Forum » General Tuning Discussion » Porsche "Dynamic Boost" antilag

Porsche "Dynamic Boost" antilag

General Tuning Discussion

Discuss all things tuning in this section. News, products, problems and results. 

Page 1

Hey folks,

I was looking through some old saved links recently and recalled an article I'd read on Porsche's "Dynamic Boost" feature. To save you reading the whole article, the function is used on their turbocharged cars to reduce lag to some extent without employing traditional antilag strategies. When the driver lifts, the ECU will keep the throttle body open and instead cut fuel, allowing the engine to continue aspirating compressed air and reducing the overall loss in compressor speed during a braking event.

The article in question:

This feels like something that should be achievable with some clever overrun boost tuning to function as an extremely mild form of antilag. Any thoughts on setup? I'm waiting on my new ECU so I can't go hands-on. At its surface I'd imagine a -100 fuel enrich and no ignition advance with the throttle being held at an unusually wide open position would do it, but I have not tested whether the fuel enrich table has the control authority to cut *all* fuel or whether it could potentially result in a dangerously lean condition under the remaining boost pressure available. Any thoughts? Or is there an easier way to do what I'm thinking of?


This is a major controls problem. Porsche like most modern ECUs uses torque based engine controls. There isn't a simple look up table of accelerator pedal vs throttle position, like most electronic throttles in aftermarket ECUs. There also needs to be a very specific deceleration shut off pattern so that you don't get driveability problems. You need to write a custom algorithm to go into this mode and still be able to accelerate and decelerate correctly. If you get this wrong, you will have a very dangerous unintended acceleration OR it just won't work (won't cut fuel, won't close the throttle).

If you're not already in the business of writing controls I would stay far away from the idea. Anything that involves keeping a throttle open could get people killed.

Certainly makes sense to me. And I absolutely take your point, but it's worth noting that *all* antilag strategies not using secondary air involve keeping the throttle open [as you already know]. I don't want to throw the baby out with that particular collection of bathwater. For the sake of clarification the car in question is a time attack car that already employs DBW-based antilag on an elite 2500. We're changing over to an M150, hence my inability to experiment at the moment.

So with the requisite and very appreciated safety and application discussion out of the way, can you think of a way to approach this experimentally?

Is it a sequential box? Are you already running flat shift arrangements/have sensors in place fo the same?

I imagine it would work best with a string of logic conditions around engine speed and MAP or even turbine speed such that it reverts to conventional antilag if the braking period too long. I imagine you would probably need a combination of some form of airflow>torque model and boost/throttle angle mapping and pergaps some ignition retard to allow a relatively smooth reinstatement of power. No doubt the motec can do all of that but I imagine to do it "right" could absorbs a great deal of time between writing, simulation and physical testing. Given it's for time attack not endurance stuff surely a more conventional antilag would be good enough? Can you run torque demand model on the motec, i recon with a couple of tables and egt safety to reinstate timing and cut throttle progressivly if things get too hot you could run with a lot of base throttle angle/boost and have ignition timing reinstate torque before throttle and boost followed allowing more progressive power on than old style systems would allow.

Lets be honest that porsche system is more about PDK upshifts, diesels don't stay on boost long when you pull fuel (without restricting the intake with a throttle).

It's class-limited to the factory H-pattern gearbox unfortunately. And you're quite correct, the current antilag strategy is doing a great job and doesn't over-stress the turbocharger in the short sessions it's used for. The car lives a bit of a double life though, as it often gets used at longer race events and track days and currently is not able to run antilag during those extended operation periods. I also just tend to enjoy noodling this sort of stuff... it's always fun to try to come up with a new strategy to employ. Particularly something like this that could have potential usefulness on a street car given adequate development and a brake vacuum pump. My current street car has an identical engine to the race car and I tend to pass developments back and forth. Ironically enough the street car *isn't* class limited and thus gets to enjoy a sequential gearbox.

Thanks for the reply! I'm going to noodle it a bit at a test day and see what I can come up with.

How are you going to test whatever solution you come up with? You will need a lot of dyno time with a dyno that can do actual road load simulation. I would be super nervous about doing anything on a road or even a track at first. Remember, this isn't pushing the pedal down and activating an anti lag mode (spark retard with more throttle or whatever). This is taking your foot off the gas and hoping that you have tuned it well enough to control torque with fuel or ignition shut off patterns. That's SUPER risky.

I bet Porsche did a large amount of testing by simulation in Matlab and then in an engine dyno or full powertrain dyno, long before anyone took it on a track much less a public road.

I'm in a fortunate position as far as that goes. A friend's shop has a very underused dyno that I have consistent access to. And test and tune days at our local tracks here tend to result in complete personal access to the track without any other traffic.

I'm sure Porsche did development well above and beyond what I'll pretend to have access to or knowledge for. But ultimately we're talking about a modified antilag strategy with less fuel and retard. The current traditional DBW-based antilag that we're employing is not a launch control type event that you seem to be describing ["pushing the pedal down," etc... unless I'm misunderstanding you] but is an overrun boost type event that keeps the throttle open when the driver takes his foot off of the gas in a braking zone.

I'll be sure to report back with any useful findings once I get underway.

If the safety discussion is out of the way, this might not be too tricky to develop using a chassis dyno with heavy rollers, especially if it also has a load cell that can simulate vehicle inertia. When doing part-throttle sweeps on that sort of dyno, it's pretty common to find that the engine won't accelerate all the way to redline at low throttle angles (just like driving a car on the track). With the drive-by-wire maps set to their traditional targets (not antilag), you could record some logs and find the highest RPM the engine can reach at various accel pedal positions. Then do some additional testing with various cut percentages and retard, which should require more DBW throttle opening in order to reach the same RPM. With some luck you might be able to find settings that can still accelerate a little plus build some boost.

I imagine this could feel pretty odd to the driver if they encounter the DBW & antilag when trying to hold constant RPM while cornering. Something like ECU-controlled nitrous might be a better way to improve power before the turbo spools, if the additional weight and complexity are an acceptable compromise. The nitrous should be easier on the turbo and exhaust manifold than all the misfires and weird combustion that are associated with antilag, and shouldn't be much more stress on the engine, assuming you switch off the nitrous when the turbo spools up to full boost.