Simulation of turbochargers - nitpicking

[ShodanCat 231]

So here's something I wanted to talk about with regard to DiRT 4's (and DR's) simulation code that handles turbochargers and how they spool. The way boost builds on the cars isn't really correct. Turbos build up to a pressure controlled by the wastegate and then the excess exhaust is dumped (either externally through a dump pipe or internally back through the car's exhaust.) However, the way turbos are modeled in these games, the turbos build boost in a mostly linear fashion, getting higher and higher as the engine increases RPMs. Instead, depending on the car, there should be a rate at which a turbo spools depending on engine load and RPM and then a pretty much hard cut off where the boost is limited. Some cars like R5s and the R4 WRX/Evo should be very fast spooling but relatively low boost and cars like the Group B Audi should take a bit more RPM to spool and have a higher boost limit.

The way it works now, you have to use high RPM in turbocharged cars in order to do things like hold a slide, when turbos should actually bring a nice surge of torque lower in the RPM range and more readily available. To get even more technical, the turbo modeling is pretty much spot on with how a centrifugal supercharger works, building boost in a very linear way that is closely tied to RPM (since they're belt driven by the engine.)

Anyway, the science behind it is as complex as you want it to be, but I think it could be improved lots with some different maths behind the simulation and would be both a nice boost in realism and make the cars even more fun to drive. :smile: A couple of simulations I can think of that get turbo behavior down pretty well is the Forza series and Seb Loeb Rally (although they both tend to err on the side of making the turbos a bit too easy to spool for playability reasons.)