THSCC Forums

I actually didn't get any meaningful amount of grading done before the tractor died. So there was no advantage.

Yes, yes, I will admit that the keyhole was much better after Kevin regraded. I still went wide on that run as well.
You can blame Whitney and Herring for allowing Charlie M and me the opportunity to get re-runs.

_________________
"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)

Hope you get a chance soon on that, Jim!

I've finally gotten a bit of time to play with the video and have a split-screen view of our 3rd afternoon runs almost ready. I think I may put the whole thing in slow motion because otherwise it runs by really fast. It would be nice if I could figure out how to annotate various points on the run, tele-strator style, but it might take me too long to figure out how to do all that...

_________________
Carl Fisher

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

OK I got the split-screen video done and up. I ended up leaving it full-speed, but adding time differentials at various places (ie. cones) which I arrived at by counting frames between when Jim got to a point and when I did, or vice versa, at 30fps. The numbers should be considered pretty rough, as differing lines meant cones passed out of view in different places, making the comparisons inexact.

Interesting to note, though that the "lead" changed twice during the run, first he had a 1 tenth advantage by the time I crossed the starting lights, then I passed Jim under braking into the "back" section, then despite this being my only "good" execution of the keyhole turn, he still took almost a second back from me there.

http://video.google.com/videoplay?docid ... 2335844112

_________________
Carl Fisher

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

Sorry. I got sidetracked trying to hack the raw file format for the runs. I'm trying to see if I can pre-process the data and perhaps make the maps a little more accurate. So far I've been able to tweak the positional data after the fact but I there is no way to load it back into the analysis program at the moment. I'm hoping the manufacturer can help but this may be beyond the sort of tech support they had in mind.

We'll take a look at the loop and the finish as soon as I get back this afternoon.



That's very cool!

Jim

Any possibility of comparing my WRX and 2.5RS sometime with this setup?

Sure, anyone could do it. All you need is an in-car camera setup for each car (presumably moving the equipment from 1 car to the other), taking a couple vids of each, then loading the footage into a video editor, lining up the starting points, then doing the split-screen.

If you mean does Jim want to lend you his in-car cam so you can do that, you'll have to ask him.
If you mean do I want to do all the messing with the video for you, then the answer is no.

_________________
Carl Fisher

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

I have a GPS datalogger that I would be more than happy to let you see the difference with. As long as I got a fun run or two in the Subie of my choice

_________________
02 Focus SVT
STF 9

Uh... I wasn't talking to you.

What's involved in the GPS setup, and how much does it cost? If it's not too much, I could just get one myself and do lots of experimentation with both cars using a variety of course setups.

If it's too much, I guess I could just trade fun runs for data.

OK, the last time we saw our heroes, they were hurtling towards certain death into the gnarly underbrush known as “the loop”. However, now that a video confirming their fates has been leaked, we’ll just cut to the chase and see how they managed to escape.

Sector 6, or what I’m calling “the loop”, was an exercise in sheer patience. The speeds through the slowest part of this section were in the 12-14mph range (that was our *good* runs!!!) If anything, this section drove home the point that if you’re going to make a mistake, don’t do it when you are going really slow!

A closer look at “the loop”:


Once again, you can see how lucky we got with the GPS maps for these 2 runs. The lines are a really good illustration of what happened out in the real world with one exception. If you’ve already studied the video in this area of the course, you should know what I’m talking about.

If you go back and look at the sector table, you’ll see that I gained ~.7 seconds on Carl in this section. We already know from the previous section that he is going faster when we enter this one and he will probably out-brake me, so is that little bit of extra distance he traveled at the bottom of the sector image that much of a killer? As usual with these things, yes and no!

Look back at the little circles in the sector image. At those respective points, Carl’s car is still braking hard trying to bleed off speed while I’ve just released the brakes and have initiated my turn-in. What is the significance of these points? Let’s look at the speed graph one last time.

Speed -vs- time graph for sector 6:


As before, the marks on the speed graph correspond to the car’s positions on the image of the loop. Notice that Carl is slightly ahead of me in the sector image at the same point in time which makes sense considering he was going faster as we crossed the “sweeper” marker. If you follow the speed graph forward from that point, you can see that Carl has to bleed off a lot more speed to finally be able to turn.

OK, what else can we do? One of the software’s internal calculated variables is a notion of time slip. It can be displayed either as a percentage or cumulatively and while it can be very useful, it has some serious limitations in its current implementation. Luckily, it works pretty well for this example so let’s take a look.

Carl’s speed, time slip and time slip rate for his 3rd run in sector 6:


The 2nd graph, time slip, shows the amount in seconds that Carl gained or lost up to that point compared to my 3rd run. Pay no attention to the goofy y-axis as that will become clear when we look at it closer. The 3rd graph, time slip rate, shows how much faster or slower Carl was traveling at that point in the course compared to my 3rd run as a percentage. In other words, if I was traveling at 16mph while Carl was traveling 12.5mph, the difference would be ~22%. The horizontal line splitting the time split rate graph is 0% so to put it in its simplest terms, if the graph is below that line, Carl is gaining time and if the graph is above it, Carl is losing time.

The red line bisecting all three graphs represents the same point in time as in the previous graphs. Putting them all together, that point is time represents the time at which Carl started losing time to me in that section. The first thing that surprised me is how far back into the braking zone it is!

Time slip graph for Carl’s 3rd run in sector 6:


There are a couple of things to note when looking at the time slip graph. The first thing is that the y-axis is a little goofy. It is in seconds and notice that it starts with Carl’s advantage from the previous part of the course. The first value is ~0.7 seconds which is Carl’s lead on me as we cross the “sweeper” marker. He then continues to gain another ~.2 seconds over the next 2 1/2 seconds at which point he starts to lose time. I’ve highlighted the time slip graph in that section so let’s look at what that represents on the course map.

Carl gains time in sector 6 over this region:


OK, so he got me for ~.2 seconds and everything is rosy, right? Let’s highlight the next 5 seconds or so on the time slip graph.


What happened here? All of a sudden his ~.2 gain turned into a ~.7 loss! That’s nearly a second lost over that one part of the course. Which part is that?


I have to say that the result surprised me a bit. As I said before, I wasn’t expecting Carl to start losing time quite so far back into the braking zone and I definitely thought my advantage would extend a little further. The fact is that our speeds at the point where our paths converge again are nearly identical so it makes perfect sense that any advantage I had would be gone.

The exit of this section was fairly routine although there is one thing I wanted to point out. Carl started accelerating slightly before me exiting the “knee” of the loop which corresponds to the circles in the image below.

Carl “steps on it”:


Corresponding point on the time slip graph:


Carl started gaining time on me again at that point but suddenly he starts losing time. What happened? I believe Carl took more of an inside exit which, while setting himself up much better for the next section, cost a little on the exit in the form wheel spin. How much time did that cost? Roughly ~0.05 seconds which is getting dangerously close to the noise floor of the system. Do I think it was worth it? Heck yes! It looks like he more than gained it back in the next section and is a good example of a little extra distance (and time) helping more than it hurts.

Getting back to the loop, does this prove that the inside line was the fastest? Nope, not even close! All it proves is that on that run Carl went a little too deep into that corner. If Carl had braked earlier, could he have reaped the benefits of greater speed off the sweeper and not lost any time in the loop? Maybe. Did his inside line through the sweeper prevent getting stopped and turned in time for the loop? Maybe.

These questions ignored and many more raised in the next installment…

Jim

The Big Finish!

This final section, sector 7, consisted of what was essentially another 2-cone slalom where carrying too much entry speed was a real killer. It was extremely important to “get behind” the first cone and it was painfully obvious on the clock if you didn’t. Take a look at Carl’s lines from his 3 afternoon runs.

Carl’s lines through sector 7:


His 3rd run, the red line, is clearly a better approach and if you go look at the sector times for those runs, it shows he was at least 1.5 seconds faster through that sector. 1.5 seconds!!! He must have been flying in there compared to his previous runs, right?

Carl’s speed as he crosses the “loop” marker:


Wrong! The image above is a graph of Carl’s afternoon runs as he approaches the “loop” marker. The circles represent his speed at that marker. Carl actually got a better run out of the loop on his first 2 runs and yet he still went faster through section 7 on his 3rd! How did he do that? He simply got back on the gas earlier.

Carl “steps on it” 3 times:


The circles represent roughly where Carl started accelerating out of the loop. Look at how much earlier he gets back on the gas on his first 2 runs! The bad news is that it looks to be mostly dictated by his line through the loop and, as we’ve seen before, that probably isn’t ideal. My feeling is that there is a compromise in the approaches in there somewhere but we are really starting to split hairs at that point.

This is another interesting detail in looking at Carl’s speed graphs from sector 6. This is also a perfect opportunity to use the speed -vs- distance graph in place of the speed -vs- time graph. Why?

Carl’s speed -vs- time for his 3 afternoon runs:


What happened? In braking in the slick stuff entering the loop, Carl accidentally stalled the car by locking the brakes too long. When he released the brakes, there wasn’t enough traction and speed left to turn the engine over and we came to a surprising halt on course. The same thing happened to me on my 3rd in the morning trying to navigate the super-slick 1st turn but I got lucky that the engine (barely!) kicked back to life as I popped the clutch. As a result of Carl’s stall, the speed graph is a little hard to interpret so let’s take a look at speed -vs- distance.

Carl’s speed -vs- distance for his 3 afternoon runs:


Just for Chris! One thing to note is that the “distance” is calculated from the discrete sampled data points so it is subject to any error inherent in the samples. Another thing to note is that the “distances” aren’t the same for each run since there are variations in the actual line driven. Just like you can spend more time in one sector from run to run, you may drive a longer or shorter distance in a sector from run to run. As with all the tools, they need to be used and interpreted in the proper context.

Anyway, the interesting thing to me is that his choice of lines out of the loop seemed to cause a drop in acceleration on run 3. I speculated in the analysis of the loop that his inside line on run 3 caused a little extra wheel spin which slowed his acceleration. If you look at the speed -vs- distance graph starting at the 0.1km mark, you’ll see that Carl’s rate of acceleration for his 3rd run is initially as good as the previous 2. Right at the 0.12km mark, you’ll see the rate of acceleration drops off compared to the other 2 graphs. Where is that on the track map?

Carl’s 3rd run starts to diverge from his previous 2:

Right there! Surprise, it’s roughly at the point where his lines start to diverge and his 3rd run really starts heading towards the inside. Coincidence? I think not! What does it mean? In this case, the penalty of reduced acceleration is more than offset by the vastly superior line entering the sector 7.

OK, that was interesting but the question I really want to answer is: “How did Carl punk me for ~.2 seconds through sector 7?”

That answer and other miscellaneous ramblings in The Big Finish, part deux.

Jim

Jim - Want to co-drive thr red wagon at the next event?

_________________
Mike Whitney
whit32@gmail.com, 919-454-5445
V10, V8, V8t, I6, I6, V6, F4t, I4, I4, I4, I4, I2, 1, 1

The Big Finish, part un.cinq

Before we look at how Carl took me to the woodshed in sector 7, let’s take another look at the previous image. Remember the g-vectors coming from the little circles? What do they tell us about those points?

The circle representing Carl’s 2nd run, the black circle, is mostly accelerating forward as shown by the (hard to see) g-vector pointing backwards along its path. The green circle has a similar rearward g-vector but it also has a lateral component showing that Carl is already turning back to the left to set up for the finish slalom. The red circle has a huge lateral g-vector that not only emphasizes how hard Carl is turning back to the left but also how early.

Wait. Turning back to the left? Yep! At first I thought the wheel spin would show up when Carl first got back on the gas (and was turning to the right) but it turns out it doesn’t show up until he needs to turn the car back to the left to stay on line. At that point, the car is fighting for near-maximum lateral traction and therefore can’t accelerate as well as the other lines. While technically that can be counted as wheel spin, I think the real loss of acceleration was more of a matter of flirting with the edges of the friction circle.

This was all dictated by his line out of the loop and really underscores the importance of not only putting the car in the right place but making sure it is headed in the right direction. In this case, the “right direction” caused a loss in acceleration but it paid dividends in the end.

Jim

Thanks for the analysis, Jim- that's really cool, and useful for me (and hopefuly for others as well)! But I still have a few questions:

1. When I made the split-screen video, it appeared that I had completely made up the second that I lost to you in the hairpin by the time we got to worker station 5. How did that happen?
   
2. If it was a "go slow here to go fast there" kind of thing (and it might not have been), can we learn anything about what was the best way through "the loop" section as a whole?
   
3. Was there anything to be learned by the differences between your line/speed through the final "slalom" and mine?

Again, thanks Jim for the time you've put into this!

_________________
Carl Fisher

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

You're welcome! It's been a good exercise for me learning the software. I've been slowed down by my obsessive-compulsive nature trying to understand and possibly tweak the GPS data but it should all come together in the next couple of months. I think I've found a way to get the positional data accurate to a few centimeters but it will require some serious testing.

As far as your questions:

1. I think there is an error in the video and it is confusing the issue. If you look at the sector times for our 3rd runs, you gained time on me all the way to the loop. I think you used the data from my 2nd run (where I was fastest in sector 1 + 2) to sync the start and then switched to my 3rd run somewhere along the line. I'm not sure how you sync'd things up but if there was some time line funkiness in there, that would explain it.

For our 3rd runs, you had ~.8 second lead by the time we entered the loop gaining time the entire way. Well, technically, I was 4 hundredths faster than you in sector 1 but then you got me by 4 hundredths in sector 2. I'd call that a wash. You lost ~.7 seconds in the loop but then gained ~.2 in the final S. That would give you a final advantage of ~.3 seconds which was roughly the delta between our runs.

2. I was planning on doing a summary at the end and I should touch on that topic at that point. Unfortunately we don't have any real data on the "high and tidy" line in the loop but I'll try to make something up and wave my hands around for added distraction.

3. There isn't a ton to learn there but that is the subject of "The Big Finish, part deux". I should have some time in the next few hours to finish that up.

Jim