Some might know this link already, but those who don't will find it rather interesting. http://www.evans-tuning.com/dynos There are really lots of dyno charts/graphs with build some specs, should help in making realistic torque curves. Update: January 2018, sadly they have taken down this resource, I hope some of you managed to save some of their stuff while it did last. HERE is information how to convert torque curve from web site to BeamNG so that you get proper power with small amount of work: https://www.beamng.com/threads/how-to-get-accurate-power-curves-very-easy.45830/#post-691768 Update: Found some dyno curves from this blog, they are here and there, you need to look older posts as there are quite many of them, along with pictures of engines and specs etc. Grab while it's there: http://www.dynosty.com/blog/
This website is my go-to http://www.automobile-catalog.com You just search make, model, year, trim level, and go to the horsepower/torque curve. It will give you an even more detailed stats than regular dyno graph such as http://www.automobile-catalog.com/curve/2008/1552355/mercedes-benz_c_63_amg.html I know this thread is 5 months old, but it is somewhat important to people wanting to replicate engines.
More like 1 year, but it doesn't matter. It was a useful post. Ot - That site is the one I use, it gives gear ratios, more engine stats, and a much larger selection.
Those are more or less educated guesses, not real torque curves, same is with their performance data, they do some guessing based on how things usually are, sometimes those are close, sometimes quite far off. That is something one has to keep in mind with that site. Often though it is only option to have at least something more than max power at some rpm. I do recommend to have other reference to compare their data though, because sometimes there is some odd fluke in their estimate code that is ending up quite far from reality.
And there aren't really any curves for makes like Renault for cars past year 1980-1990. It is really weird as renaults are quite popular, but they've got basically all of the models of Dacia Duster sorted out
Member atv_123 has made nice tool which can calculate torque curves too, asking him nicely might help some people too. Also one can use software like desktop dyno, to get those calculated curves. rFactor2 guys have this kind of tool which for some parts is usable for BeamNG too, atv_123 tool is kind of similar for BeamNG, but might be nice if there would be official tool like this :wink: https://www.studio-397.com/modding-resources/rf2-physics-calculator/
Yeah they stated that the datas were software calculated and may not be accurate. No idea about the accuracy of the stock engine dynos as they are hard to find. According to your experience, what cars on that site do you find the data is very inaccurate?
It isn't so much that it is inaccurate... It's just that it is nothing more than an educated guess. If you look at a real torque curve... From a bone stock 2011 Ford Mustang GT, and then this torque curve of the same car from automobile catalog... (I know its small, I apologize) As you can see, the calculated graph is pretty darn accurate... It captures the definite overall curve and does a pretty darn good job doing it too. The entire shape is there, it looks pretty good, so what's the problem? Well, the problem is accuracy, because, as everyone here knows, good enough is really only the first step. Look at the real curve... look at all that texture in there...It is imperfect with all its little bumps and larger lumps in it... that is what gives the engine its character... It feels alive... It feels mechanical. With the current state of simulated torque curves, engines feel as if it is an electric motor that has been designed to feel like a torque curve... it is perfect... it is 100% identical every time you step on it... the machine doesn't feel alive then... it feels like nothing more than a representation of something that could be so much better. Wow... I put a bit more feeling into that than I expected to have... oh well... makes the point all the better. Engines, as refined as they are becoming nowadays, are imperfect. They can end up running slightly better or slightly worse at all sorts of different RPM's for all manner of different reasons, usually due to harmonics in different intake and exhaust runner lengths and the change in air/fuel mixture and burn efficiency that is caused by it. This means that, depending on the weather outside, the humidity, the temperature, the altitude, the barometric pressure (also influenced by altitude), oil viscosity and type, battery voltage and load on the alternator, and all sorts of other crazy factors, the torque curve of an engine will never be the same twice. Even under perfect circumstances on an engine dyno... while you can get very, very close... you will still have just a shade of differentiation. I am slowly getting my engine calculator closer and closer to that holy grail of engine calculation. Right now I am at a very similar stage as automobile catalogs simulation where you just factor in bore, stroke, STP, fuel type, compression ratio, number of cylinders, intake efficiency, boost pressure, and boost efficiency. As said before, that gets you pretty close... but I want to go further. My next step that I am working on will take into account cylinder configuration, firing order, cam lift and timing, actual stoichiometric calculations at every conceivable rpm the engine can hit at a resolution of 50 RPM's or possibly better, intake runner length, exhaust runner length, you name it, I am going to try and have it... I am hopeful that that will give the kind of detail we are all looking for.
A bit natural to put feelings on passion Point taken, thanks for taking the time to explain. Would rate informative if ratings were still there lol. Good luck on the project, I hope I will see the day it's released.
What I know atv_123 and Diamondback knows their stuff really really well, I know just a little, but can agree 100% on what atv_123 wrote. I know some reallife experiences from dynos and cars, what little I remember anymore, but even data from dyno can sometimes be too smoothed, like curves you see in sales brochures, they are too perfect, but they want to sell impressions instead of giving actual data. But it all is just starting point or at best something to aim for (real curves). Think it like when in modeling a car in Blender you have blueprints, you can have blueprint of general model of car or specific model of the car, yet still you have to make something that is not exactly like blueprint, but captures general looks of that specific car model. After countless of iterations you end up something that might be okay, but in case where you have turbo and supercharger in same engine, it is endless iteration loop.
My opinion is that it is, but that depends probably a lot from FPS, without smooth solid 60 fps even FFB starts to fail and then all that work we put creating vehicles to best of our abilities might be in vain.
If the jBeam torque curve is very precise (every 10RPM, for example), beam will benefit from that kind of torque curve. This comes at the cost of FPS, though, as more information has to moved into RAM and the CPU has to calculate power from that curve.
The internal torque curve has a resolution of 1rpm, giving it equally detailed data points to begin with does nothing to the performance. Increasing the max rpm on the other hand increases the size of the overall data structure, so that could in theory affect performance. (only talking about runtime performance here, massive amounts of jbeam will also take longer to load/require more memory at init)
the only issue i have here is that if i put in the values directly. The engine torque has less power, so i have to up the power to get the "Torque Curve" UI app to show the same amount of torque/power etc. this is very painful to do and rather annoying! Takes a lot of fiddling as the friction is dependant on the RPM and etc.
I was educated about this recently on mod support, I guess I should put this on 1st post too, really should be on Wiki: https://www.beamng.com/threads/how-to-get-accurate-power-curves-very-easy.45830/#post-691768
Do you guys see this as such a big issue? I mean technically we could implement something that reverse calculates the friction values so that what you input in jbeam is actually flywheel torque. Who has accurate flywheel data anyway though? (key point being "accurate" here)
For me at least when I make turbo car torque curve reverse method would be easier. I think other games uses reverse method, BeamNG might be one of the few or only one having it this way so maybe modders coming from other games might find reverse method more natural? I don\t know about accurate torque curves, but I do try to get real dyno sheets whenever possible. But this is just my point of view.
You might find this handy https://www.beamng.com/threads/flywheel-torque-power-spreadsheet-calculator.45889/