Geoffrey

I spent the last two days at Lime Rock Park in CT. Friday I didn't drive because it rained all day and I don't have rains for the car, but yesterday it was Ok to drive, although not the best conditions. Yesterday I finally felt comfortable with my car. I fixed the suspension issue I had at Mosport, raised the rear and lowered the front of the car to add more rake and realigned the car. It really felt good and the twichiness I felt at Mosport is gone. I believe air was getting under the car, hence the rake adjustment. If I had more time, I would have made some suspension changes at the track. I would have softened the compression and rebound due to the bumpy track surface and probably tightened the rear sway bar one hole. I also didn't mount my new slicks and continued to run on the original used ones and they are heat cycled out so they don't stick nearly as well as a new set. The racing ABS is simply fantastic and I really noticed it on this bumpy track, especially in the two hard braking zones at Big Bend (turn 1) and the Uphill (turn 4).

I also finally put out the MoTeC beacon so I could collect some lap data. The MoTeC dash has 8mb of logging memory and a very sophisticated software package to help with analysis. I let my friend Bob Scotto, who has a 964 cup car, drive the car and am able to compare data between his driving and mine. He is a much better driver, able to do a low 58 at LRP in his car. He did a 100.2 in my car (which takes some getting use to) and I did my personal best at 59.5. However, when you compare the data, you can see that even with his slower lap time, he is much faster in the back section of the track. You can begin to do analysis as to why he is faster in the back section of the track and that can help identify areas to work on. Every graph is configurable for data points and many other items too lengthy to talk about here. The system uses a lap beacon, wheel speed, and g-force sensors to do all of its computations which makes it more accurate than a GPS based system.

Here are the graphs:
1) Track Report - this graph can show one lap, or a reference lap with one or many laps. In this case, I laid out Bob's best lap and my best lap which graphically shows throttle position and identifies gear, max section speed. The inside lap is my lap set as the reference lap, and the outside is Bob's best lap. You can see that for turn 4 and turn 5 he is on the gas earlier and longer which is giving him a faster corner exit speed.

2) Section Report - this graph lays out both laps and shows data for each section. You can see his minimum corner speed for turn 5 through turn 6 is much higher than my laps. This is consistent with him getting on the gas earlier and longer as well as braking less.

3) Section Times - this graph shows the section times for each lap. The blue cells are the fastest section times for all of the laps. The green shows a section that is within a certain percentage of the fastest section time. The eclectic lap is the fastest theoretical lap based on the fastest section times. Shown here includes 6 sections from lap 31 (bob's lap) and 6 sections from my laps.

4) Braking Data - These are just channel vs time plots that show the two laps overlaid and we can begin looking at detail data for the back section of the track. You can see he presses the throttle earlier and the corner speed increases. You can also see he is braking later.

5) Cornering Data - again, various channel vs time plots that show corner data. The car is generating almost 2g of cornering. I have not yet connected my steering position sensor, so there is no data for steering angle and computed oversteer.

enjoy.

DrJupeman

Geoffrey, that is awesome data. Beating Scotto must be fun, too! (I assume by 100.2 you mean 1:00.2, not 100 seconds... )

Geoffrey

"Beating Scotto must be fun, too!"

Well, I wouldn't exactly say that. I could not keep up with him when he was in his own car. My car takes some getting use to, especially when it needs to be revved to 7800 rather than 6700 of a Cup Car.

Geoffrey

Here is some additional data. On lap 46 I spun in turn 3 (the left hander) Here is the data showing wheel speeds and associated g-force. Unfortunately, I don't have my streeing potentiometer installed yet, so I can't see how the steering wheel is turned. You can see that as soon as the car reached 2.25g it began to spin and I was not fast enough to catch it.

The second graph is of the brake zone of turn 1 (big bend) which is bumpy and causes wheel lockup when threshold braking. My car has racing antilock brakes and you can see the FR lockup and the ABS kick in to release just the FR wheel. You can see the same for RL

The third graph is at the crest of the uphill which shows wheel spin as the car gets light and you are full throttle. I did not have the traction control active as I have yet to find that the car really needs it on a dry surface. I need to spend some time with it in a controlled environment to adjust it. The traction control is based on throttle position vs ground speed with a trim for lateral g-force and a dash mounted trim **** to adjust the amount of wheel slip permissible before the ECU activates the traction control function.

NineMeister

I can't comment on the track as I don't know it, but from looking at similar datalogs from my car I usually see longitudinal G forces (braking) of up to 1.3g on road legal race tyres whereas you look to be finding around 1.0-1.2g. Since you seem to be pulling a comfortable 1.8-2.0g in some of the bends I would say that with the added confidence from a new set of slicks, braking later and deeper into the bends will be your highest priority since this will make the straights "longer".

Does the new Motec software show a "friction circle"? If so have a look at the vector g in the corner entry points, with 2g of grip available it means that you can have 1.4g of cornering and braking at the same time, but most amateur racers (and I include myself in this group) rarely run at the edge of the grip circle in this zone.

Geoffrey

Hi Colin, Corner entry speed is the single area of improvement I can make right now. As I have more and more confidence in the new car I'm working on that which may be shown by the data. I don't brake as hard and as late as I can as I'm focusing on consistently increasing the corner entry speed. Then I'll be able to adjust my brake points and brake pressure. I also have some bad habits to break from my turbo car which locked up the brakes easily, especially on this very bumpy track, so I tend not to brake very hard. Lots of new things to learn. I find your observation interesting about the brake g-forces and will have to look at this when I go back on Sept 14. I know that entering turn 1 (big bend) I'm standing on the brakes enough to activate the ABS. I am going to reduce the compression dampening in the front to help with the bumps and to get some weight transfer to the front of the car to help reduce understeer on corner entry. I'll also probably fit the new slicks I have and do a comparison. I just haven't because I wanted to get use to the car on the old worn out hard slicks that break away easier. Here is the friction circle. I had avoided posting it due to the conversation about a year ago with Colorchange on the Racing forum.

NineMeister

I would say that if I had to pick one aspect of my CS that took the most time to learn it would be braking late and trailing the brake into turns, even on the limited "margin for error" narrow hillclimbs that I compete at. However, even with the stock RS ABS system, I still try to maintain the habit of generating weight transfer on application of the brakes by feeding them in harder and harder as I approach the turn in point and then reverse the weight transfer back onto the rear by slowly releasing the pedal effort as I approach the apex, whereafter my right foot is firmly on the loud pedal again.

Looking at my own friction circle from a Racelogic DL1 logger I can see that even now I rarely run on the edge of the maximum vector grip, occasionally I get there but normally I see around 90% of the theoretical limit, which I suppose is my way of avoiding lunching a £50k race car that I cannot afford to repair.

That said, on my road legal race tyres I usually hover around a 1.2-1.3g limit (the attached sample is from my winning run at a recent Goodwood sprint), so it is mighty impressive to see your car pulling close to 2g on your worn slicks and when you do get to grips (sic) with the brake performance you will be knocking more seconds of your best lap time.

Geoffrey

The tires I'm using are the Michelin GT3 Cup specific slicks someone gave me. As best as I can tell they had 6 heat cycles on them and were used in the IMSA cup series here in the US. I then ran them 3.5 hours on the test day (4 heat cycles), 3 days at Mosport (12 heat cycles), and last Saturday (5 heat cycles). They are definately heat cycled out. I have a new set of Michelin slicks, however, they are the generic S8A and S7A compound in the same GT3 Cup sizes. These tires should be slightly better for my car since it is lighter and the tires are constructed for a lighter vehicle than the GT3 Cup car.

If I recall, the DL1 has built in g-force sensors and is a GPS based data logger? Your data is interesting, but then it is a hill climb vs a circuit track so that may be why it looks a little different. I'm interested in the math of 1.4g long/lat as the theoretical maximum. I think I understand it, but I'm not sure. The I2 software is great because I can click on a view of the track within the friction circle then highlight any single dot and it will show the position of the car on the track where the sample was taken. In looking at some of the samples, it is interesting to see that the highest g-forces occur on a flat section of the track, and not on the uphill corner (turn 4) where the car compresses significantly. I suppose the guardrail 2ft from the edge of the track might discourage me from pushing it there. The corner marks the beginning of the slow section of the track for me and where I have the most room to make up.

I made some suspension adjustments tonight and programmed the traction control to get that function going. I also have begun looking at how to install the shock position sensors and the steering angle sensor which I want to do over the winter.

NineMeister

The maths of the theoretical lat/long g comes from pythagoras. Imagine a 2g vector at 45 degrees from each axis and where x=y:

2.0^2 = 2 x^2 thus x^2 = 2.0 or x = sqr2.0 = 1.414

OK?

If your car was capable of sustaining a maximum of 2.0g in both braking and cornering, the theoretical perfect corner on the g-circle would follow a 2.0g radius from braking to full lateral, then unfortunately would follow an elipse to full acceleration since the car would not have the power to register more than 1.0g in any gear other than 1st (unless you have been to RS Tuning ).

Geoffrey

"unless you have been to RS Tuning"

Got it, I thought that was where you were going.

Geoffrey

After getting rained out on Thursday at LRP (I stole my wife's 944S2 while she was sleeping...but that is another story), Saturday was a near perfect day once the track dried out. This was the last nice track day there for the season and I turned my fastest time in the LAST lap of the the last session wahoo. Thankfully Bob was there and I was able to follow him around the track and I figured if he can do it in a 964 Cup, I can do it in my car. In looking at the data between my last fastest lap (59.5) and Saturday's fastest lap (58.7) I can see that I was carrying more speed into the uphill and downhill which made the difference. Everywhere else is about the same. So, this is good in that I turned a respectable time, but bad in that this car should be capable of low 56 times so I have a ways to go.

TD in DC

Thanks for sharing. Quick question. Why are the G plots so spiky? Is that because the data is dirty? Also, would any of you be willing to look at some of my data and help me learn how to interpret it?

Thanks.

Geoffrey

The g-force data is raw data and I haven't applied a filter to the data channel in the software. Also, I hard mounted the sensor on the tunnel with a piece of rubber between the sensor and the tunnel thinking that would be enough to dampen the vibration of the car...it wasn't, especially with the solid transmission mount right under it. I need to unscrew it and use a piece of velcro to attach it and that should make the sensor pick up less vibration. That is how it is typically done, but they get knocked around as people are in and out of the car and when belts and such hit them. I wanted to find a better way, but I'm not sure this is it.

What data system? Sure.

TD in DC

Awesome. It is a DL-1. I would like to e-mail you two files. One from a session in April, and one in a session from the weekend before last. If you could take a look at it and walk me through it so that I can see what there is to be learned, if anything, I would be very grateful. Thanks.

If you like, pm me your e-mail address and I will e-mail them to you.

DrJupeman

Geoffrey, 58.7 is stout in any car. Well done!