THSCC Forums

As some of you know, we put a DL-1 data logger in Carl’s Neon this past event and collected GPS and accelerometer data from each run. I’ve just started playing with the results and while I was fairly skeptical how useful it may be, it turns out there is a ton of good information in there.

The first thing I wanted to check out was the course maps generated from the GPS data. Keep in mind that we started the logger before we got to the start line and we stopped it long after crossing the finish line so the following “course maps” contain extra information at the beginning and end.

This is what was generated by my third run in the morning:



Here is Carl’s third run in the afternoon:



It looks pretty good so far. The maps resemble the actual courses and you can clearly see that we almost made a complete loop by shutting down the logger after pulling back to into grid.

Here are the two runs on top of each other showing that despite the time interval between them, the “GPS drift” has been kept to a relative minimum:


Pretty boring so far, huh? More to come…

Jim

The analysis software is a little clunky especially for “open lap” type events. Essentially, you place 2 markers that represent the beginning and end of your lap and name them specifically “start of run” and “end of run”. You can then place any number of other markers you wish to look at specific parts of the course. Finally, you define sectors which span from marker to marker with “Sector 0” defined to be the entire course. I broke the afternoon course into 7 sectors and it was interesting to see where time was lost and gained throughout the runs.

Here are all of our runs for the afternoon with markers and sectors (I forgot log my first one):


Note that the label for Carl’s 1st run is missing from the image above but there is actually a line representing his run. The software is a little buggy in spots. And clunky.

Clockwise from the top of the course, the blue markers are labeled: “start of run”, “launch”, opening”, “back section”, “slalom 2”, “sweeper”, “loop” and “end of run”. Given that, the sectors are defined as “Sector 1” = “start of run” to “launch”, “Sector 2” = “launch” to “opening”, “Sector 3” = “opening” to “back section”, etc. Remember that “Sector 0” is defined as “start of run” to “end of run” and is therefore the total course time.

With all that setup, the software will now display course and segment times. The following image shows our runs with the fastest time highlighted in each segment. Ignore the time in the first column as that just represents the size of the logged file. In addition, the overall times will be *slower* than the posted results due to the fact that the markers aren’t in the same place as the beams we used on the real course. I set the “start of run” marker to be where the cars just start moving and that is a few meters behind the actual start. The “end of run” marker is probably a hair short of the true finish as I used a spot where all the cars were still accelerating. I’ve found that the start position has the biggest effect on the overall time as would be expected since the car is moving the slowest and that more than accounts for the difference. The good thing is that it is all relative and therefore doesn’t matter when comparing the data. Even better, comparing the calculated times -vs- the actual times, the delta is consistent with an apparent error in the low 100ths of a second. That’s definitely good enough for sliding around in the dirt!

5 Afternoon runs:


Looking at the data, it’s obvious that Carl’s 3rd run (“A_Carl_3”) was the fastest. We both rode along for each other’s runs and even before we hit the finish on that one, I knew Carl was on a flyer. He placed the car perfectly in most spots and drove through the sweeper like a man possessed! Interestingly, he went to the inside of the slalom after the sweeper on that run (and only that run) but still managed to carry more speed than either of us did on any other run!

Starting with sector 1 and 2, the first thing that jumps out at me is that short-shifting might be the ticket in Carl’s car. On run 2, I purposely shifted into 2nd fairly early and it seems to have an advantage which carried into sector 2. I’ll have to play with that more to see if I can validate that data point.

From the first hard braking zone (start of sector 3) and all the way to the sweeper (start of sector 5), Carl pretty much murdered me. He was much more aggressive and it shows in the times as well as the video. He also nailed the last S (sector 7) on his last run in a perfect balance between line and speed. This shows up very well on the track maps as well as the speed graphs which we’ll check out later. I took my revenge in the loop (sector 6) and we’ll check that out as well.

One of the “games” you can play with the data is to try and figure out what would be the ideal lap. While it is easy enough to just add the best segments together, it can be misleading as it may not be physically possible to execute. After looking closely at the intersection between the different data sets, it looks like adding the best segments in this case is fairly accurate. With that said, if we managed to piece it all together for a single run, we would have been in the 1:00.8 range for about a 1.3 second gain. Applying our “real world delta” from before and we wind up with a best “real world” theoretical time of 58.6 seconds. The best time in our class for the afternoon course was turned in by Chris Suich at 59.052 seconds showing how good of a drive Chris had!

Bench racing, anyone?

More to come...

Jim

WOW, I cant wait to use my MaxQData. Keep the geek stuff coming!

_________________
02 Focus SVT
STF 9

So where does the time go? Carl was much faster than me through the back section and while his fastest sector 3 time was on lap 2, I’m going to compare his lap 3 with my lap 3. Without going into detail, the GPS position traces (maps) can be a couple of meters apart between runs but occasionally you get a pair that are *very* close. The current software doesn’t allow you to apply any sort of positional offset to a map yet but they claim they are working on it. In the meantime, it is more of an annoyance than a hindrance and doesn’t affect the results to any real degree. I’m working on my own solution to that as well so things can only get better.

Anyway, if you look at the table at our 3rd laps, you will see he beat me by ~1/2 second in sector 3. Let’s see if we can find the time.

Here is a better view of sector 3:


The red and black circles just past the “opening” marker represents our cars at that point in time. That point actually marks a time in which we were dead-even! The little lines sticking out of the circles represent the horizontal and vertical g-loading and if you look closely at the image, you can just make out what is happening. In the software, you can zoom in to take a closer look as well.

If you notice, with the exception of a small slice right near the “Sector 3” label, our lines through this section are nearly identical. Even through that one section where we diverge, it is only by no more than a meter. Look at the scale in this next image to see how close we really were. This happens throughout the entire run!

Our lines through across the “back section” marker:


How about if I told you that we got on the binders at nearly exactly the same time just after the “opening” marker and we got back on the gas at roughly the same time charging towards the “back section” marker? And technically he ran a longer line around that section by taking a *slightly* wider arc near the beginning of it! OK, we did all that *and* we drove the roughly the same line, how in the heck did he beat me so badly?

Speed through section 3 -vs- time:


Simple! He flat out-drove me through there. One of the features of the software is to compare a sector between 2 or more runs with time relative to the instant each run passed the starting marker for the segment. This is the basis for its “best theoretical lap” calculation as well as its time-slip functionality. It essentially allows you to compare sectors between runs even though they are happening in different relative time-frames.

If you look at the above graph (Carl is in red), you’ll see that we both started bleeding off speed at the same time but then he didn’t bleed of nearly as much and had a much higher average speed from roughly the 2 second mark to the 6 1/2 second mark. How did he do that? Carl wasn’t afraid to pitch it in there a little to rotate the car while I was taking more of a traditional slow-then-turn approach. You can see this in the lines coming out of the circles in the view of sector 3.

Remember I said that the circles were at the point where Carl and I were roughly even? If you look at that image carefully, you can see that I am already pulling lateral g’s (black line sticking out of the circle toward the top of the image) meaning that I am turning in anticipating the inevitable understeer from my unfortunate approach. Meanwhile, Carl’s circle is still slowing in mostly a straight line but he has already set the car up to be flicked into the turn. As a result of his approach, the car is rotated just enough so that he can get back on the gas and blast through the turn leaving me in the dust.

So what happens at the exit of the back section? From the data, it appears that Carl and I get back on the gas at roughly the same point in the course. The really interesting thing is that at that point in the course, we were going nearly identical speeds and our heading was the same. So how did he gain another couple of tenths right there? If you look at the speed graph from roughly 8 seconds on, you see a bunch of dips in my curve as I accelerate. What was I doing? Getting loose, of course! Carl was busy accelerating ahead while I was sawing at the wheel and contemplating my naval or something. The video confirms that Carl was reacting much faster to the car than I was and as a result, he corrected a bunch of fast micro-slides while I corrected relatively few slow macro-slides. Winner? Carl!

Next I get my revenge in the loop…

Jim

Actually, before I get my revenge in the loop, let’s take a quick peek at the 2 sectors before it, 4 and 5. 4 was perhaps the least exciting part of the analysis as the time made through the slight chicane there was largely determined by the speed you carried into it off the back section. Either there really wasn’t enough of a feature there to upset the car much or Carl and I handled it consistently well each run. Of course, I lean towards the latter…

If anybody is reading along who wasn’t there, the sweeper in sector 5 fed straight into an optional 2-cone slalom. OK, it wasn’t much of a slalom but you had to pick which way you wanted to go at the first cone. With the exception of Carl’s 3rd run , we went to the outside of the first cone every time. Carl went to the inside on his third run and the sector times clearly shows that he was fastest through there on that run. So, should we automatically deduce that the inside was the way to go through there? Not so fast!

Close up look at the sweeper, sector 5:


Here again the GPS did a great job of distinguishing between Carl’s red inside line and my black outside line. Notice how the “sweeper” blue marker is defined very close to where our paths cross. This should be very close to the mid-point of the 2 cones that made up the mini-optional slalom. OK, so how did Carl’s inside line gain ~1/10th of a second on my outside line? This time, the answer is a little more subtle than it initially appears.

Speed -vs- time graph for sector 5:


One thing is clear, Carl absolutely kicked my butt coming out of the previous section! We’ve already established that nothing interesting happened in sector 4 so his extra speed is solely based on being smoother coming off the back section. We both got on the gas way back there at the same time but he managed to keep the car going the proper direction longer during that period. Like I said, he flat out-drove me through the entire back section.

Look what happens next, however. At ~1.8 second mark, our speeds are roughly the same. Where did that occur? It happens a good bit before our paths start to diverge. Carl had to bleed off a bunch of speed to get the car pointed towards the inside while I was still building speed I was missing from the previous section. Our speeds cross again as Carl and I hit roughly the same steady state speed which we then hold throughout the rest of our respective lines through the sweeper.

If you look at the circles representing the car in the previous image of sector 5, you’ll see that I’m pulling a lot more lateral g’s than Carl while he is busy accelerating forward. That point directly corresponds to the point selected in the speed graph at the 3.6 second mark. In order to fully illustrate the section from where our speeds first cross to where they cross again (~1.8 seconds to ~3.6 seconds), I created the following image. This was done by simply highlighting the section on the speed graph which shows some of the power of the software. It’s still clunky though!

Highlight of ~1.8 to ~3.6 seconds from the speed graph:


OK, let’s start from the beginning of this section again. We know that Carl came blazing in but if you look at the speed graph, sometime *before* the above highlighted section he had to turn the car to make the inside line. He starts turning around the .9 second mark and spends the next roughly 2 seconds bleeding off enough speed to make the turn. Meanwhile, I’m taking the longer outside line but I don’t have to slow down yet due to the fact that I can carry a touch more speed out there.

Now imagine if Carl had taken the outside line on that run. I believe his speed graph would keep climbing until roughly the 3 second mark at which time he would have had to bleed off very little to make the outside line. By rough calculations, I’m guessing that would have netted about an extra 2 to 3 tenths that neither of us found that day. Does that mean that it is always faster to go ‘round the outside? Nope! But I think on that course during that period of time, the outside was a little quicker.

While it is a little hard to see from the speed graph above, we can also learn something from the tail end of the sector. The graph along with the other data shows that I got off the gas a touch sooner than Carl but we wound up getting on the brakes at roughly the same time. That means I coasted for a bit before applying the brake while Carl jumped right on them. This could be the difference between a left-foot braker’s possible quicker reaction time (Carl) -vs- a right-foot braker (Me) or it could be that I was just coasting. I tend to lean towards the “coasting” theory at the moment as I do remember trying to be extra gentle with my inputs.

No matter what happened, the end result was that Carl was carrying a bit more speed right at the end of that section but as you will soon see, sometimes you have to go slow…

Jim

First, Jim, thanks for sharing this data- it's very cool! And it's kind of you to pick the one place out of the whole day where you didn't kick my ass to talk about...

Hmm, I'm not so sure about that. You've got all the numbers so would know better than me, but since it does look like I was carrying more speed into Sector 3 than you, I should have been making time on you through there. I also agree that we got back on the gas at the same point on the course, but since I was going faster, I got to that spot sooner. If you look at the final "dip" in our velocities before we start accelerating, you see mine comes a couple tenths before yours, then we both accelerate at about the same rate. My point is, you should end my speed v. time plot a few tenths before yours if you want them to represent the same spot on the course- my "end speed" appears higher because I'm probably 20 feet further down the course by then...


You sure did! An excellent display of braking control, especially in your first time in the car. Obviously you killed me there, even discounting my 2nd run where I stalled. And of course you also put the wood to me in Sectors 1, 2, and 7 as well...

_________________
Carl Fisher

Be Cool to the Pizza Dude:
http://www.npr.org/templates/story/stor ... Id=4651531

If you can hide this thing on Vincent, you can sell the results.

_________________
RedLanternRacing, builders of the fastest cars in last place.
Success is not an option.
1981 Scirocco, "Vera"

Jim,
Definately some cool stuff. Very nice analysis.

Can you plot speed vs distance instead speed vs time?
I think that would give better head to head comparison at given points along a sector.

I struggled with the keyhole turn-around and always took it wide and steady. I see your best sector 6 run was a turn early, cut across the muck and turn out. I wonder if I took that track if I could get another .5 sec... hmm... I guess we'll have to wait until November to find out.

Be sure to save the data, since Four Oaks it now a semi-permanent course. It'll be fun to see if we can improve times (given similar track conditions).

_________________
"Ruttin' is racing"

Chris Suich
Apex, NC
AutoX 2012 Nissan Leaf (Quietly changing the world)
AutoX 2003 Mini Cooper S (on sabbatical)
RallyX 1993 Nissan Sentra - "Le Tigre"
FunX 1970 Camaro LT1 ('95)

No way we would have had enough traction (street tires) in the General to take the keyhole "wide and steady".

On my slowest run I blew the braking zone and slid out wide and almost came to a stop! Cutting it in as tight as possible way the ticket in a 1WD car.

_________________
Vincent Keene
'06 Ford Mustang GT (track rat)
'15 Dodge Charger R/T (yeah, it's got a HEMI!)
'07 Ford Fusion SE (205,000 miles and counting)
'98 Chevy Z-24 (retired)
'93 Acura Integra (Team SWB 24HOL Car)

I didn't word that section very well. The speed graph is a little misleading in that the faster person’s trace appears to be compressed since they were in fact gaining time on the other trace. There is a better graph for looking at the data but I was saving it for a later section. Let me see if I can explain what I was talking about a little better. Take a look at the following image.

Point on the back section where we both start gaining speed:


First I should say that that is not the point where we both got back on the gas. I really can’t tell that from the data we collected and what I was really referring to was the point where we both started gaining speed. Believe it or not, at that point, we are both traveling roughly the same speed (less than a 1mph difference) and the car starts accelerating at roughly the same rate.

Here are the 2 points on the speed graph:


If anything, it shows that I was accelerating earlier which was a direct result of my conservative approach to the previous corner. Let’s add a new marker across that point and see what happens to the exit of that section. I also added a new sector "8" which spans between the “back section mid point” and “back section” markers.

New “back section mid point” marker and sector 8:


Speed -vs- time for sector 8:


Look at that! You passed that point going a *hair* slower than I was but managed to make it up by the time we were exiting the back section! My initial better acceleration was due to the fact that the car was pointed more or less straight ahead at that point (conservative entry!) while you were still managing your entry slide. What were our sector times for the new sector 8? I did a 5.02 and you managed a 5.07. So my conservative entry netted me a whopping 0.05 seconds while you still beat me by 0.52 seconds for the entire back section! The vast majority of that time was simply in the corner entry approach which I’ve highlighted below.

Carl kicks my butt in this section to the tune of .5 seconds:


OK, so from the “back section mid point” marker to the “back section” marker, we were a wash, right? Not really. While the time was roughly the same, your exit speed was much greater and it was carried all the way to the sweeper. That was specifically where I was misleading in the paragraph you quoted as I was looking ahead to the advantage you gained off the corner even though it hadn’t shown up on the clock yet. Let’s take a peek at our speeds just entering sector 4.

Speed -vs- time just after crossing the “back section” marker:


It’s pretty clear that you did a much better job maintaining your momentum out of the back section. If you look closely at the video, you’ll see that you were much quicker on the wheel catching any mini-slides and generally keeping the car pointed in the correct direction. It was subtle but it adds up and definitely shows up on the stopwatch.

At this point I should say that I’ve now had this product for exactly a week so I’m still getting to know it. If there is anything that looks off or you think I’m interpreting something incorrectly, please let me know. I’m here to learn as much as anything else.

With that said, any interpretations are subject to change without notice at any time!

Jim

Vincent is spot on. I'll show that in my next section I'll post later this afternoon when I get a chance.

I can plot speed -vs- distance and have in a few instances (but not obviously posted). Each graph has its own characteristics and limitations and for what I was initially trying to say, the s -vs d graph didn't illustrate my point very well. Another issue that is "distance" is calculated as the software has to interpolate across a discrete set of data points. Little errors tend to compound on themselves during the calculation and it winds up being a less accurate that what I would like. If I get some extra time, I'll try to explore that further for you Chris.

I definitely plan on keeping everything and hopefully we can all learn something from it. My main goal is to try and solidify the general feelings we have about going fast out in the dirt. I think most of us have a pretty good idea what is fast in most cases but I think we will find a few surprises here and there. I'll summarize everything when I'm done and it should make for a decent cliff-notes version to this madness at the very least.

I was starting to think I was talking to myself so it's good to hear some feedback. I'll share whatever I have as long as people are interested.

Jim

Well I'm sure interested! Over the past year or so I've either picked up some bad habits or really not kept up with the learning curve, 'cuz I've gone from being fairly often in the Top 5 overall to only being in the Top 5 in class on a good day!

_________________
Carl Fisher

Be Cool to the Pizza Dude:
http://www.npr.org/templates/story/stor ... Id=4651531

I'm sure Chris will admit that the result of that run was a little skewed as Kevin had "fixed" the course as he often does for the poor little AWD boys.

Not saying that Chris didn't have a good run, but the back section which gave most of us the most trouble was re-graded before his run.

_________________
Vincent Keene
'06 Ford Mustang GT (track rat)
'15 Dodge Charger R/T (yeah, it's got a HEMI!)
'07 Ford Fusion SE (205,000 miles and counting)
'98 Chevy Z-24 (retired)
'93 Acura Integra (Team SWB 24HOL Car)

Hrrrmmmm Good catch, I didn't realize that. Regardless, his 2nd run was still fast and if he went to the outside of that section to achieve that, it is all the more impressive. We'll check out that section in detail and maybe we can estimate how much either approach costs or gains.

Anybody else care to share what they did there and how they think it worked for them?

Jim

Jim, this thread is missnamed. You should change the word "quick" to the word "exhaustive"