There is never enough time on race day, especially when there are multiple cars in the garage
What to look at when you just have 20 mins
By Trav Mays
Having come in from an outing, you have de-botted and de-suited, taken the covers off your car and have had a look over it for any obvious issues. Whilst it is still cooling down and you have rested and chatted, you look up at the clock and think to yourself you have about 20 mins before you need to start prepping the car again, you wanna look at your data, what should you look at first? Or maybe you are someone looking after a number of cars in the same race and you can only spare about 20 mins with each, what then, what should you be looking at?
Vehicle Vitals
Assuming there is nothing wrong with the car, excessive oversteer or understeer, braking, etc, the first thing to do is download your data and have a look at all the car's vitals across the whole session. The main ones you should be looking at are
Below is a screenshot of a typical engine vital signs page, using the first 3 laps of the Motec i2’s sample Calder Park data. I have included just the first 3 laps because of a lack of room, the screenshot would be too small if I had included the whole session. As you can see I have set it up so RPM is along the top, and then I have the temperatures second, the pressures 3rd, and the volts 4th. Laying it out grouped like this allows for a less busy screen, less Y-axis points, it also makes it easier to see if something has gone wrong in one of them, a drop in coolant temperature would be quite noticeable when compared the rest despite them having different temperatures.
- Engine RPM,
- Engine oil and water temperatures,
- Transmission and differential temperatures if you can,
- Engine oil and fuel pressures, and
- Battery voltage
Below is a screenshot of a typical engine vital signs page, using the first 3 laps of the Motec i2’s sample Calder Park data. I have included just the first 3 laps because of a lack of room, the screenshot would be too small if I had included the whole session. As you can see I have set it up so RPM is along the top, and then I have the temperatures second, the pressures 3rd, and the volts 4th. Laying it out grouped like this allows for a less busy screen, less Y-axis points, it also makes it easier to see if something has gone wrong in one of them, a drop in coolant temperature would be quite noticeable when compared the rest despite them having different temperatures.
First check is the vitals of the car - data: Motec i2's sample Calder Park
Another way to quickly view this data, especially if you don’t have alarm channels set up, is to open up a channel report. Below I have the channel report for the above session, the first thing you’ll notice is that I have laid it out exactly as above, the different channels are in the exact same order, this helps with going back and forth between the two pages to check things out. The other thing you have probably noticed is the colours, the pink or coral colour represents the highest value across the row whilst the blue highlights the lowest. For example, if we look at the min row for the RPM channel, we can see that the highest min was in lap 3 (2346 rpm), whilst the lowest value, surprisingly, was on the in lap (0 rpm).
Another way to view the data is with a channel report
I have put for each channel the min, max, range, average, and standard deviation, I believe these 5 give us an excellent snapshot as to what’s going on, especially the standard deviation. In the image above, you can see that the engine temp standard deviation for lap 3 is highlighted with a white border (I did this by clicking on it), it is a lot higher than any of the other fast laps, so might be something we should look into. If we follow the column up, the high standard deviation is backed up by the range (difference between min and max) value being considerably higher than the previous two laps, 6.4℃ as opposed to 2.6℃ and 2.0℃. If we continue to the top, we can see that the lap time posted was about 1.5 seconds slower, so that might explain it, but lap 1 was a similar time and didn’t vary this much. To investigate further lets add a couple more rows, the temperature at the start and at the end of each lap.
Channel report with Start and End values
We can see that lap 3 had the highest start lap temp of the racing laps, 94.4℃ and if we look at the track map below, we can see that at Calder Park, after the start/finish line we have a long straight, this would help to explain why lap 3, despite its relatively slow lap time it achieved the highest engine temp. If we then have a look at the temp at the end of the lap, the lowest of any except the in lap and we have solved the mystery, which would have taken about 20 seconds. We could have also seen this in about the same time on the engine vitals page, but this method is better for people who aren’t visual and are more numbers based.
Motec i2's track map of Calder Park, Melbourne
Driving
We have checked over the car’s vitals, solved any mini mysteries we want to look at, and have about 15 mins left, what should we look at first? If you’re using Motec i2, I would recommend opening up the speed trace across all laps, have a look at anything that stands out, minimums and maximums, slopes, inconsistencies etc, zoom in on anything that doesn’t look right and look further into it.
If you are using Aim’s Race Studio, I would open up the split report and have a quick look at the standard deviation for each section. What we are looking for is high values, the higher the value the more inconsistency. You’ll just need to be careful doing it this way as standard deviations can be thrown off by one large value, for example if you had a spin or lock up in a sector, the standard deviation might be quite high but your times, except for that one time, could be quite consistent. Having a look at the min, max, and average values will help with this. Once you’ve had a quick look at that and seen the areas that could see the most improvement, I would open up the best, worst and an average lap to quickly compare what you did differently. The rest of the analysis process will go along the same lines as the Motec analysis described below.
If you are using Aim’s Race Studio, I would open up the split report and have a quick look at the standard deviation for each section. What we are looking for is high values, the higher the value the more inconsistency. You’ll just need to be careful doing it this way as standard deviations can be thrown off by one large value, for example if you had a spin or lock up in a sector, the standard deviation might be quite high but your times, except for that one time, could be quite consistent. Having a look at the min, max, and average values will help with this. Once you’ve had a quick look at that and seen the areas that could see the most improvement, I would open up the best, worst and an average lap to quickly compare what you did differently. The rest of the analysis process will go along the same lines as the Motec analysis described below.
Aim Split report using Aim Webinar data, link to the video and data here [1]
Using the split report allows for a quick but powerful look over your data, especially when you have a large number of laps.
Example
As stated before, the first step is to open up the speed trace for all laps and check for inconsistencies. Below I have put a screenshot of laps 2, 3, and 4. Laps 2 and 4 were fairly close in time, so let's focus on them. From a quick look, the top speeds and minimums are fairly consistent, the slope of the speed trace under braking and accelerating is fairly similar, so a pretty consistent lap, we do however have one discrepancy highlighted. Because this is such a short outing, I’m pretty sure the driver would know what happened there, but for us, we have no idea, so let’s have a closer look.
First step in analysing the session is looking at the speed trace
The next step is to overlay the laps, open the delta, and zoom in on that area. The section of track we are looking at here is quite large and challenging, it's the S bends and the hairpin before the front straight, see below. The driver is clearly finding it hard to link this together.
Calder park race track map with the section analysed highlighted with orange lines
If we have a look below, we can see a close up of the section in question, lap 2 (the fastest overall lap) is in blue, lap 4 (the fastest section lap) is in green. We can see that they have approached the first corner a bit more cautiously in lap 4 than in lap 2, braking more gradually and 40ms earlier. After taking the first part of the S bend at the same speed, the second part is where lap 4 starts to get its lead back. Lap 4 is a lot smoother through this section, allowing them to take the next corner at 14km/hr faster. Lap 2 continues to be more aggressive on the throttle and brake through the next section, it manages to get the speed up 6 km/hr as they enter the hairpin, only to have to jump on the brakes harder again and lose all of what they had gained, resulting in them going through the hairpin 6km/hr slower and finishing the sector 0.25 seconds slower.
Clearly something has happened here, to find out we need more info, lucky for us we are given everything we need for a quick analysis by Motec. If we have a look at the steering angle channel the first thing that stands out is two massive corrections, the driver needed to catch the car as it oversteered twice in this section. The second thing is how smooth the green line is compared to the blue, as you know, smoothness equals quickness.
If we go one step further and have a look at the braking going into the corner, we can get a bit of a clearer picture. Below I have added the front brake line pressure bias and the front brake line pressures to the image above, some interesting things going on here. As you can see, we have a wild swing in brake bias, going from 70% front biassed to just 38% (62% rear brake biassed) for lap 2 (an insane level of change, this car must have been a handful), whereas in the same section on lap 4, the brake bias only shifted 10%, maxing at 55% and dropping to just 45%.
The extreme level of rear wheel bias occurred as the driver was coming off the brakes and turning into the corner (trail braking), it peaked just before max steering angle, causing the rear to start to step out with the driver catching it. Clearly something is going on in the braking system. So let's have a quick look.
The first graph we will look at is the rear vs front brake line pressure graph, here we are looking for a tight distribution, clearly this doesn't have one (if you’d like to learn more about this or see what a good distribution is supposed to look like, head over to my post on this type of analysis here). The points are coloured for each lap, dark blue is our fastest lap, lap 2, it also happens to have the wildest distribution. The craziest point is at the bottom right, here we can see that the brakes have been applied to a point were we see 2687 kPa (26.87 bar) on the front but only 787 kPa (7.87 bar) on the rear, this is just before the first corner, after that massive straight!
The first graph we will look at is the rear vs front brake line pressure graph, here we are looking for a tight distribution, clearly this doesn't have one (if you’d like to learn more about this or see what a good distribution is supposed to look like, head over to my post on this type of analysis here). The points are coloured for each lap, dark blue is our fastest lap, lap 2, it also happens to have the wildest distribution. The craziest point is at the bottom right, here we can see that the brakes have been applied to a point were we see 2687 kPa (26.87 bar) on the front but only 787 kPa (7.87 bar) on the rear, this is just before the first corner, after that massive straight!
The next grah we are going to look at is the front brake bias vs the front brake line pressure, again we have the points coloured for the individual laps. Unfortunately the brake bias is a recorded value and it wasn’t recorded at a very high frequency, so we don’t have as many samples to draw a conclusion from, just 324 samples as opposed to the 1274 samples in the graph above.
Whilst we don’t have as many samples, we do get a pretty clear picture that something isn’t right. It has been recommended by Claude Rouelle, from Optimum G, that we should be aiming for a variance in brake bias of about 5% at 1380 kPa (13.8 bar, 200 psi). We however have a variance of about 28% and if we include the outliers our brake bias variance blows out to 44%. We never actually reach a variance of 5%, the best we get is about 8%, clearly something is going wrong here. The driver did an amazing job given the circumstances, the inconsistent braking would explain why they were a lot less aggressive after lap 2 and braked 40m earlier for the S bend.
Whilst we don’t have as many samples, we do get a pretty clear picture that something isn’t right. It has been recommended by Claude Rouelle, from Optimum G, that we should be aiming for a variance in brake bias of about 5% at 1380 kPa (13.8 bar, 200 psi). We however have a variance of about 28% and if we include the outliers our brake bias variance blows out to 44%. We never actually reach a variance of 5%, the best we get is about 8%, clearly something is going wrong here. The driver did an amazing job given the circumstances, the inconsistent braking would explain why they were a lot less aggressive after lap 2 and braked 40m earlier for the S bend.
As you can see, this quick 20 minute analysis gave the driver and or team something to look into, clearly there is something wrong with the braking system, figuring out what and fixing it will help to not only bring down lap times, but will most importantly, reduce the chances of an accident. If you’d like to learn more about what potentially is going on in this brake system, check out our other post which is dedicated to this type of analysis here.
This brings us to an end for today, I hope you enjoyed it, I had fun writing it. If you enjoyed this post and would like to read more, we have some other posts that might tickle your fancy, otherwise sign up so you don’t miss any future ones. As always, if you have any comments, suggestions, or noticed I made a mistake, write it down in the comments section or get in contact via email or using the contact page, I’d love to hear it. Thanks for reading and I hope you have a great day.
By Trav Mays
Previous Post Next Post
References
No comments:
Post a Comment