Why Data Analysis Isn’t Just For Nerds
If you’ve watched just about any pro level endurance race on TV, you’re familiar with heavily accented announcers talking to us about how cool and awesome pro level teams are. One of the cool and awesome things we hear about is called telemetry, an ongoing live report of the car’s health and performance being sent back to the pits for analysis by teams of engineers in real time. While it may sound silly to some, the information plays a large role not only in making the cars faster, but also in keeping them reliable. At the grassroots level we don’t quite need live telemetry, but recording data for further analysis is one of the best things you can do for yourself as a driver, and for your car. In the last 10 years data systems have become incredibly affordable, while being able to perform functions advanced enough for professional racing teams. Even an entry level system used correctly will not only save your engine one day, but it will surely cut a half second off of your lap times. Costing only as much as a set of tires, isn’t that worth it? Let’s discuss.
I know I’ve made some pretty big claims here, but if you’re willing to put in the time to learn the software and know what to look for, my claims are not only certain, they are conservative. I know that I’ve personally improved tremendously over the last few years by learning how to analyse and improve from the numbers and graphs. There are a ton of systems out there, but I would personally get one that is also a display dash. The best thing about having a digital dash is that you can program most all of them to alarm you when a certain parameter gets outside of your pre-set safe boundaries. This is, in my opinion, the most important feature. No driver is able to watch all of the gauges constantly during a race. Because most engine failures on the racetrack could have been avoided by stopping when the tell-tales throw in the towel, we’re talking about cheap insurance here.
Simplifying the driver’s job is only the start. When we look at the data after a session we can see where the driver has made mistakes, where he has “left time on the table” and by comparing different laps against each other, we can see what the best potential lap could have been, and plot new strategies for the next session. I’m getting excited just thinking about it. On numerous occasions I can remember going through my best laps second by second against each other looking for tenths. It’s like hunting for time. I recommend starting with a simple set of sensors, and then expanding the system as your skills improve. Let’s look at a simple system.
The bare minimum to get any useful information would require the recording of RPM, vehicle speed, throttle position, a lat/long G-sensor, and lap time (beacon or GPS). Some systems can calculate longitudinal G-force from vehicle speed, but that’s backwards. Since we also use the data to make sure the engine isn’t a grenade with its pin pulled (all engines are grenades, don’t be fooled), we also need to monitor oil pressure, oil temp, and water temp at the bare minimum. With this information we can get a good sense of the ability of the driver, how often he is on the limit of the tires, how hard he brakes, how smooth he is, how early and aggressively he is picking up the throttle and much more. On the engine end of things, we can observe and record trends such as water temp vs oil temp, water temp versus ambient temp, and we can also be on the lookout for oil pressure loss during corners, which is a no-no and indicates the oil pickup is no longer drawing oil!
So what do you do with this information? Well, let’s consider viewing speed and throttle position at the same time. Say that on lap four exiting a corner our Stig got on the throttle earlier, but at a slightly slower speed. On his best lap, in that particular corner, Stiggy got on the throttle much later but his mid-corner speed was higher. He felt that corner was awesome and bragged that he was on fire. Well, the data nerds noticed on lap four that the car gained 4 mph in straight speed, and while the best lap was 0.2 seconds faster in the middle of the corner, the driver lost 0.5 seconds by the end of the straight. We’re talking three tenths in just one corner, and I see it all the time. Another example for speed plus throttle position would be to compare braking zones. I’ve observed a difference in 10 meters braking from 140 mph-40 mph to be worth roughly 0.150 seconds. If you’re paying attention, you’ll realize that we’ve already found a potential 0.450 seconds.
Let’s move onto the accelerometer. By looking at the longitudinal Gs versus speed we can see how consistent the driver is in braking. As soon as a driver goes to heel-toe he will naturally change his brake pressure. Almost all drivers, including myself, are guilty for never coming back to the same initial pedal pressure. Adding that pressure back would mean more deceleration, equalling the opportunity to start braking later. Spending less time going slow. Faster lap. The vector of lateral G and long G can be calculated into a channel called “total G” that can indicate how well you balance the car from braking into mid-corner.
And that’s just the tip of the iceberg. Eventually you’ll want more information to help improve car setup, and to get a fuller picture of what the driver is doing. Integrating data with video is a fantastic option, as that data won’t tell you if you’re in a pack of twenty cars or if you’re out there all alone. Data systems are worth the investment if only for the warnings they can display to alert the driver. When you then consider how much money you spend to gain half a second, you’d be silly not to have a data system.
Of course, it won’t do the work for you. You need to learn how to read the data and improve from it! Nerd it up!