Civetta Scintilla

Neither the FCV nor the Scintilla has a center differential, the AWD systems are all clutch-based, Race version or not.

Edit: You can manually swap to an AWD transfercase with a center differential in the parts menu. It's not factory installed on either GTx nor Amateur Rally Asphalt.

 

literally unplayable (TM) because "Race AWD" in this vehicle is in-name only. It's just another Clutchex. In comparison, older vehicles like the SBR4, Sunburst, ETK-I, Pessima, etc, etc. All have true Race AWD transfercase, they are all center limited slip differential with user-adjustable torque distribution, locking percentage, and preload. Newer vehicles such as the K-Series and the Bastion which have clutch-based splitshaft AWD only, don't pretend they have Race AWD, they just "honestly" used "AWD transfercase" even on race versions. Only the FCV and Scintilla have the "Race AWD transfercase" named on another Clutchex AWD. Center diff Race AWD was added a while ago, I forgot to double check, you can select it in the part's menu.

The problem: For Clutchex AWD, the Scintilla, unfortunately, handles the worst. The "AWD transfercase" on K-Series and Bastion, and "Rally"(in-name only) AWD on FCV, still distribute a good chunk of torque to the other axle when accelerating rapidly, cornering or not, because they have identical wheel diameters. The Scintilla has a larger rear wheel diameter, thus, in normal driving, the front wheels have a higher angular velocity, so in most cases the Clutchex AWD can't be engaged. The rear wheels must have much more significant slip for it's angular velocity to exceed the front wheels so the center clutch can engage to transfer torque to the front, compared to the Bastion.

High wheel slip triggers the Traction Control. Theoretically, during aggressive acceleration with cornering, the AWD should engage first upon rear wheel slip to redirect torque and reduce oversteering, the throttle should only be reduced if wheel slip persists even with AWD engagement. As I explained before, the e-diff logic and stability/traction control are controlled by separate sub-system each that don't communicate with each other. Therefore, it's very often for the AWD Scintilla to reduce throttle before enabling AWD during corner exit, causing unpredictable handling.

You may experience on AWD Scintilla (especially Amateur Rally - Asphalt (Sequential)) randomly:
Power oversteer: neither AWD nor TC engage, feels like any other conventional RWD vehicle with too much power.
Balaced: AWD engages correctly before TC.
Lift-off oversteer or fishtailing: TC cuts throttle aggressively without the e-diff engaging AWD

This makes the handling of asphalt rally Scintilla worse than the RWD Bolide rally version. Fortunately you can manually swap to a center differential AWD and mitigate the problem

The Hillclimb SBR4 has an adjustable center diff AWD transfercase, while the street AWD versions have an open center differential with a fixed 25:75 torque distribution

 

Note: Using the term ABD to refer to auto brake differentials, the automated braking of only 1 output of a regular L/R differential (left/right)

Key points get the list dots and indents.

Sooooo, I quick-peeked into 0.38.6 to see what has changed, here's my ramble on the various systems discussed on the past pages cos I got bored:

• A small part of me thinks that BNG should have locked itself to a key earlier time period like the 90's, the early 00's, or even the 2013 0.3 release year, so cutting-edge vehicle tech would be less of an issue but I digress. (And a bit late for that. But still, modern open world game guidance elements in a stylised desaturated/monochrome/pixelated/flip phone UI would be a funny combination.)

• ABD by itself shouldn't fully eliminate inner-tire-fire (ITF), especially at higher vehicle speeds. It is unclear if modern systems compensate for steering angle too, my current belief is that they still work by mapping a allowable amount of L/R speed difference that increases based on overall vehicle speed, before deactivating entirely at above highway speeds. Excess engagement at higher speeds could both cook brakes and potentially cause instability, note the micro delays added by hydraulic actuation of brakes IRL.

TBH, my future intent for personal use is to mod all vanilla factory Scintilla configs with a rear clutch LSD with ABD, like Huracan/R8 gen 2. Yes, ABD compliments non-active clutch LSDs in road car applications, Porsche has combined them all the way since ABD introduction with 993.
Can also fine tune handling balance by playing with preloads and greater lock-up during deceleration, a bit like 911s.

The only ABD applications that appear to target full equalisation, and only at low speeds, are certain off-road tuned systems. See the front wheels during roller tests of modern full size Land/Range Rovers in particular, the rear wheels may benefit from the optional active clutch LSD so have to confirm the spec first.

• While Mclarens use ABD since 7-speed models use open diffs, they go further and incorporate permanently active yaw "assist" by brake, often referred to in an excessively generic manner by non-critical press as torque vectoring by brake (TVB).

(Don't get me started on some pros who mix up the core functions of ABD with TVB or throw around the Viscous term with Haldex-style AWD systems)

Can't find any objectively measured data, but during high speed cornering, I think only TVB, and not ABD, would attempt to target speed equalisation and further over-slowing of the driven inner wheel.

TVB can be incorporated IRL on pretty much any differential type, and the key unique "trigger" compared to ABD is steering input. This engagement of the inner driven wheel brake on initial turn-in would pre-empt the unloading of the inner wheel that results in ITF, and can also interface well in just small amounts with anti-dive/squat geometry to manage the roll of the vehicle, as much touted on the 2022+ MX5.

See VAG's XDS for open diffs, Porsche's PTV and PTV Plus for clutch LSDs and active clutch LSDs respectively, and the toggle-able brake-AYC feature on the Evo X complimenting the actual active S-AYC. Even selective L/R engagement virtual diffs like Magna's (current A45, Golf R, RS3) are compatible with braking of the disengaged inside wheel to force additional yaw.

• There currently doesn't seem to be consistent application of understeer correction in Beam's ESP/ESC, go fast and turn any direction with increasing lock (but not too abrupt that the rear end loses it) and there isn't the characteristic braking of the inner rear wheel.

In a sense, TVB is an evolution of this aspect of ESP/ESC without the focus on stability and engine throttling down, and only acts on drive wheels.

Related side note: Back in the 0.1x.x era, I kludged in a simulated ABD using the old TCS code by adding a check between the L/R on the same end of the vehicle. The vanilla TCS code compared diagonal wheel speeds and would thus fail split-mu scenarios. I have zero idea how to kludge anything into the current dynamics system and am too lazy to research further atm.

Barely related side note: TVB is occasionally undesired among certain professional drivers for potentially making a car feel artificially unstable and inconsistent, and aggressive application combined with ABD on a FF platform can destroy brake systems as speculated during the 3rd gen Focus ST's sub-par result during MotorTrend's BDC 2013, possibly why the official Quaife option was later added and the petrol 4th gens outright got a standard active clutch LSD.

• Moving on, there is support, and presence, of a transfer gear ratio to compensate for different F/R tire diameter with the Scintilla, though haven't checked the lua's calculations if all the rotational velocities are being accounted for accordingly.

On a side note: if AWD transfer clutch thermals are ever implemented, would be interesting to try out a Focus RS/GR Yaris/Corolla style forced over-speeding of a rear axle on one of the FF based AWDs just to see how fast the transfer clutch could overheat...

• But in general, engaging the front axle would become a balancing game of stability and traction against the turn-in "assistance" from only using the rear axle. BMW M's middle "4WD Sport" setting is not recommended by default for setting lap times for example. This was also part of the debate back with the 997.1 Turbo when comparing the "feel" with the R35 GT-R, not counting the latter's standard rear clutch LSD with TVB. (And why PTV was invented for the 997.2 Turbo, then PTV Plus with the Panamera, Cayenne 2.0, and PDK 991s but digressing here)

Another curious theory I've held is that the pre-facelift Aventadors and Huracans (being the first gens with Haldex AWD and also silly fast and heavy) frequently suffered from brake fade partly due to high front axle engagement to aid stability/traction, not helped by ABD being used for the open front diffs, so the facelifts modulated the Haldex effect, possibly in a similar manner to 997.2. The ESP/ESC equipped Gallardo used ABD on the open front diff (all features being the first for Lamborghini), but the viscous coupling (not viscous LSD) AWD system inherently limited the max possible torque transfer to the front so it may not have been an issue then. Meanwhile the VT Diablos and all Murcielagos were among the last to use clutch LSDs with front wheels, and feature above average amounts of preload possibly to restrict yaw off-throttle, perhaps reducing spin-out crashes with the less experienced who may lift abruptly mid-corner.

• The best part is that all this is just the stability software, drivetrain, and diffs. The wheels, tires, chassis rigidity, suspension kinematics, alignment, and ten-million other factors also play a part in how the Scintilla drives. XD

 

You can easily add ADB to mechanical LSD by copying the "electronicDiffLock" controller from open diff to the LSD. However make sure to relax the control logic a little bit so the eDiff won't brake the axle where the LSD sends more torque to during acceleration. Here's an example (it's for locking differential but it's the same for an LSD)
If the preload torque is zero or very low, the LSD is like an open diff when coasting or slowing down, however the eDiff in this game doesn't work with zero throttle, you may want to write your own lua program

The AWD in this game is proactive, however the clutch is always disengaged with zero throttle. I made a mod (now outdated) a while ago, the center diff will lock under heavy acceleration and braking on tarmac mode, you can scavenge the lua from the mode and take a look.

This is another complaint about this car, the suspension tuning

In other words, if you change camber, the toe changes on its own, and when you change toe the camber changes on its own. It's much worse in Scintilla than most other vehicle like the Pessima. For serious suspension tuning you'll need to write equations and solve them, to get the correct amount of the tuning slider for your targeted alignment angle

 

Unfortunately this part is realistic and may be the reason why the tuning menu uses percentages rather than standard units

 

Please elaborate how camber/toe change on their own with toe/camber tuning is realistic(describe or provide links to technical articles)
and how can we figure out the equations like
camber = coeff_c2c * camber_set_perc + coeff_t2c * toe_set_perc + c_0
toe = coeff_c2t * camber_set_perc + coeff_t2t * toe_set_perc + t_0
4 unknowns with 2 equations good luck figuring the coefficients out. We might need to measure the angles of the suspension geometry by digging through the jbeam files or hiding the mesh of the vehicle and showing its beams in-game. This is simplified as I don't even know whether they are linear or can be approximate to linear. c_0 and t_0 also change if you change the coilovers

 

Don't have a specific technical paper, but do have an old racing book which mentions that tail-chasing due to interacting adjustments is common in chassis setup. Maybe not camber and toe directly affecting each other, I haven't specifically heard of any real-life cases of that, but for example ride height can affect both.

 

Doesn't the front driveshaft have a smaller drive ratio to compensate for just that? Or am I thinking of something else? I'm sure there's a part that has a 1.0 and a 0.x ratio that you have to switch out depending on the wheels you selected.

 

There is one, but it's not tuned very precisely, or the sensor still takes the wheel rotation into account without multiplying the driveshaft ratio. I can't really tell

 

I noticed even the DCT Scintilla stalls when heavy braking at speed and the wheels lock up. Understandable for manual/sequential gearboxes if you don't engage the clutch, but DCT too? Just wondering if that's realistic behavior.

 

There's a problem with DCT programming, originated from years ago. The DCT only slips the clutch in 1st, reverse and during shifting. When in 2nd gears and above, the clutch is fully locked. However, there's an "anti-stall" mechanism that opens the clutch when the RPM is too low. The problem is, the "anti-stall" is a separate sub-program that behaves like an on-off switch or "emergency bypass" or "fail-safe" that overrides existing logic rather than tuning within the shifting logic. Back then, the "anti-stall" RPM was set to slightly above engine idle RPM, which is good for anti-stall, but since it overrides the current running logic, decelerating gently (low shifting aggression leads to low shifting RPM) from a high gear may reach that RPM, preventing the DCT from downshifting "natually" (due to being overridden) and you can continue to decelerate gently all the way to stop with the DCT still in 5th gear. To solve this, the "anti-stall" was adjusted from slightly above to slightly below stall RPM. This prevents the gear hang from occurring during gentle deceleration but it lowers the safety margin if the driven wheels slip and stop rotating. A supercar with a high-rev engine naturally has a lower inertia than a "big block V8". Unlike the flywheel in manuals and the torque converter in automatics that both contributes to the engine inertia due to connected to the engine, there's no equivalent thing in the DCT model. These two factors makes the Scintilla very easy to stall, combined with the "anti-stall" logic of the DCT.