TRev's BRZ RTR Autocross Build

[SVThis]

This is sort of off topic, but is a helpful reminder that most of us already know but sometimes forget when we are out on the course. My son sent this to me today and I assume it was written by Andy Hollis. I need to read it several times as my first season with the car so far isn't up to my expectations. I came from a 4th gen ESP Camaro that I was very succesful with to the FRS on street tires which I'm struggling with.

ANDY'S TOP TEN AUTOX DRIVING TIPS

[Just got back from a weekend of teaching Evolution schools and thought I'd share some stuff that I must have said a thousand times.]

1) Position first, then speed. Positioning the car perfectly is more important than trying to attain the highest potential speed. For example, you will drop more time by correctly positioning the car nearer to slalom cones than you will by adding 1 or 2 MPH in speed. Same with sweepers (tight line). Same with 90-degree turns (use all of the track). Also, position is a prerequisite for speed. If you are not in the correct place, you will not be able go faster. Or at least not for very long!

2) Turn earlier...and less. To go faster, the arc you are running must be bigger. A bigger arc requires less steering. To make a bigger arc that is centered in the same place, the arc must start sooner (turn earlier).

3) Brake earlier...and less. Waiting until the last possible second approaching a turn and then dropping anchor at precisely the correct place so that the desired entry speed is reached exactly as you come to the turn-in point is quite difficult to execute consistently. Especially when you consider that you get no practice runs on the course, and the surface changes on every run, and you aren't likely to be in exactly the same position with the same approach speed on every run, etc. Better to start braking a little earlier to give some margin of error. And by braking less you can either add or subtract braking effort as you close in on the turn-in point. This will make you consistent and smooth.

4) Lift early instead of braking later. Continuing with the philosophy of #3, when you need to reduce speed only a moderate amount, try an early lift of the throttle instead of a later push of the brake. This is less upsetting to the car, is easier to do and thus more consistent, and allows for more precise placement entering the maneuver (remember #1 above).

5) Easier to add speed in a turn than to get rid of it. If you are under the limit, a slight push of the right foot will get you more speed with no additional side effects. On the other hand, if you are too fast and the tires have begun slipping, you can only reduce throttle and wait until the tires turn enough of that excess energy into smoke and heat. Don't use your tires as brakes!

6) Use your right foot to modulate car position in constant radius turns, not the steering wheel. In a steady state turn, once you have established the correct steering input to maintain that arc, lifting the throttle slightly will let the car tuck in closer to the inside cones. Conversely, slightly increasing the throttle will push the car out a bit farther to avoid inside cones. It is much easier to make small corrections in position with slight variations in the tires' slip angle (that's what you are doing with the throttle) than with the steering wheel.

7) Unwind the wheel, then add power. If the car is using all of the tire's tractive capacity to corner, there is none left for additional acceleration. At corner exit, as you unwind the wheel, you make some available. If you do not unwind the wheel, the tire will start to slide and the car will push out (see #6 above).

8) Attack the back. For slaloms (also applicable to most offsets), getting close to the cones is critical for quick times (see #1). To get close, we must move the car less, which means bigger arcs. Bigger arcs come from less steering and require earlier turning (see #2). Now for the fun part... When you go by a slalom cone and start turning the steering wheel back the other way, when does the car start to actually change direction? Answer: When the wheel crosses the center point (Not when you first start turning back!) How long does that take? If you are smooth, it takes .25 - .5 seconds. Now, how long is a typical person's reaction time? Answer: about .5 seconds. Finally, how long does it take to go between slalom cones? Answer: Typically on the order of 1 second. Given all of that, your brain must make the decision to begin turning the steering wheel back the other way just *before* you go by the previous cone!!

Since this is a mental issue, a good visualization technique to get used to this is to think about trying to run over the back side of each slalom cone with the inside rear tire of the car. To hit it with the rear tire (and not the front), the car must be arcing well before the cone and the arc must be shallow. Attack the back!

9) Hands follow the eyes, car follows the hands. 'Nuf said.

10) Scan ahead, don't stare. Keep the eyes moving. Looking ahead does not mean staring ahead. Your eyes must be constantly moving forward and back, and sometimes left and right. Glance forward, glance back. Your brain can only operate on the information you give it.

Bonus Tip: Don't forget the stuff in between the marked maneuvers! Too often we think of a course as series of discrete maneuvers. There is typically more to be gained or lost in the areas that are in between. Pay special attention to the places where there are no cones.

[TRev]

Here is a WOT log of 93 octane. First thing I noticed is the resolution is not very good!
https://docs.google.com/file/d/0B9Y8...it?usp=sharing

I didn't expect the commanded O2 to be that rich (as low as 10.9) in the high RPM range. Because the ODB2 logging only goes to 12.1, I am going to attach my Innovate clamp on wide band when I get a chance and find out where it is really getting to. This is the first direct injection car I have logged. Being NA though I can't believe 10.9 would be optimal for performance and must be a safety number that is on the rich side and robbing some HP.

I won't have time prior to Lincoln and Peru to test thoroughly enough to try any type of blend. I will keep updating as I get more info.

[xwd]

The cars run rich above 6000 RPM, that's a known fact. There are plenty of stock dynos which show this, it sits at a the same AFR until about 6K and then it drops down to redline. There are definitely top-end gains to be had by going leaner, most of the Ecutek tunes are around 11.5:1 or 12:1 all the way to redline.

My only concern point would be at around 4500 RPM where the car makes peak torque you might suffer some knock if you go too lean.

[TRev]

Originally Posted by dwx

My only concern point would be at around 4500 RPM where the car makes peak torque you might suffer some knock if you go too lean.

As long as the ECU continues to try to reach the commanded value by adding fuel, I don't think it will be an issue as the car should start leaning out up top first by either running out of injector or pump. I am basing that guess on the fuel flow rates in the log.

[xwd]

I'm not sure what value the "commanded AFR" is in your log but when you are above a certain RPM and throttle position the car runs in open loop based on fuel trim tables, it no longer uses the O2 sensor. If you are doing a WOT run, it's never really in closed loop mode.

The fuel trims used are based on closed loop values, so it may compensate but it's hard to say. Your best bet would be to fill up with a blend, reset the ECU, let the car learn how much fuel to reach stoich on the blend in closed loop by driving around at less than 50% throttle position for a few days, then see how it reacts in open loop. Generally anything above 60% throttle you are in open loop.

People have been running up to 8psi or so with turbo/SC on E85 without running out of pump and injector, that's not the issue. The issue is the ECU can only compensate so much when in closed loop, so as you increase the E85 blend percentage you hit the limit on how much the ECU can compensate to get to stoich. On the WRX you could run about 35-40% before you ran out of adjustment room and run lean and throw a code. It could only adjust the stock tune +/- 25% and I think the BRZ is the same. Guys were able to circumvent that by increasing fuel flow, usually by increasing static fuel pressure, which of course we can't do in stock/RT. Bumping the static pressure from 43 to 56 or so PSI let you run a 50% blend.

Here is a good thread which references another thread related to Subaru ECUs and their CL/OL logic.

http://www.ft86club.com/forums/showthread.php?p=774955

[TRev]

Originally Posted by dwx

I'm not sure what value the "commanded AFR" is in your log but when you are above a certain RPM and throttle position the car runs in open loop based on fuel trim tables, it no longer uses the O2 sensor. If you are doing a WOT run, it's never really in closed loop mode.

The fuel trims used are based on closed loop values, so it may compensate but it's hard to say. Your best bet would be to fill up with a blend, reset the ECU, let the car learn how much fuel to reach stoich on the blend in closed loop by driving around at less than 50% throttle position for a few days, then see how it reacts in open loop. Generally anything above 60% throttle you are in open loop.

People have been running up to 8psi or so with turbo/SC on E85 without running out of pump and injector, that's not the issue. The issue is the ECU can only compensate so much when in closed loop, so as you increase the E85 blend percentage you hit the limit on how much the ECU can compensate to get to stoich. On the WRX you could run about 35-40% before you ran out of adjustment room and run lean and throw a code. It could only adjust the stock tune +/- 25% and I think the BRZ is the same. Guys were able to circumvent that by increasing fuel flow, usually by increasing static fuel pressure, which of course we can't do in stock/RT. Bumping the static pressure from 43 to 56 or so PSI let you run a 50% blend.

Here is a good thread which references another thread related to Subaru ECUs and their CL/OL logic.

http://www.ft86club.com/forums/showthread.php?p=774955

My background is with standalone tuning (AEM and DSMLink). What you are saying all makes sense to me and I am familiar with closed/open loop. I have never logged a car with a factory wide band sensor though. I am not sure what the commanded AFR is either. It may be a term specific to Torque.

The O2 values I am seeing at WOT closely matched the commanded value until I started adding the blend. The problem I ran into is that O2 under WOT wasn't consistent with the blends I tried even after making sure to give ample time for the blend to cycle through the fuel system. I could get a few pulls in the 12.7-12.9 range and then another pull was in the mid 13 range. Timing wasn't affected until things went too lean though. From a seat of the pants feel the torque dip was definitely less apparent and I don't think it was a placebo effect as it was too significant of a change. Without consistent AFR values, not something that I am willing to risk using.

[TRev]

Ran a local event this weekend. Our events are usually only 20-30 cars. 22 cars and placed first on index by 1.382. Missed FTD by .209 to an ASP EVO on R-comps.

[Kido1986]

I need a chance to benchmark my driving against other twins. Any chance you want to take a Florida vacation one weekend? I've got an open bedroom and empty car if you want an autocross vacation over the summer

A local National BS trophying driver put a full second on me in my own car this weekend (if I ran in the same heat he did, I could have likely cut off .2 seconds but that still leaves him .8 faster)

[TRev]

Originally Posted by Kido1986

I need a chance to benchmark my driving against other twins. Any chance you want to take a Florida vacation one weekend? I've got an open bedroom and empty car if you want an autocross vacation over the summer

A local National BS trophying driver put a full second on me in my own car this weekend (if I ran in the same heat he did, I could have likely cut off .2 seconds but that still leaves him .8 faster)

I would love to if my budget was larger and more importantly I had the vacation time. I like Florida even in the summer Do you guys run in January or February

You are doing the right thing IMO. Have various jacketed drivers drive your car during test runs/events or even better, see if they will co-drive a whole event and hopefully get some data acquisition from you both.

That was one of the best things I did. I was blessed to have an 8 time National Champ and Solo Triad Award Winner drive my car for an event. Over 8 runs, I was .041 behind him. When we looked at the data, I had the 1-2 shift advantage and would carry more speed up until near corner exit. He was left foot braking and on the gas sooner so he had more speed in the straights.

[Kido1986]

Originally Posted by TRev

I would love to if my budget was larger and more importantly I had the vacation time. I like Florida even in the summer Do you guys run in January or February

You are doing the right thing IMO. Have various jacketed drivers drive your car during test runs/events or even better, see if they will co-drive a whole event and hopefully get some data acquisition from you both.

That was one of the best things I did. I was blessed to have an 8 time National Champ and Solo Triad Award Winner drive my car for an event. Over 8 runs, I was .041 behind him. When we looked at the data, I had the 1-2 shift advantage and would carry more speed up until near corner exit. He was left foot braking and on the gas sooner so he had more speed in the straights.

We run 20+ events (not counting the second, local smaller group Martin) over all 12 months (The only upside to this hellhole of a state hahah)

He liked my setup and didn't want to change anything. Im running 1 full turn in front, 1.5 turns in the back, 35psi square (sometimes using 35.5 in rear) on RS3s). His synopsis was slight corner entry understeer and slight throttle on oversteer but not enough to warrant changes right now.

[TRev]

Running significantly more rebound in the rear this weekend. Started the first two runs today at 1 3/4 from full soft front and 1 1/4 from full soft rear. The car was a bit too loose so I went a quarter turn softer in the rear on my third run and dropped ~.5 seconds. The course in reverse at Nationals last year was my undoing so it was nice to put in a decent run today.

I only have video from the S4 because my GoPro froze on my last run.
[ame]http://youtu.be/jBY0x0liQoE[/ame]

[Kido1986]

Great driving man! Course looked difficult but fun. Slalom ending into a tight sweeper is evil hahaha

What pressures were you running and what RPM are you launching? I can see the RPM on the video but where are you aiming for?

[TRev]

Originally Posted by Kido1986

Great driving man! Course looked difficult but fun. Slalom ending into a tight sweeper is evil hahaha

What pressures were you running and what RPM are you launching? I can see the RPM on the video but where are you aiming for?

Thanks, though I had a few spots that I could have improved on. Was a slight bit late going into the slalom but since it was evil and had the tight sweeper at the end, it was better to come in hot and be a bit late IMO. I got a bit late on the second curved wallom and pinched myself a bit at the finish. I think I could have cut another tenth or two off fixing those couple of things.

I am running 32psi all around and launching is mostly by listen and feel. Between the 5-5.5K range.

I got my alignment tweaked today. Goodyear has a 3 year alignment deal that lets you get up to six alignments in that time. It was about $180 with tax so that is only $30 per alignment!

Their machine showed my toe to be a little more conservative then the last print out I had. Only -.01 left front and -.02 right front. I am now at -.05 left front and -.06 right front. Back showed previous as .04 on both sides and is now .11 left rear and .10 right rear.

I primarily just wanted to up the toe in on the rear to help allow for more throttle sooner on corner exit.

My camber is also showing up a bit different too though I didn't mess with it as it is max'd out up front. I have -1.1 left front and -1.5 right front. Rear shows as -.6 left rear and -1.7 right rear.

[Kido1986]

Holy-rear-camber-variation! I thought mine was bad at -.8 and -1.1!