BRC 1955 - The Golden Age [Qualifying R5]

[TheBobWiley]

I never realized flatplanes gave better performance. I wondered why so many people used them over the crossplane which is smoother (its all about that comfort stat right?).

[strop]

Actually it depends on balancing. XP is also slightly heavier and may or may not give lower peak output. I've found setups where XP would beat FP and vice-versa.

[Leonardo9613]

Flatplane v8s can rev higher, since they don't need balancing shafts. So they engine is lighter and has fewer moving parts. The downside is worse smoothness.

[Packbat]

...well, I just looked up locking differentials on Wikipedia, and I think I see why I chose not to put an automatic locker on my car:

• Pros: Automatic action, no driver interaction necessary, no stopping for (dis-) engagement necessary, continuous driving even in unforeseen road condition changes
• Cons: Increased tire wear and noticeable impact on driving behavior. During cornering, which half-axle is uncoupled is dependent on torque direction applied by the drivelive. When the torque direction is reversed, the speed of the driveline is suddenly forced to change from the inner to outer axle, accompanied by tire chirping and a large jerk. During cornering, the automatic locker is characterized by heavy understeer which transitions instantly to power oversteer when traction is exceeded.

...I'm really mad about this. 32 cars out of 40 entrants used automatic locker differentials, and I didn't because I saw - correctly! - that there was no physical way in which they could make my car faster on the racetrack. No. Physical. Way. A car that switches abruptly between heavy understeer and massive oversteer on cornering cannot be faster on a racetrack - it cannot pull higher cornering gs, it cannot burn less of its tyres, and it cannot have anything approaching the driveability of a car with a simple open differential. Real life racing cars of this era either had primitive limited-slip differentials that would be a lot more expensive than the automatic locker or they had heavy-duty open differentials, and the same should hold here.

This needs to be fixed for the 1965 BRC. The Automation devs need to fix the cornering gs model (no way an automatic locker should pull 1.2 gs in a tight circle test) and Der Bayer needs to fix the BRC tyre wear model. And if the cornering g issue is not fixed in Automation's cornering model in time for the 1965 BRC, locking differentials need to be banned from the competition.

[GenJeFT]

...well, I just looked up locking differentials on Wikipedia, and I think I see why I chose not to put an automatic locker on my car:

• Pros: Automatic action, no driver interaction necessary, no stopping for (dis-) engagement necessary, continuous driving even in unforeseen road condition changes
• Cons: Increased tire wear and noticeable impact on driving behavior. During cornering, which half-axle is uncoupled is dependent on torque direction applied by the drivelive. When the torque direction is reversed, the speed of the driveline is suddenly forced to change from the inner to outer axle, accompanied by tire chirping and a large jerk. During cornering, the automatic locker is characterized by heavy understeer which transitions instantly to power oversteer when traction is exceeded.

...I'm really mad about this. 32 cars out of 40 entrants used automatic locker differentials, and I didn't because I saw - correctly! - that there was no physical way in which they could make my car faster on the racetrack. No. Physical. Way. A car that switches abruptly between heavy understeer and massive oversteer on cornering cannot be faster on a racetrack - it cannot pull higher cornering gs, it cannot burn less of its tyres, and it cannot have anything approaching the driveability of a car with a simple open differential. Real life racing cars of this era either had primitive limited-slip differentials that would be a lot more expensive than the automatic locker or they had heavy-duty open differentials, and the same should hold here.

This needs to be fixed for the 1965 BRC. The Automation devs need to fix the cornering gs model (no way an automatic locker should pull 1.2 gs in a tight circle test) and Der Bayer needs to fix the BRC tyre wear model. And if the cornering g issue is not fixed in Automation's cornering model in time for the 1965 BRC, locking differentials need to be banned from the competition.

That's not a complete universal problem with the auto lock, it depends on the type. If you have a basic one then yes they do. But then you got the diffs like Gov-Lok and the ZF on volkswagons that act as a open diff until you apply enough power to get wheelslip at which point they lock. I think the game is simulating the 2nd type over the first.

[Packbat]

Here is a video about the Gov-Lok locking differential. When it locks, it locks both wheels to spin at exactly the same speed - which is what produces the behavior I'm talking about.

Here's Automation's pop-up tooltip on the automatic locker:

The text is unfortunately cut off in my game.

Automation pop-up automatic locker.png (14.74 KiB) Viewed 3097 times

As for the ZF differentials from the 1936 Auto Union race cars (and early Volkswagens), they were the primitive limited-slip differentials I was talking about - this thread discussing the operation of cam-and-pawl type of differential explicitly points to the way these allow intermediate amounts of torque transfer between the equal-torque of an open differential and the equal rotation speed of a locked differential as a difference between these differentials and automatic lockers. Their major disadvantage over modern geared LSDs (which would seem to be the closest analogue) is that they wear out very, very quickly.

That said...

...the same thread includes someone talking about using an automatic locker in vintage race cars and setting lap records.

I'm still deeply skeptical of the implementation in-game - I would like to see the logic behind the relative tameness [edit: and cornering g] penalties for automatic lockers and for straight-up snap oversteer - but finding evidence that there existed actual racing automobiles using automatic lockers is enough to make me no longer want to quit BRC over this.

[Der Bayer]

I just want to add that the cornering is calculated off-power. So as long as there is a differential all cars do basically the same.

I do agree that it could have some influence on tyre wear. But calculating why and when increased tyre wear is to be expected is something I currently have no good idea for.

[Packbat]

The additional tyre wear would come when cornering under power, and correspond to the rear wheels slipping while locked because they weren't spinning at the right speed for the turn. I think you could probably measure it in a similar way you measured tyre wear due to understeer or oversteer.

...is it wrong for me to want to have taken a vehicle dynamics course in college just so I could make more informed suggestions to the devs? Because it's killing me that I can't fine-tune the suspension to ensure optimal weight transfer during corner entry and exit. I literally spent more hours (like, double the hours) tuning my suspension settings in Forza Motorsport than I did driving races, and having nothing but a single yaw rate curve to go by in Automation is agonizing.

[Der Bayer]

But! Big but! Currently the track simulation (even my improved one) are pretty simplistic when it comes to combining longitudinal, lateral and vertical dynamics. Basically they are all simulated independently with only little interaction. No where close to reality. Vertical dynamics are not simulated at all on the tracks. And already the calculations are slow as fuck. I want to and will improve the simulation further and further. But if you want to have the effects of weight shifting in the simulation, you really need a) a full scale vehicle model (doable when I find time) and b) PROPER car controlling AI (we are miles away there). No wonder you think you cannot get the weight transfer on corner entry and exit right: It is not simulated at all.

The problem with the differentials is: When is it locked? We don't have any parameters given. And they should differ from car to car (probably depending on engine power and tyre choice...). I would have to make a really good estimation for the parameters from car to car in the background. Not really possible. Especially because the effect of differentials currently is not really simulated, except for a very basic approach during acceleration.

More yaw rate graphs? Remember: Tycoon game.

[Packbat]

I get that it's a tycoon game, really - but when there are spring, damper, and sway bar settings available for me to change, I assume that they will affect handling. Like, if the game had nothing but a slider that went from "wallowing" to "twitchy", I wouldn't be bothered by the lack of information, but I actually know what springs, dampers, and sway bars do in an automobile. And not just in the sense that I could walk into any university and take their senior- or graduate-level vehicle dynamics class - I have the engineering background for it, all I would need is a quick linear algebra review - but in the sense that I can get in a car in Forza 4, drive a lap, and say, "Hmm, my car needs stiffer rear sway bars, softer rear springs, and ... mm, maybe a little less front damper?" As it stands, Automation gives me all the tools I need to make the car handle any way I want it to, but not the information to know how it is handling - and I guess that could be fixed by taking away the tools (in the same way that you don't give the player full detailed control over the shapes of the cams, just a slider), but it seems wasteful to have all these parts and not make them work.

[Der Bayer]

They do work properly, but for the Tycoon mode (that's what they were planned for initially, test track was just an additional gimmick). Every slider is a compromise between sportiness, drivability and comfort (and it tells you how good you are approaching that). It's more complex than wallowing and twitchy. The test track of course is not on the same level as Forza. And the advantage Forza has is that you can "feel" the car. This is something which is very likely never going to happen in Automation.

I also want to add the way car testing (generating the stats) in Automation has gone: It was pretty slow in the beginning, with car testing taking a few tenths of a second (definitely an annoying timespan when you were making small adjustments) in the beginning, too much complexity for the vehicle model was not really allowed. I tried to keep everything as simple as possible while still giving the planned sliders a realistic and significant influence on the stats. I and the devs were pretty pleased with how things worked out (almost no balancing was necessary, the only major change was the additional bottoming out stat), although the lateral, longitudinal and vertical dynamics tests were separated from each other, each building on a model which was just as complex as necessary. Now after some optimisations by Caswal calculations are so much faster that I am thinking about improving especially the cornering model. Currently we are running a single-track vehicle model for that, with a two lane body roll model. A two-track cornering model would be more realistic (giving wheel offset a bigger influence) and could maybe solve the "sweet-spotting" in the cornering calculations (some of you did notice that and Killrob mentioned it in his most recent Let's Play). If only I had more time at the moment...

[Packbat]

That's helpful to hear. I've been complaining because I love the game and I want it to be better. I think you've told me some things that make it much clearer to me how this works now, and hopefully that will mean I can be more helpful and less ranty in the future with my suggestions.

[Killrob]

Thanks Martin for taking the time to explain this in detail. Considering our current level of the simulation and resources, I think things are just fine the way they are. Racing is NOT part of the game in any way. The strong penalty on drivability when choosing the automatic locker diff hurts the car stats enough already and it seems balanced to me so far (in tycoon gameplay).

[Packbat]

...god, this is my foot in my mouth, isn't it? This tastes awful!

I didn't understand the tooltip describing the handling of the automatic locker. When I read "These handling issues make an automatic locker diff a poor choice for any vehicle requiring good" &c., I thought it meant the car would suffer a 30% penalty to driveability, not a 3% penalty. But that confusion doesn't excuse my entitled, arrogant, condescending attitude in the rest of my comments on the last page of the thread - especially not my refusal to hear people explain to me (as they have before) that simulating vehicle dynamics is not trivial, however hard I try to make it sound that way.

I'm sorry for being an ass and I'm sorry for derailing the thread.

[TheBobWiley]

Someone owning up to their not-so-nice posts..... my god, what universe am I in?