The data said it all: I was too slow down the straights.
The likely culprit? I was also too slow through the turns.
This was the quick interpretation from the VBox Motorsport data-acquisition crew when comparing my laps to one of my competitor’s.
VBox Motorsport had invited us, along with several other journalists and influencers from the motorsports world, to an intensive training session at Circuito Maspalomas. It’s on Gran Canaria, nowhere near here.
[Live thread: Learning how to best use Racelogic’s VBox products]
We’d mostly focus on the company’s HD Lite, a clever, easily deployable data-acquisition system with built-in HD video capture and a full suite of data tools that can be read and interpreted in VBox’s Circuit Tools software.
We plan on transitioning our track testing to this platform, as it should dramatically improve the level of analysis we can share.
Anyway, back to me being slow.
“I don’t feel tardy” was, of course, my first reaction.
But then we looked deeper at the speed traces. Turns out the initial analysis–which was completely valid and certainly the most statistically likely cause of my lap time deficiency–turned out to not be entirely correct.
When looking at the speed trace, we could see that my entry and mid-corner speeds (in red) generally outpaced my competitor’s (in blue). And there wasn’t the telltale stairstep past the minimum corner speed that usually indicates getting on the gas too early.
Instead, we saw smooth negotiation through the corner followed by an acceleration trace that climbed at a steady rate.
My competitor’s curve, though, had a more convex shape. It bulged upward, showing a greater rate of acceleration out of the corner.
It also showed an upshift from second to third much sooner than mine–followed by a shift from third to fourth that I never made because I never got within 500 rpm of needing to make that pull to fourth.
My competitor’s acceleration rate began to trail off by the end of the straight, while mine continued to limb fairly linearly. By the braking point, our top speeds on each straight were almost identical, but they were in fourth gear and I was in third.
So why, then, were they hanging as much as half a second on me down those straights? If our terminal velocities were identical, how was their bulged acceleration curve worth so many tenths? And what was causing it to begin with?
To answer the latter question, we started with the techs who were maintaining the NB-chassis Miatas we were driving.
“Are these cars all identical?” we asked. “One seems like it may be down on acceleration.”
We found out that a couple of cars–mine included–were equipped with a 3.90:1 final drive from an NB Miata originally fitted with a six-speed transmission, while the rest of the cars had the 4.30:1 rear from the five-speed cars.
All the cars were equipped with five-speed transmissions, giving the 4.40:1 cars a real mechanical advantage through the meat of second and third gear. By the time I got well into the powerband in my leggier third gear, it was time to hit the brakes.
Now we have to answer the question of why two cars with identical terminal speeds at the end of a straight–and I even had superior terminal speeds on occasion–were getting outrun by cars with less top speed. The answer is the area under the curve.
Top speed is great, but if you take longer getting there, you ultimately spend more time at an average lower speed for a given distance traveled. Whoever has a higher average speed over a given distance is going to cover that distance faster, regardless of the terminal speeds at the end of that distance.
A more complicated question follows: What’s ideal?
Looking at our acceleration curves, another 500 feet of straight would have likely seen me wringing out my tall third gear and overtaking the rate-of-acceleration contest while my competition was struggling in the middle of fourth gear. But we didn’t have another 500 feet of straightaway.
When comparing two speed graphs in Circuit Tools, we could easily use the delta graph–showing where one car was gaining or losing time compared to another–to see that we were losing lots of time at the beginning and middle of each straight. We were then losing time at a lower rate toward the end of the straight–right on our way to overtaking the delta battle had the straight been longer.
So, what’s ideal for a given straightaway is not just the configuration that produces the greatest maximum velocity but the configuration that produces the greatest area under the acceleration curve.
But I didn’t have access to another mechanical configuration, and frequently you won’t, either. So, how do you use this kind of information to make the most of what you have to work with?
In my case, I concentrated on maximizing any advantage I could gain by my occasionally higher top speeds while mitigating whatever advantage I could find when the situation arose.
This strategy took the form of heavily concentrating on threshold braking, holding that speed as long as I could, and taking as deep a braking point as I could get away with, even if that meant adjusting my approach angle to be able to brake later and harder.
On corner exit, where my competition had their most clear advantage, it meant taking a particularly defensive line to break as much of my opponent’s momentum as possible. Obviously blocking is not allowed, but that doesn’t mean you have to use all of the track on the exit of a corner, either.
Being overtaken by the car with the more favorable gear ratio might be inevitable, but by throwing off an opponent’s exit enough to slightly delay their throttle application could hold off that pass long enough to hold position down a shorter straight.
And the lesson here is that these strategies are easier to employ when you have the benefit of truly knowing where your deficiencies are. That’s where data acquisition comes in.
Without the data, all I really know is that my lap times are similar to everyone else’s yet I’m getting dog-walked down the straights.
With the benefit of the data acquisition, I can at least pinpoint specific areas where I may have had a momentary advantage so I can focus on maximizing the time I spend in those moments and leverage them as much as possible.
The other question this situation brings up: Could we have figured out this issue without leveraging data acquisition?
Yes, of course, but the mystery-solving curve could have been much, much longer and required a lot more trial and error and experimentation. Looking at the objective speed traces massively streamlined the process and pointed us to the right questions that much more quickly.
Also note that we didn’t use any esoteric functions of our data equipment. A simple speed trace was the most valuable tool in our arsenal when it came to solving this mystery.
Comments
BTW, I spotted a couple of minor typos in the article -- the OEM rear end ratio for the NB 5-speed is a 4.3, not a 4.4. Also, at one point the article says that the other car was finishing the straight in 5th, which doesn't seem to be reflected in the data (starting in 2nd, 2 shifts puts it in 4th).
As for the conclusion:
Also note that we didn’t use any esoteric functions of our data equipment. A simple speed trace was the most valuable tool in our arsenal when it came to solving this mystery.
I would argue that an RPM trace wold have solved the mystery sooner, because the gearing ratio discrepancy would have been immediately apparent in the data instead of something you had to infer from shift points. :)
Tom1200
PowerDork
10/6/23 12:00 p.m.
Mmmm what happens when you don't have access to the other car's data???
codrus (Forum Supporter) said:
BTW, I spotted a couple of minor typos in the article -- the OEM rear end ratio for the NB 5-speed is a 4.3, not a 4.4. Also, at one point the article says that the other car was finishing the straight in 5th, which doesn't seem to be reflected in the data (starting in 2nd, 2 shifts puts it in 4th).
As for the conclusion:
Also note that we didn’t use any esoteric functions of our data equipment. A simple speed trace was the most valuable tool in our arsenal when it came to solving this mystery.
I would argue that an RPM trace wold have solved the mystery sooner, because the gearing ratio discrepancy would have been immediately apparent in the data instead of something you had to infer from shift points. :)
Thanks for spotting the typos. I'm a zillion miles away but I'm sure the CFla-based web crew will get th ecorrections made.
As for an RPM trace, I don't completely disagree, but an rpm trace wouldn't necessarily instantly reveal a higher mid corner speed the way a speed trace would. If you just looked at an rpm trace, you'd also see lower rpm on corner exit, and it could be just as easy to work under the assumption that that also meant lower cornering speed, hence less exit speed, hence less straightaway speed.
Ultimately I think the real takeaway message is that people tend to get really intimidated by data, and they can get overwhelmed by the sheer amount of information they car produce. It's a fair and reasonable fear. But the reality is you really don't need many channels to produce a TON of usable info.
JG Pasterjak said:
As for an RPM trace, I don't completely disagree, but an rpm trace wouldn't necessarily instantly reveal a higher mid corner speed the way a speed trace would. If you just looked at an rpm trace, you'd also see lower rpm on corner exit, and it could be just as easy to work under the assumption that that also meant lower cornering speed, hence less exit speed, hence less straightaway speed.
Ultimately I think the real takeaway message is that people tend to get really intimidated by data, and they can get overwhelmed by the sheer amount of information they car produce. It's a fair and reasonable fear. But the reality is you really don't need many channels to produce a TON of usable info.
Yeah, an RPM trace by itself is not useful. RPM + speed trace viewed together would make it very apparent, though.
A speed trace by itself will identify problem areas and will let you infer reasons as to why the area might be a problem. IME, though, to really solve those problems you want the driver input traces. Throttle position and brake pressure will confirm (or reject!) your inferences (in the case presented above, a throttle trace could have definitively ruled the "late to throttle" argument). TPS and brake pressure traces will also give much better feedback when you go back out to try to fix your pedal work. You can make a similar argument for steering angle, but I don't find that I actually use it all that often -- I guess steering is just easier to get right than brakes or throttle! :)
I would say that 95% of the time I'm looking at data I'm using the three primary traces of speed, TPS, and brake pressure, plus the derived trace of delta-time when comparing two different runs.
PT_SHO
New Reader
10/6/23 2:06 p.m.
My '06 car apparently does not have brake pressure or steering angle available in the OBD-2 stream. All I have seen that is useful (using Solostorm) is throttle position and RPM.
I will look into whether I can graph out the speed trace. Right now I just use the course segmenting feature to look into where I was quicker, aka shorter segment times, to pick out which section to focus on. So I don't see an advantage of the VBOX for my situation, though I welcome comments.
A truism generality from drag racing is that cars with similar horsepower will have similar trap speeds, but that other aspects of setup will drop elapsed times. Clearly if JG was only in third and competition was in fifth, there was some massive difference in gear ratios. But if a specific corner was in his /optimum/ powerband for his transmission, his rear end would give him an advantage AT THAT POINT over the other car that was not, and if the transmissions were identical other than rear end ratio then the drop ratio between each gear would be the same. If the drops between gears are too large to stay in the optimum torque range (hard to believe with a 6 speed) then the higher numeric transmission would have slighly tighter drops and that too could end up as an advantage.
Displaying 1-5 of 5 commentsView all comments on the GRM forums
You'll need to log in to post.