Well I went for a job interview at Toyota F1 to write their realtime F1 simulation system - described as being 'like a playstation game'. Drivers use it to learn circuits. Engineers use it to test new components, systems and setups etc etc. Data is constantly fed into the sim from the car telemetry which is used to improve the sim, it's a cyclic process. I don't know or care if Tristan has more knowledge in this area but we both clearly have more than you.
F1 technology filters through to the consumer market and happens quicker than 1 lifetime. That was simply the point I was making, don't know why everyone on these boards feels the need to have petty arguments about every tiny comment. The words 'get a life' spring to mind....
Technology does indeed seep down within one lifetime. But the fact is they are using one car (each year, although not huge changes between cars anyway), and spend millions of man hours on that one vehicle.
How can Joe-Bloggs, with his nine year old car of questionable quality (bush wear and tear, tyre degradation, chassis fatigue etc) get anything CLOSE to accurate data out of it in a couple of hours (when he doesn't even know 99% of the required data to input, despite 'lots of research') when the F1 teams sometimes struggle too.
If you did go for a job interview at Toyota, then you must be either cleverer or more intelligent than any of your posts, technical or otherwise, have shown. In which case I'm sorry for assuming you are what you type like (a moron). Having said that, clearly you didn't get the job...
Alex has quite a bit of first hand race engineering experience too - he knows quite a bit of stuff too (and has put me right on more than one occasion). I'd go so far as to suggest he knows more than you and mikespeed combined, judging from his technical posts, his ability to discuss technical stuff, and his ability to remain light-hearted during a discussion (and his spamming ability - but he learnt that from yours truly!).
I didn't get the job but that was because they didn't think I'd stick it out in Cologne for more than a couple of years before going to get a job with a team in the UK. I blitzed the technical interview
I'm not interested in an argument, all I said was, quote:
"F1 teams test components and setups in simulators. I don't see why consumer sims can't or wont eventually be used in the same way."
Sometimes it's not what people write but they way others interpret them that is the problem. I will say this, when a ford fiesta has as much telemetry gathering capability as a modern f1 car, consumer sims will allow simulations. It will happen, that's my opinion and that's the last post from me.
My advice - if you DO have the technical ability (and therefore language) to discuss this sort of stuff (and bear in mind I know several people from uni that worked at F1 teams from machinists to designers, so I know that a machinist doesn't necessarily understand all the technical stuff), can I advise you to stick to it, and use the language. Bamboozle me with 'tech-speak' every time something like this comes up - who knows, we might both learn something.
Contributing to a 'technical' thread with "GO **** YOURSELF TRISTAN - and I'm sure I speak for most people on this board, probably everyone you have ever met and worked with and your mother too." hardly lets us see you have much technical abilities, and [from memory] other posts on technical matters have left me equally unimpressed.
Have you ever played around with X-Plane? It's an interesting flight simulator that it allows you to do just that for flying machines. Prototypes have been made of planes and copters using it's simulating capabilities and for airfoil simulation it's truly impressive what it can achieve.
However, as soon as you do start playing around with it you'll see that it was made from the ground-up to be like this - and it has taken several years to do this. And it's not all that simple to use either.
To change the aim of a piece of software at any given point is pretty hard. I imagine it's even harder for a simulator.
Currently as it stands LFS isn't that much of a thorough simulator, in the strict sense of a datacrunching and information outputting piece of software - like you'd find in the labs of some automotive production company, to actually give you usable data - I believe it's model could also mislead you if taken to extremes or just give you simplistic answers which could be deduced by reasoning to start with.
In a far-fetched way it would be like saying you could use Crayon Physics to solve architectural problems. It just might give you a sort-of close idea for a generic concept but you could do that just as well and faster in your mind's eye. And the more you study the more your mind's eye is able to reproduce.
Wasn't the majority of the real design work done in CAE programs similar to the commercially available ones that I've seen at uni open days and am looking forward to getting my hands on? Maybe the real time simulation is fed largely from spreadsheet type data outputted by said packages (although at a much greater level of fidelity and not bodged nearly as much as all commercial sims have to at some point, I include LFS in that)? Do they actually use the data analysis from the real car to drive the simulation, surely the computer models offer a more detailed view of what's going on and the real life outputs simply show the accuracy of the model? IIRC there are about 200 different data outputs on an F1 car, which is a lot for looking at what's happening to it but, given only a handful will ever be relevant, not a lot for finding out exactly what's causing the result and to scientifically diagnose a solution.
Regardless of what F1 teams do, and I'm sure useful real time simulation is a future possibility (and I'm prepared to listen to someone who knows better than me whether it's already happening), I very much doubt there will ever be useful real time simulation available in a commercially available product. An F1 car is a pretty simple thing to model compared to an old banger with a big V8 stuck in it, the F1 car has a lot less tat stuck to it than a production based car and the finished thing should be exactly what's built on the computer. If you tried to put an existing car into a sim you'd have to start scanning and precision measuring every single component and it would still then be a rough guesstimate as to what's best and the finer adjustments would be achieved through fairly rough adjustment of adjustable racing suspension and testing, which would be many times less expensive.
A plane is a much simpler system to simulate though, the only hard things it requires are CFD and stress analysis. The rest of it is pretty simple stuff really and with testing it's entirely possible to build a light aircraft with pencil and paper, in the same way you could go and build a car that drives perfectly safely down the road or pretty well round a race track using existing knowledge and a limited grasp of physics. You couldn't get away with pushing the envelope in the way an F1 car or jet fighter does though with this type of technology
Yes they do. It's an iterative process. Real life data is fed into the computer program. By comparing results of the model to measured real life data, the model is improved. It's similar to an AI neural net.
In this way don't need to know every detail about the materials and structures down to atomic level to produce a highly accurate model. The model continuously improves by simply running the simulation and driving the real car and processing the data. Sure, there has to be a starting point for the software which involves in depth modelling but if you have enough modelling points (telemetry sensors), the computer models can create themselves with relatively simple starting points.
Aircraft autopilots use this method to learn how to fly aircraft. They are able to cope with changes in their environment, not simply because of a set of human designed rules. They have brains of their own - quite scary really!
AM i the only one that just hopes to see a post like this from Tristan? I don't know are you all to prod of yourselves to actually read but this guy knows stuff. LOOOOOOOOOOOOOOOADS of stuff. So i cherish every post that is filled with information. I found out a lot of new stuff from this thread. Didn't read the whole thread (only 3/$ of the first page, then the piss stuff started). I have actually learned a few things from the rare posts from Tristan i've read (don't read him that much, sadly, where can i catch you Tristan? ), which will hopefully come to use in the future (aiming for a Facutly of mechanical Engineering...).
Interesting stuff but it sounds like the real time simulation is simply being based off non-realtime computing that's refined with the real telemetry, much like I suspected. So the product in discussion producing useful outputs wouldn't just be a real time simulation, but a full array of commercial (and privately developed) highly specialised CAE programs requiring a good understanding of physics and with a huge learning curve, definitely not a couple of hours of work or within reach of the home enthusiast.
He's plastered across the forum, although constructive Tristan posts are like a needle in a haystack
Agreed, but it's hard to say when a computer program become a 'real time' simulation, not just rule based - when it's modelling atoms etc? It's something that can be debated forever.
Don't forget that you're reading this on what would probably be considered a super computer 10-15 years ago. There's a bit of freeware software called Space Simulator which NASA would've killed for in the 60's and they had a budget 10's of thousands time larger than any F1 team. If you look at AGI's Satellite toolkit you'll see what kind of thing is available now commercially.
Agreed that you'll only ever be able to model down to a certain level and modeling any mechanical component at an atomic level isn't realistic and even if you did you'd never be able to get the advantage because of limited tolerances available in even the most precise manufacturing techniques of today.
The thing is though you'd have to be modeling on a level that is completely different to any commercially available sim and you'd be looking at a comparable level of detail to programs that are always going to be serious engineers tools rather than general entertainment.
as someone who
1) doesnt take forums too cerally
2) PREACHES that lfs is just a video game
this thread is wayyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy too serious for my blood. some of you people need to lighten up and realize not everyone plays lfs with hopes of being michael schumacher, and some people genuinely dont care, they just wanna play video games.
im not the best car builder tuner on earth, never claimed to be, but after having my tonsils removed and being told to sit inside ways of resolving my immense boredom get thought of, and i posted one.
but seriously, it blows my mind if any of you actually do any real racing because of how much you seem to care about your e-persona. i knew this guy once, he told the same story about how he won a spec race at road atlanta in 1987 and he was the rain man. its 2008 and he tells that story every day. he posts on the internet like some of you do. hes smart and all, but he just takes shit wayy too seriously.
As someone who
1) Also doesn't take forums to seriously
2) Knows absolutely that LFS is a 'videogame'
I still hope that LFS, in it's attempts to emulate reality to some extent are done so in a technical manner from the ground up, rather than quick bodges to keep the punters happy. There are hundreds of games designed for your intellect, which should keep you happy and away from here. There are not many games that attempt the same thing as LFS in the same way. Most sims, for example, don't care HOW something is achieved, just that the end result isn't far off. Hence rFactor is quite convincing 80% of the time. But it's that last 20%, when you are nearing the limit or slightly past it, that rFactors 'bodge it' approach falls over and reveals a completely dead simulation underneath. You like it? Great! I hate it!!!
I'm no Schumacher, and some people don't care. But YOU also have to appreciate that some DO care. And you'll find more of the latter in sim forums like LFS, nK (okay, maybe not nK), X-Plane etc. Niche simulation markets. More people care and aren't just into playing 'videogames' (which is a term I last heard in 1987 - what decade are you living it?!).
That's because you just made up the term 'car builder tuner'. No one in the history of mankind before or since has described themselves as a 'car builder tuner'.
Couldn't you have chosen something more constructive, like catching MRSA or SARS?
Have you not noticed - I don't give a monkeys about any 'e-persona'. I like technical stuff. I like driving vehicles (rather than being driven by them), and I like simulations.
What's that got to do with this? Have you tried taking stuff seriously - you'll find EVERYTHING in life is more absorbing and complex than you ever imagined. Day to day stuff, like... timezones!.. are vastly more complicated than you think about them. Get a life and learn stuff - contribute to the human race in some way, rather than bumming around wasting peoples' time.
Ah, a cooked breakfast man.
How much awesome? I can cope with no spaces in sentences for effect, but at least make an effort to punch the keyboard accurately. You'll look less dense as a result - see how southamptonfc pulled himself back from the brink of idiocy? It's not too late for you to do the same...
@mikespeed
i get where you were coming from with the idea.
The problem with Tristan is, it seems he can't think on the fundamental level. I have that problem sometimes, when someone asks me a programming related question its hard to just give a answer and convey importance of certain aspects without giving a 20 hour lecture. same with him he can't discuss spring rates without talking about a 100 other variables that you might not even know about much less care about.
so while your idea sounds good, to him he just sees all the unaccounted variables and dismisses it.
Try one of the old LFSTweaks or Mechaniks - you CAN change engine, CoG and a host of other variables. And in the setup you can put whatever springs you like on a car. And it won't handle like the real car - partially because LFS remains incomplete (and no one will deny this), and partially because those variables are only the tip of the iceberg.
Do I go on a 20 hour rant about other variables you don't care about? Be that as it may, you HAVE to care about them to get the desired function. Why do you, when discussing programming, go on the same rants about other aspects of the programming that the other person hadn't considered? It's because they NEED to consider it.
I don't give a monkeys if you didn't know or don't care about the vast complexity of the original suggestion. But the fact is it IS a complex scenario, and without proper attention paid you won't learn anything.
Besides, you can double the spring rate of one of the cars in LFS (that closely matches your real life car) and see what it does with the limited variables we have now.
i would not go so far as to say they "NEED" more like "really should" all depends on the situation you can satisfy
i think its really subjective wether or not you'll learn something. i doubt LFS will ever teach you something you didn't already know.
for someone who knows substantially less, they might learn a great deal. Now am not saying his scenario of plugging in variables from his car will net him results that are actually useful. His suggestion at heart was to be able to edit variables that give the cars their characteristics IE Cog and all that lalalalala
"but how many of you have wondered how some different rate springs/whatever would feel but dont feel like taking the time to swap everything go to the track and do it when you have LFS already on the computer"
To paraphrase "I would like to input my own car into LFS to see what drastic changes to setup data, such as doubling the spring rates [taken from later in his replies], will have on my car".
I have discredited this of being any use whatsoever because his car, no matter how carefully he measures the data, will not be accurately simulated to the degree required to get any useful data out - Stupid Inputs = Stupid Outputs. He would be no better off using, say, the UF1 and doubling the spring rates than he would putting 'his car' into LFS. And that sort of stuff is well documented, easily visualised, and pretty simple to calculate.
Lfs has a pretty good feeling of the cars when you practise a lot and you know what to change on yr setups and so lets say 30% realism?? but it would be impossible to add all movements, forces and so into the sim. and if it was possible, it would take ages to recreate your 13 slivia with an LS1 engine! and there are other facts that change yr cars handeling like weather, Human Gforces, dust/oil on track. if you wan't a realistic car, then the enviorment must be even as realistic otherwise it just won't work. best thing to do is go out in the weekend, get some spare springs with diffrent stiffnes, other tires, brake pads/discs, coilovers, strutbars, enginemounts, gear ratio's and more.. test the diffrent parts and write the diffrences in a notebook. this is maybe expensive but it is the ONLY way to learn. an yes i know what i'm talking about.. :rally_dri
I take back what I said.
), which will hopefully come to use in the future (aiming for a Facutly of mechanical Engineering...).