Hey guys, Ive been tinkering with the engine.jbeam and have gained a level of understanding on how to change torque and gear ratios. What I want to do is match real world engines, but to start I though to use my Xsara 1.4 as an experiment to see how to get everything to match up, so when I come to put my other car into the game I will have to knowledge to do it properly. I have the gear ratios and dyno graphs available, my engine peaks at 118lb torque at 3300rpm and peaks at 73hp at 5500rpm, I can match the torque but I currently don't know enough about changing the jbeam to bring the peak hp down to 5500rpm. As mentioned this is a personal learning exercise, so I can better understand how Beamng for making accurate performing cars. Any help greatly appreciated Phip
if your programmed torque curve matches the original exactly, then the peak hp should also match the real car (as it is itself calculated from torque and rpm). One thing to note. You've specified torque in ft.lbs. The game uses Nm
Something to keep in mind: the friction, and other variables, will force you to put in higher torque figures than are listed in the dyno chart, to reach the desired torque/power. This is because our torque values are put in, then engine/axle friction is applied, then the horsepower at the wheels is made, whereas in a real car, it's measured after all of the friction effects have already taken place. So you can't exactly put real values in; you need to adjust them a bit (increase them) to get real results.
That's very informative, thank you! I'll convert lb to Nm and input those, most dyno systems can show wheel output but also estimated crank output. That would be a better measurement to use if available considering Beamng applying resistance after the torque data. Within the engine.jbeam file there are other engine related measurements, such as "inertia" "friction" "brakingCoefRPS" "burnEfficiency". Would I be correct in believing inertia is the engines resistance to rotational change, friction is the base resistance of rotation, brakingcoefrps and burnefficiency however I'm a bit lost on. Most cars I'm interested in have some level of real world stats available so being able to input those as close as possible really sparks my interest. Much appreciated
burn efficiency basically is the gas consumption/MPG. I forget if the lower the number, the better the gas efficiency, or vice versa. you can monitor and test by using one of the debug menus, it will tell you MPG and % of gas left in the tank. I also think there may be an app that you can snap to your HUD. I also think the braking coef is basically the engine drag, so once you let off the gas, the engine will "engine brake" itself more if the number is higher. At least that is what I have experienced when I fiddle with my jbeams.
I think there's something wrong with your source numbers, or maybe there's a typo somewhere. 118 lbft at 3300 rpm is already over 73 hp. Wikipedia says 111 or 121 Nm for a Xsara 1.4, both 55 kW. Torque at 5500 rpm should be 55kW * 9549 / 5500 rpm = 95.49 Nm, where 9549 is the factor for converting between kW and Nm. I usually set the friction value to 10% of peak torque, then add the same amount to each torque value to compensate for the friction. brakingCoefRPS has basically the same effect as friction, except it rises with RPM while friction is constant. I'd start with 0.10-0.15 brakingCoefRPS for this engine.
yes I think I may have been using a slightly inaccurate data sheet. Also the 1.4 engine has been around for a long time in various models of tuning so finding my specific version wasn't quite as easy. I have found this curve which best matches the spec of my car http://www.automobile-catalog.com/curve/2004/2032460/citroen_xsara_coupe_1_4i_x.html I have dropped the approximate gear ratios into the 5 speed box and speed at specific revs matches well ie. 4th gear, 2000rpm, 30mph I have noticed that there is a lack of pull from about 60mph with an inability to reach much over 86mph on gridmap. The first thing I considered is the drag coefficient of the hatch as that of course wont match the xsara, I have also considered the variables mentioned earlier in this thread too and am tweeking to refine. The benefit I have is I can jump in the real car and compare directly, everything I learn will hopefully be transferable to other cars I'm hoping to one day make so they are much closer to their real counterparts. also I wondered if there are any Beamng specific quirks that I need to consider and adapt to that I haven't considered? I know the developers are aiming for perfection, but im sure at this stage of development there may be things I simply need to bend around
Rolling resistance is still too high, and there's a strange tire resonance thing that can cause weaker engines to get stuck in a "pit" at a certain speed, well below their proper top speed. The Covet used to only be capable of 87 mph, but after I gave it just a couple more horsepower, it was able to get over this "crest" and reach 100. Also, make sure you're using gear ratios from the transmission that's mated to the 1.4 whose performance you're emulating.
thanks guys, for the moment I have compromised and matched revs to speed, 1st gear 5k at 20mph, 2nd gear 4k at 30mph ect matching my car's performance. I have been unable to find in the jbeam where the final drive gear is referenced so that is the compromise just for now and I will address it if I can locate the gearbox final drive. Could someone please explain how Beamng measures the drag coefficient, I notice some cars have whole digits eg "dragCoef": 4, the xsara has a claimed real world drag coefficient of 0.32, how is this comparible to the in game measurements? Unsurprisingly dropping the hatch's drag down to 0.32 has lifted the top speed cap to around 110mph just 3mph over the factory claimed 107mph top speed.
Our dragCoef value is not the same as the measured drag coefficient in a wind tunnel. I'm not sure what the units are, but the range of 5-10 gives a realistic reduction in top speed (about 20-30 mph) versus a car that had no drag. Putting 0.32 for dragCoef is effectively the same as putting 0.
fair enough. Something else interesting that came out of this is the McLaren F1 has the same drag coefficient as the xsara coupe, I didn't think that would be the case!
Drag coefficients are a funny thing. Some cars with very upright windshields have low drag (Citroen DS, for example), while seemingly sleek cars (like the Lamborghini Countach) are about as aerodynamic as a house. Also, the F1 most certainly has some more downforce engineered into it for stability, since it's a supercar, which would increase the drag coefficient.
I spent some time a while ago thinking through the front end lift issues of my Toyota MR2 and attempting to design something to reduce the lift at speed that happens to this car. As I learnt more I found most cars develop negative downforce and almost all sporty modifications and designs are to reduce that lift as much as possible with varying success. Very few actually generate downforce, just successfully negate the lift that seemingly all cars produce at speed. I found the following thread very interesting as some fellow MR2 owners started using CAD airflow simulation to see the reality of the CFD rather than hypothesize http://www.mr2oc.com/showthread.php?t=410768 Their findings really helped my understanding of lift and downforce with a car I know inside out.
