch.colepearntweetlis

Cole Pearn, the crew chief for the 78, tweeted that they were off by 0.01″, which is one hundredth (a.k.a. ten thousandths) of an inch. We don’t know if that’s the actual measurement or not, but it at least gives you a bit of an idea of the magnitudes we’re talking about.

So if my experiment went according to plan, I would present all of this information to the drivers and crew chiefs so they would understand the capabilities and limitations of the measurement.

The Rules Change

NASCAR this week did away with P2 and P3 level penalties in the LIS. Thinking about it graphically, it looks something like this:

bspeed_laserinspection_ruleschange

Now: everything from here out is my own speculation. My guess is that cracking down for small violations given the nature of the LIS is a losing proposition. I suspect that the variability of the measurement is such that splitting hairs between P2 and P3 and being ‘close enough’ opens up questions no one wants to hear asked.

I wager that the values at which P4 penalties begin are large enough that

  1. Being under that value doesn’t give you a whole lot of advantage on track
  2. It is large enough that a team can’t blame being out of compliance on the machine accuracy or the accuracy of the measurement process.

Personally, I think the change is good. NASCAR backed itself into a corner, realized it, admitted it, and found a way out. And I like the way out because it does give the teams a little more room in which to operate.

 

A Nit

Stop saying “LIS System”! LIS is an acronym. It stands for Laser Inspection System. It’s like PIN Number is wrong!  Stop!


Also published on Medium.

3 thoughts on “NASCAR’s Laser Inspection Process: Confidence and Accuracy”

  1. Is a random distribution bell curve applicable as what’s being measured isn’t a series of unrelated random events but rather a range of movement within set parameters? If, for instance the starting point for your tests was an in-compliance car that measured in the exact middle of the acceptable range, the bouncing of the car between tests should result in the car moving X amount in one or the other direction from that middle of the starting point. There wouldn’t be the long tail of possible events in each direction – in other word, shouldn’t 100% of the events fall within a set range?

    And this goes to NASCAR’s setting of the acceptable range – by establishing what is effectively a mid-point and an allowance on each side of that mid-point, they’re deeming a car as non-compliant if – for whatever reason – it ‘bounces’ outside the range.

    It would be similar if NASCAR set both a minimum and maximum speed on pit road (limit is 45 MPH, +/- 5 MPH) – you would be non-compliant if you’re at 49.999 or 50.001.

    1. Outstanding question, Steve! I think a curve like this is a valid expectation for this type of measurement. You’re right about the middle representing the ideal in-compliance car and the breadth of the curve tells you how much variation in a measurement you can get as a result of the movement of car parts relative to each other.

      You’re right that the tails wouldn’t go out forever. No matter how hard you bounce the car, all the measurements will fall in a certain range.

      I don’t know how NASCAR sets up their ‘acceptable range’. The laser inspection is a totally different kind of animal compared to the way they’ve previously had to establish rules. A spoiler angle is pretty easy to define. If you measure it a hundred times, you’re going to get very, very similar measurements each time.

      With the laser inspection, the question is how accurate and reproducible it is under the conditions.

      This is one of those things that may raise more questions than it answers (and not in a good way!)

      Thank you so much for reading!

Leave a Reply

%d bloggers like this: