Race Studio 2 Analysis

[GSpeed]

After much playing around, we've got a good way of detecting ABS engagement using an Aim datalogger like a Solo DL, MXL2, or MXG and Race Studio 2.

Channels Required
4 Wheel Speeds
Brake Pressure

Step 1- Calculate Wheel Speed Rate of Change
You need to create two math channels that calculate the rate of change of your average wheel speeds at the front and rear of the car. Thankfully, Race Studio 2 has a nice little deriv() function that does all the hard math for you.

dFrontWheelSpeed = deriv((TOYO_WHE_FL+TOYO_WHE_FR)/2)
dRearWheelSpeed = deriv((TOYO_WHE_RL+TOYO_WHE_RR)/2)

Step 2- Compare Wheel Speed Rates to Brake Line Pressure
This next function looks at your wheel speed rates and brake line pressure, and determines if an ABS event is probably occurring. The values used here are TOYO_BRAKE_PREX > 800psi, and wheel rates greater than 0 (i.e. accelerating). The theory is that under heavy braking events, the wheels will not be accelerating unless the ABS pump is stepping in and allowing wheel slip.

ABS_ENGAGE = bit_or(bit_and(IF(GT(TOYO_BRAKE_PREX,800),1,0),IF(GT(dFrontWheelSpeed,0),1,0)),bit_and(IF(GT(TOYO_BRAKE_PREX,800),1,0),IF(GT(dRearWheelSpeed,0),1,0)))

It's a little awkward to read through, but it's an "OR" statement containing two "AND" statements we described above. We use two "AND" statements to check for positive wheel rate changes at the front or rear axle. If an ABS event is detected, it returns a value of 1, otherwise it returns 0.

Based on our experience, it's better to set a relatively high sample rate even though the wheel speeds are only read at 10 Hz. That gives you a nice sharp square wave on your graph, and sharper transitions on the GPS plot. Not great form from a programming standpoint, but for this application processor load isn't a realistic concern. This is calculated once, not continuously.

What does this tell me?
In short, this shows you where ABS is engaged on a race track. By looking at a GPS plot showing the "ABS_ENGAGE" channel we just created, you can visually see where you're engaging ABS, and then compare that information to other laps. Your fastest lap, instructor's lap, someone else's lap, etc.



It's a lot of work for such subtle differences on the map, but you can see the brief instances of color on the otherwise blue track. You may be invoking ABS somewhere you shouldn't be, or vice versa. It's just another data analysis tool to keep in your theoretical box to help you get more out of every session on track.

If anyone has any other ways of logging ABS events with the OBD channels, please post them up. More discussion is always a good thing. There's always different ways of doing things, but this is the first way we've found that consistently gives coherent data. We haven't validated it with a hardwired ABS signal, but based on driver notes and feedback, it seems to be accurate enough for our purposes.

[Sleepless]

THAT is awesome! Thx

[philooo]

I think you guys are going to make us all so much faster !

I personally have a feeling my ABS is engaged, on the track, much more often than i'd like. Being able to see it visually will really help me know if I am paranoid ... or not

[philooo]

Quick question you say race studio 2, I thought the MXL2 was using their new file format and thus RaceStudio3 ? no ?
Or is it simply that you are more skilled at the Race Studio 2 coding ?

[GSpeed]

Whoops, my bad. Race Studio 2 Analysis.

[AZP Installs]

Very Nice!

-mike

[Sleepless]

Note that for me, the brake pressure variable is TOYO_BRAKE_PRE, not TOYO_BRAKE_PREX as shown above.

Great stuff, thanks so much!

[Sleepless]

Next challenge, probably not hard for you @PST:

- Create formulas for a similar track map to show the need for an LSD

Refer to: http://www.ft86club.com/forums/showthread.php?t=68310

[GSpeed]

Originally Posted by Sleepless

Note that for me, the brake pressure variable is TOYO_BRAKE_PRE, not TOYO_BRAKE_PREX as shown above.

Great stuff, thanks so much!

Good to know. What device are you using? Solo DL, right?

Originally Posted by Sleepless

Next challenge, probably not hard for you @PST:

- Create formulas for a similar track map to show the need for an LSD

Refer to: http://www.ft86club.com/forums/showthread.php?t=68310

Sure, that looks like fun! I still need to read through that full thread, I've seen it referenced a few times. I threw this together.

Ch: RearSpinLR
Desc: Shows rear wheel speed differential when throttle position is over 40%. Negative values indicate left rear wheel rotating faster, positive values indicate right wheel rotating faster.
Ch. Req:OBD2
Notes: Accuracy restricted by low refresh rate (10Hz) across OBD network.

IF(GT(TOYO_PEDAL_POS,40),TOYO_WHE_RR-TOYO_WHE_RL,0)

Ch: RearSpin
Desc: Shows rear axle speed difference compared to front axle speed. Positive indicates rear wheel spin.
Ch. Req:OBD2
Notes: Accuracy restricted by low refresh rate (10Hz) across OBD network.

((TOYO_WHE_RL+TOYO_WHE_RR)/2)-((TOYO_WHE_FL+TOYO_WHE_FR)/2)

Ch: RearSpinLR_Filter
Desc: Filters channel "RearSpinLR" to only show values when rear wheel speed differential is greater than the speed difference between front and rear axles. In other words, this filters out completely breaking the rear end loose and focuses on single-wheel events.
Ch. Req:OBD2
Notes: Accuracy restricted by low refresh rate (10Hz) across OBD network.

IF(GT(abs(RearSpinLR),RearSpin),RearSpinLR,0)

Here's that last one plotted over a lap:



There are a lot of small spikes there, so let's overlay the throttle position ("TOYO_PEDAL_POS").



Almost all of those spikes coincide with throttle application, which combined with the low magnitude and short duration make me say we can ignore them. Additionally, they're also the outside wheel for each of those corners, which reinforces that decision.

That large-ish spike at about 4500', though, does appear to be a wheel spin instance. It's mid corner, throttle solidly applied, and on corner exit. So let's add steering wheel position and see what happened there.



Yep, looks like a dab of oversteer there. We can tell from the steering wheel angle, it was a right hand corner, and the left rear wheel broke loose for a split second. Judging from the GPS trace, the driver probably went a bit wide and took more curb than he should have there. It would be good to watch the video and see how that corner went down, but hey, it's quittin' time. My workweek is over.

Myth: Plausible.


Jake

[Sleepless]

Yes, using Solo DL.