I confirmed that autoclutch is an issue if speed is too slow for the current gear and heavy throttle is used. While at 50mph in 6th gear in the LX6, I slowed to below 10mph, and took off again. With autoclutch on, the engine rpms quickly increase to 3000 rpm and overheat the clutch. With autoclutch off, it works like it should (engine rpms stay in sync with speed, and the clutch isn't overheated.
Looked into these, and they cost $250(USA) or more. Small aerobatic RC helicopters on the otherhand, such as the AXE CP and the Blade CP+ are selling for $200(USA) in my area so I may buy one of these. These helis are aerobatic enough to do inverted hovers. I've made my choice now where to spend my gaming / hobby money next.
Neither of these has to be that accurate if they are only used for basic skills. As I stated before, the software used to design aircraft (and air foils) is much more sophisticated. Aerodynamics at mach .6 or higher gets really complicated (affects commercial aircraft), and tables are used in addition to formulas as part of the algorithms used in such software, and also in the software for ballistics, which go up to mach 4.5 for bullets, higher still for missles. Fluid dyanmics is very complicated when dealing with turbulent flow, and again tables are used.
Thanks for the info, that's what I was looking for.
A larger motor with less gearing would do the trick, the ultimate being a direct drive servo / inductive motor, although if the torque was too high, it could break the wrists of a player if the control function went bad.
I fixed my previous post. Real world physics is usually quite complicated, and the closer a simulation gets to reality, the more complex things get. For racing sims, just getting the data can be tough. Aerodynamics is fairly well known subject, but the high end programs that accurately model it are fairly expensive. Apparently some companies are releasing tire behavior data now. I'm not sure where chassis data is going to come from. All of this stuff requires extensive tables and equations to do a proper job.
I have the impression that aircraft design software (aerodynamics, frame dynamics, engine properties) has gone well beyond what is available for racing cars, probably due to the difference in budget when designing an Airbus 380 versus a Ferrari Formula 1 car.
In the case of real racing cars, simulation software hasn't reached the point that it can be trusted, so they toss a car into a wind tunnel and see what happens, gather telemetry while the cars are running on a track, and make adjustments as required. Despite all of this, design problems still occur, leading to problems like the Mercedes Lemans car that did 2 1/2 back flips with a 1/2 twist (the 1/2 twist allowing a wheels down landing was fortunate, as the driver was uninjured):
Few things in the real world occur like an equation (technically a small set of differential equations). For this discussion, one issue is hysteresis. Near the cricital slip angle, the reaction depends on the path, whether the crictial slip angle was approached from a smaller or a larger angle.
The issue mentioned by GT5 developers is that the "magic" formula didn't work. Hysteresis is one of the reasons it doesn't work.
A table based approach will always work if there are enough tables and the tables are large enough, and cover all the state transitions. Considering the amount of memory available in a typical PC, there's not a significant restriction on table size. Higher order equations can be used for interpolation, resulting in a combination of table and equation based process, essentially the table is a very large set of equations.
The main issue in either case is obtaining the data. The quality of the output depends on the quality of the input (sampled data).
Most real world simulations use a combination of tables and equations. Ballistics is a classic example. The first usage of computers was to interpolate data to create tables for mortars, back during WW2.
My last edit didn't take place. Corrected the post to read "lag logo" instead of "ghost car".
It is very good. The previous standard was NR2003, 42 players drafting on an oval within feet of each other (closer than that and lag induced collisions would and did occur, taking out many cars in true Nascar tradition, but without the post race fist fights, a feature that never made it into NR2003). Both games seem to have a good predicition algorithm used for short lag periods, and cars just dissappear if the lag is longer.
Ok, I forgot about that, make sense now. How much lag does it take before LFS turns a car into a lag logo (ghost car with no collision)? I rarely play online anymore, and during my short online testing with patch Y, I didn't notice any cars being turned into lag logos, but maybe the players have better connections now.
I tried spinning cars in LFS; conclusion, spinning doesn't cause the clutch to overheat in LFS. I tested the LX6 and FO8 at Kyoto Ring; did a few 140mph spins, no effect on the clutch. I brushed walls to make the spins worse, still no effect.
I then drove the car backwards at Kyoto and rammed the AI cars; conclusion, crashing can damage the clutch, allowing it to slip and overheat when trying to move the car after a crash. The clutch overheats quickly, but not instantly. This seems reasonable, as the clutch plates or springs could be damaged in a crash such that the clutch wouldn't engage properly.
I wasn't able to reproduce the instant solid red CT bar in single player mode, although I saw it happen a few times in multi-player mode. I'm not sure what was different. It was the instant red that didn't make sense to me. Mabye the multi-player crashes were just more servere, and / or moved my car enough that the clutch just got overheated very quickly, so it was just a very fast transition to solid red instead of an instant one.
Spinning didn't matter, just getting T-boned at high speed and knocked sideways with no spinning pegged the clutch bar as well. Apparently certain impacts in LFS overheat the clutch, which seems like a quicky fix to model damage until a better damage model is implemented.
Where are people getting the idea that not clutching in time in a spin is going to cause the clutch to overheat? The main reason it's important to clutch during a spin is to keep from killing the engine, especially on race cars without a starter.
The reason clutch overheating is associated with spins is because a driver may forget to downshift after a spin and tries to take off while in a high gear and slips the clutch too much. I'm not sure how often this would happen in real life, maybe there are some racing cars where the clutch would overheat before the driver realized that the car was still in a high gear.
I just realized I left something important out of my original post, which is why I created this thread.
Why does a high speed impact peg the clutch temperature bar?
I was wondering why no one had responded to this, until I just realized that I left it out in my original post.
I tried out patch Y on an oval server and there are some players that like to ram other cars between races. When my car got hit by another car at high speed, the clutch temperature bar was pegged solid red and the car couldn't move. Now by all rights, my car should have been totalled, but why was this implemented via clutch overheating, or am I mis-interpreting what "CT" stands for?
I wouldn't say it's exceptionally bad. I realize I'm not a racing sim expert, but my lap times are within 5% to 7% of world record times with most games, a trend I first noticed with GPL which I played in spurts to get my rank to a -12.
I don't know where this ranks me in general with racing sims, probably above average, but obviously not "expert".
OK, that sounds reasonable, not perfect, but better than the rest. Not being an expert myself, I can only go by what I read, within reason (S1 was not as great as many claimed, but since version S, S2 is very good). I got the impression that the sim racing experts that prefer NR2003's physics don't race the ovals, or even the cup cars, but probably the Trans-Am cars on road courses.
Even I knew that the all or nothing traction in GTR was a problem, that was never completely resolved. The tire model just isn't quite right. A real GT driver made a comment that GTR2 was more realistic with traction control set to low instead of off, so in his opinion, there's still an issue with the tire physics. I bought Race 07, but I haven' really checked out the feedback on this game.
I seem to recall this being done, but don't remember the outcome. However, skill sets in one area don't always translate well to another, even when very similar. For example, some world class Superbike motorcycle racers weren't competitve on the 2 strock based Grand Prix type motorcycles, while others were. Top table tennis players aren't competitive in tennis or vice versa. I wouldn't expect a sim racing expert to do better than a normal street car driver with a natural but undiscovered talent for driving race cars after a few laps of training for both.
The most valuable point made here was the issue with the "magic" formula. I've always thought that interpolation of table data has always been the way to go, since tables are how real world data is sampled. I seem to recall Todd Wasson coming up with his own tire phsyics model, but don't remember if he mentioned any specifics, other that weaknesses in the "magic" formula. It's very rare that a curve fit and resulting equation with a few variables is ever going to be as accurate as a table with dozens or hundreds of sampled points.
I'll probably be buying a wheel and pedal set soon, maybe a G25. Recommendations welcome here. I had two issues with the Momo racing wheel and pedals. Pedal pressure was too light, especially for my clumsly left foot if I try to left foot brake (I left foot brake in my wife's car, which is an automatic, just because I wanted to learn how to do this). The other issue with the Momo wheel is the motor inertia prevents fast wheel movments, no matter how small, which is an issue when at the limits.
You're right, I overreacted in my second post. I based that statement on what I read in the other thread. Since I can't possibly know how strong the clutches are on all the cars emulated by LFS, I have to go by the feedback of others, and for some cars and players, it seems that the clutch overheating is excessive. In hindsight, I should have left out my second post, and waited for more feedback in this thread before offering any opinion. The title is correct, it's just asking if the clutch heating is realistic, since there seemed to be a few complaints about it.
I was probably biased because I still feel that it takes too long for small hot spots on a tire to cool down. The segmented tire model seemed like a compromise, but I'm not sure how complicated it would be to model variable size contact patches that would depend on the load and tire pressure during the period of tire slippage.
I know that, which is why I'm surprised that he considers LFS with PC based controllers to be the equivalent of a simulator. I've never read Tristan's opinion of how closely LFS's FOX simulates the Reynard he drives, and I'd be curious about this.
Regardless of the price, it still stands that many racing game experts and a few real world race drivers consider NR2003 to be the best. Tristan is a real world race driver and considers LFS to be the best. I don't know how much time he's spent with NR2003, perhaps with the GTP mod. I'm guessing that most of those real world race drvers that think NR2003 is best are Nascar drivers, so they are biased. The racing sim experts shouldn't have any such biases though, because of the variety involved in all the racing games currently available.
I've read that many racing sim "experts", and a few actual real world race drivers consider Nascar Racing Season 2003 to be the most realistic racing sim ever made, including LFS or any ISI based game. Used copies of the NR2003 are still selling for over $70 (USA) at Ebay, new ones for even more. There are many track and a few car set add-ons for the game.
This doesn't mean there's no room for any other racing games. I bought LFS, GPL, NR2003, and many ISI based games, like rFactor and Race 07. I enjoy them all. I also enjoy arcade racers like the Need For Speed series, Toca Race Driver 2 and 3, and even little known "budget" games like Ford Racing 2.
Other than knowledge about aspects of race car driving like induced understeer, and learning track layouts, I don't think that any of these games would prepare a person to drive a real race car, because the games can't truly reproduce what happens in a real car when at or past the limits of traction, and the feel is never going to be the same as a real car. To get the feedback a person needs in a real car can only be experienced with actual track time. I don't think any sim "expert" would be able jump into a Formula 1 car and click off sub 1:50 lap times at Spa.
You missed my point then. There are real world racers that left foot brake and do clutchless shifts with a 3 pedal race car, with no assists. It's a different skill set.
"Some people shift with the clutch, some people shift without," Hemmingson explains. "There is no best way; they just do whatever works best for them. People that drive right foot gas, left foot brake usually find that it works better for them not to use the clutch. Then there are other people that drive heel/toe and use the clutch, but there is no set rule because people are better at different things.
"All you've got to do, basically, is burp the throttle. If it isn't loaded really hard, you can pull it right out of gear, and it will go right into the next gear--if it's done quickly. In a situation like going down into a turn, burping the throttle just unloads the torque on the transmission slightly, which allows the transmission to align itself with the higher rpm of the next lower gear, and it will go right in."
One coder? Better hope he never wins the lottery and retires. I feel a bit sorry for the guy. He's been at it for years now, probably long hours, and I wouldn't be surprised if he's sufferenting from occasional bouts of burnout.
Although it seems that removal of assists, such as removal of auto-cut / auto-blip gave the coder more stuff to do, unless there was some conflict between the auto-cut / auto-blip and the new clutch model, and those assists would have to be tweaked in order to work with the new clutch, in which case removal is easier.
In spite of the name, it's a technically a beta release, since it's available to the general public (alpha implies only available to testers within or contracted by a company):
Back on topic. So is LFS better because of it's physics, or because assists have been removed? If it's the assist removal, think I could sell patches for the other racing games to remove the assists?
Regarding assists in general, why not consider what other simulation oriented racing games have done, which is basically to include assists, but allow servers to specify which assists (or even views in the case of NR2003) are allowed? The assists can always be tweaked so that using them would be slower than not using them (except when the real world equivalent has assists, like traction control and no lift sequential shifting on a Formula 1 cars, or active suspension used in some GT series).
Why should LFS be different than every other simulation oriented racing game? I can't imagine that a thread like "Join the we're glad LFS is harder and more realistic now club" would exist for any other racing game. It's no wonder that many outsiders consider a large part of the LFS community to be elitist (some also claimed LFS S1 had the best physics ever). Note, harder isn't always more realistic.
Getting back on topic about realism:
Since the gear ratios for the cars in LFS are fully adjustable, I would assume that the transmissions being simulated are racing type transmissions with dog gears, as opposed to the syncro-mesh type gears used on street cars (I'm not aware of a racing class that allows adjustable gears, but requires synco-mesh transmissions). With dog gears, usage of the clutch is optional, it depends on the driver, and length of the race (drivers will be more conservative in a 24 hour endurance race than a typcial 2 hour race). With proper throttle modulation by the driver, cars with racing transmissions can be shifted up and down without the clutch, and some very good race car drivers take advantage of this and left foot brake, which can require just as much skill as heel and toe type downshifting. Hmm, are there force feedback shifters?
These deviations are implemented to compensate for the lack of physical feedback that is experienced in a real car. One example would be exagerrated tire scrub and squeal sounds.
I created an assists and realism thread, why do the assists oriented posts keep showing up in this thread asking if the clutch heating in LFS is realistic (as opposed to exagerated or diminished compared to real life). I didn't intend this to be a thread about the removal of auto-cut or auto-blip, because I had the impression that in addition to player that formerly used auto-cut and auto blip, there are also players with "proper" wheel and pedal setups, that never used auto-cut or auto-blip, were having issues with the clutch overheating.
I've only run one test myself, doing full throttle no lift upshifts at 9000 rpm in the LX6 at Blackwood, and the clutch temperature bar only showed a very thin sliver of orange.
I didn't test the other cars, and based on the posts here, some think it's exagerrated, and other think it's OK. My guess is that it may depend on the car.
He wants to use a racing game to simulate a radio controlled model car. He wants to view the car from outside the car, so it will be similar to driving a model car. This is why he refers to antenna and camera view.
update - Ghost car - OK, Kegety has not made a ghost car mode for version Y of LFS.
If the front tires are steered too much, he wants to know why the car doesn't just go in a straight line, instead of still turning. My guess is that he has seen a model car that was able to steer the front tires so much that the model car would go straight if the front tires were turned all the way.
Then he asks why rFactor is different than LFS. It is because the car and setup are different in the rFactor demo and the LFS demo.
The current rFactor demo has a 60 minute limit on game play. The LFS demo does not have a limit on game play, so LFS is better.
To Lerts
Do a internet search (ask, google, yahoo) for:
radio control car simulator
rc car simulator
virtual radio contol car
virtual rc car
You will find more free demos that are for model cars.
This could be related to which side the driver is on (left or right). I've noticed that the front tire on the side opposite the driver seems to get hotter than the front tire on the same side as the driver when braking.
The other thread was about clutch oveheating. This thread is about assists and to point out the fact that cars in LFS represent race prepped cars because of the setup options, and are not typical road cars as mentioned in a few posts here.
Regarding the removal of auto-cut and auto-blip, I was never affected because I already have macros for auto-cut and auto-blip for my joysticks (the virtual throttle and brake axis are separate with my controllers software). I was just trying to be an advacate for those actually affected, since some of them might be unwilling to post about this because of the type of responses posted in this thread.
Racing brakes, being designed for racing, won't fade unless there is some type of failure. The stock brakes on many standard road cars will quickly fade after just a few laps under race conditions.
Thanks. I was also thinking that if the inside rear tire loses grip, then the rear end of the car is getting less overall grip than the front end, and this would result in an oversteer reaction, but this would depend on the suspension setup and the difference in roll stiffness between the front and rear end of the car.