2014 FSAE-Australasia

[Menisk]

Below is a bunch of longitudinal slip data from the UQ 2014 car in both acceleration and autox driving. I'll start with acceleration. The first plot is some of our best runs when we were testing at lakeside. We found the fastest way to do the runs was to drop the clutch at 14k RPM and use the inertia of the engine to generate the slip quickly as the tyres were hooking up hard and in turn self regulating the slip. For the record track temps were somewhere in the order of 55C and we were smearing the rubber off the tyres during autox we had so much heat in them. We literally had treadwear markers disappearing as the surrounding rubber filled them in as it smeared across the carcass of the tyre.



This next plot is from the acceleration event at comp. The day was warm but the track temps were definitely not as high, by the end of the event the tyres were plenty sticky as we'd be doing multiple runs in succession and I had our driver weaving, braking and accelerating on the coast back to the start line between runs. The surface seemed dusty to begin with and we definitely weren't hooking up nearly as well. Realistically by the end there really couldn't have been much dust left with all the cars clearing it. My only conclusion is the surface was rougher and we were relying on the mechanical adhesion of the tyres far more than the chemical adhesion on a smoother surface. Here we would have been better winding back the launch RPM and starting from a lower slip. We initially responded by dropping the RPM back to a 12k dump and as the temperature came into the tyres and we started hooking up a little we made the wrong decision of putting the RPM limit back up under the poor assumption our problem that day was one of tyre temperature. Because the surface was crap we didn't end up with the tyres limiting the slip themselves and staying in their optimal range. Most of our experience has been that our 13" Hoosiers love slip. We won the acceleration event in 2013 with a 10k RPM limiter dump in first and put down a time only 0.05s slower doing a 14k dump in second without a gear shift.



The key to winning acceleration is as much normal force on the rears as possible, the power to back it up and holding the tyres in their optimal slip range (15-40km/h for us). The reason ECU beat us was likely a bit of static rear weight bias (we were 50/50 due to the shitty job of packaging an F4i motor in a minimum wheelbase) but also because they had a different link position that offered a heap of anti-squat that drives the rear wheels into the ground for a brief moment at the start line. We intended to do the same but unfortunately lacked the test time.

The next set of plots are from endurance at FSAE-A. I'll start with the plot of slip measured as a ratio. You can see a general trend and you can see that the fx-sr curve falls over as you add lateral force (the different colours). However it's messy and doesn't seem to have an awfully nice correlation.



I've ended up finding that slip measured in km/h is a far better indicator of longitudinal force. There is still noise to the curve and that is to be expected as the points are generated off different patches of road, some good, some bad. there are some interesting things to take from the picture below. The first being wheelspin is not necessarily a bad thing. You can see that the peak longitudinal force comes later in the curve as you start adding lateral load to the tyre. If you've got a driver that's running around the edge of the friction circle and putting power down as hard as your car can, he will be spinning up tyres and the back will be stepping out a bit. It's not an indication of a loss of grip so much as the tyres finding the new operating zone which needs a bit more slip angle and a bit more slip ratio. If you watch any of the cars on LC0s that are going fast you'll note that the lack of cornering stiffness means they need to drive the cars incredibly sideways. It's not an indication of a lack of grip, it's just found in a different place.

[Jonny Rochester]

Thanks for this analysis Menisk. It is obviously from real experience, and well written in a language I can understand. Communication is important as well! A description with Fz and Fy etc. is too hard for me to follow, and even a "full" mathematical model can never be complete to describe something so dynamic as drifting a car, the driver taking his own judgement to find best grip etc. Thanks for trying to bridge this gap.

[mech5496]

I will echo Jonathan, that was a great input, thanks!

[Tim.Wright]

In my experience, the laptime sensitivity of R% of a given car depends heavily on the track geometry. It favours tracks with long straights and the advantage continues also when you are power limited.

The price you pay is less cornering performance (due to tyre load sensitivity) less stability (cornering stiffess is also load dependent), a slower lateral acceleration response to steer (generally considered bad feedback to the driver).

Consider also that most of the time you are traction limited in FSAE is when you are cornering. Weribee is not Monza, like anything R% is a trade off...

[Tim.Wright]

Though, I will concede one point. With amatuer drivers, you aren't so likely to fully exploit gains in cornering performance and stability. But if you give them better traction and they will use it.

[MCoach]

One of my favorite plots I ever made looked at various radii and comparing tire angle vs lateral acceleration.

It doesn't mean much on it's own, but it's neat. Combine it with a yaw plot and now you can start doing some serious calcs and estimations.
Semi related here is the time vs velocity and yaw on a different set of radii. You use get an idea of how stability changes over different velocity and corner radii when using wheel angle, yaw, and accel measurements together. Second plot, red is V, green is yaw, units are undesignated. These are old pictures so not very relevant to our current design.

[Z]

Menisk,

Thanks for the data.

I think there are some other things that can be taken from the above plots, but on the other hand I am not sure of all the background details that go with all those little dots, so I better not speculate.

I would be interested to know if you can extract the number of "wasted wheel revolutions" you had in the Enduro event. So, the number of times the rears spun more than they had to, to generate peak-Fx, coming out of the corners. This can then be used to estimate "how much fuel wasted", and "how many FE points lost".

I am mostly interested to know if this is a BIG points difference, or not worth fussing over? Note that Monash might have had Traction-Control (???) so NOT lost too much fuel, and ECU were 35 points behind, and Auckland another 10 back. So, for ECU and Auckland, could less wheelspin = less fuel wastage = change of podium positions?
~o0o~

Tim,

The price you pay is less cornering performance (due to tyre load sensitivity),
less stability (cornering stiffess is also load dependent),
a slower lateral acceleration response to steer (generally considered bad feedback to the driver).

* The TLS is fixed by putting larger tyres on the back, and smaller on the front. See red car, previous page, for hints.

* I reckon AutoX style courses benefit from less stability.

* Slower lateral response, with SAME steering-ratio, is an inevitable consequence of increasing the distance from CG to front-wheels. But solution is simply to speed up the steering-ratio (and increase max-lock), as I suggested on page 7.

With amatuer drivers, ... give them better traction and they will use it.

Yes! And so easy to use. Just stomp on the skinny-pedal, and go fast!

Z

[Tim.Wright]

* Slower lateral response, with SAME steering-ratio, is an inevitable consequence of increasing the distance from CG to front-wheels. But solution is simply to speed up the steering-ratio (and increase max-lock), as I suggested on page 7.

A faster ratio doesn't decrease response time to a given lateral acceleration. It just means you have to move the wheel less to get there. The transient response time from ay=0 to any nominal value will remain the same.

Regarding the tyre sizes, you could equally put larger tyres on the front and rear and enjoy more cornering grip. Like I said, the correct direction to go (prioritising traction or cornering speed) will depend on the track geometry and can be understood from even a simple lapsim.

The worst case for stability is under brakes. While instabilities may help maneuverability on a tight track once you get to the point that the driver can't control the car under brakes, then some of your top speed advantage is going to be thrown away by him braking earlier.

[RenM]

I am mostly interested to know if this is a BIG points difference, or not worth fussing over? Note that Monash might have had Traction-Control (???) so NOT lost too much fuel, and ECU were 35 points behind, and Auckland another 10 back. So, for ECU and Auckland, could less wheelspin = less fuel wastage = change of podium positions?

Traction Control is not very fuel efficient with an engine that uses intake-manifold fuel injection. The only way to control the power output fast and reliable is by changing the ignition and thus deliberately making the engines efficiency worse. TU-Graz ran Traction Control some years ago and their fuel efficiency really got bad.

[Z]

Happy hangovers, all!

2015, and a brand new set of Rules!
~~~o0o~~~

A faster ratio doesn't decrease response time to a given lateral acceleration. It just means you have to move the wheel less to get there. The transient response time from ay=0 to any nominal value will remain the same.

Regarding the tyre sizes, you could equally put larger tyres on the front and rear and enjoy more cornering grip. Like I said, the correct direction to go (prioritising traction or cornering speed) will depend on the track geometry and can be understood from even a simple lapsim.

The worst case for stability is under brakes. While instabilities may help maneuverability on a tight track once you get to the point that the driver can't control the car under brakes, then some of your top speed advantage is going to be thrown away by him braking earlier.

Tim,

All three of your points are typical of the sweeping, and lame, QUALITATIVE excuses that Engineers use to AVOID CHANGE.

Sorry to be so blunt about this, but although those excuses are nonsense they will certainly convince some students that "more R%" is a very dangerous option to take.

Taking these in reverse order.

3. "... once you get to the point that the driver can't control the car under brakes, ..."

I have driven many cars with extreme R% and they brake just fine (ie. just set the brake-balance to "extreme-R%"). Putting it another way, you are saying that Porsche 911s are crap on a racetrack.
~o0o~

2. "... the correct direction to go (prioritising traction or cornering speed)..."

This suggests an exclusive-or option. Namely, EITHER the car is good in a straight line, OR it is good around corners, but it CANNOT BE GOOD AT BOTH. Although this excuse is beloved by FSAEers, it is nonsense "and can be understood from even a simple lapsim".

(Interestingly, even very small children know that the fat kid has to sit closer to the middle of the see-saw to get the right balance. It is beyond me why student engineers cannot see this. They seem to think that a see-saw can only "balance" if there are two equal sized kids at equal distances from the fulcrum. Groaaannn..., education system down the crapper... )
~o0o~

1. "The transient response time from ay=0 to any nominal value [is worse with high-R%]."

This sweeping generalisation is oft-repeated (ie. it appears in RCVD!), but is very misleading.

Consider a SINGLE given car doing the "step-steer" transient-response test that usually goes with the above claim. Fit one lateral-G-accelerometer to the very nose of the car, and another at the very tail of the car. Do the step-steer test and check the two lat-G curves. The nose-lat-G curve rises very rapidly and very high, before settling down to its "steady-state" value. The tail-lat-G curve starts off in the NEGATIVE direction, before slowing down, reversing, and eventually reaching steady-state.

The point is that any lat-G measurements made towards the tail of the car always have slower "transient-response" than those made towards the nose. This is simply a kinematic side-effect of where the measurements are taken. But the different lat-Gs, and different transient-responses, are of THE SAME CAR! The car as a whole does not change performance! This is a good example of "cogitatio caeca", namely getting lost in the alphabet soup of algebraic analysis, and not being able to see the forest for the trees.

Anyway, there are many other factors also involved, but if you are still convinced that more-R% makes for a slowly responding car, then please give a QUANTITATIVE argument, theoretical or empirical, supporting your view. Please include all the relevant parameters and assumptions. I will then show where you went wrong, and how Porkers can actually get around a racetrack quite quickly.
~~~o0o~~~

RenM,

Yes, I guess something like ETC (ie. drive-by-wire) would be needed to make TC help with FE.

When the Fuel scores come out we might know more.

Does anyone know when all the detailed scoring will be available???

Z