Launch control. Flatshift. How to tune?

[GintsK]

Hello!

Can you share experience what is best way to get stable boost on start line and lost boost at shift as little as possible?

Today I try set up launch control for Eclipse 4wd. Problem was get constant boost level for period more than second. As boost is builded it starts rising. And it rise up to regulated boost for driving. We need about 1.3bar for launch, but our boost is adjusted to 1.65..1.7bar.

I see reason for this in  ign retard relative to spark table (not fixed) for launch control - more boost ->more retard (based on spark table) what cause more boost.

Other problem is flat shift. Original function looks designed for other purposes (sequential gearboxes). An we stays at launch control for this. But looks like it cut all sparks if rpm is higher than launch limit. And it did not promote turbine to spin.

Any suggestions, please?

Gints

[MWfire]

If you use vems to control boost, you can adjust 1.3bar when your throttle is 87.5% open(boost scaling vs tps).

[GintsK]

If you use vems to control boost, you can adjust 1.3bar when your throttle is 87.5% open(boost scaling vs tps).

Hm. I use 1.1.27 with new strategy simple boost in "closed loop". With this kind of control I can get very close to desired WOT boost at any situation with no overshoots, fluctuating etc. But not with scaling. Boost scaling vs TPS did not work as desired - it tries manage boost with short shots to 0% DC. It cause strong boost waveing.

And other aspect - there is no time to adjust boost with TPS on start line. Given time is 1 to 5sec.

Gints

[rob@vems.co.uk]

I downloaded 1.1.35 yesterday to start testing it and saw that it had:

* shiftcut input thresholds for picky sequential gearboxes:
 - shiftcut_potlow: activates below this lo-threshold (19.6mV unit, use 95 for a normal input switch that is appr 1.9V threshold )
 - shiftcut_pothi: activates above this hi-threshold (19.6mV, use 255 with normal input switch so it never activates)
 - shiftcut TPSthreshold: shiftcut will not activate below this
 - shiftcut igncut (0..100%) not necessarily 100% as earlier
 - shiftcut retard (0..60 deg ATDC)
* boosttarget decrease above adjustable MAT
* boosttarget decrease above adjustable EGT
* if overrun_fuelresume + 1000 == overrun_fuelcut than partial gradient fuelcut will be applied (0..100%) in the 1000 RPM range
with an ALS engine with always-open throttle, this is smoother on the environment and the exhaust than an igncut based idle

Which to my mind suggests that the issues you're having have been noticed and have had some fixes applied.  I dont know how good 1.1.35 currently is, but I shall be bench testing it tonight because it is meant to have missing tooth and N+1 triggering supported, and it would be nice to aggregate the configs I have for one firmware.

[lugnuts]

Those are some exciting changes from the new software.

Gints, the fix for launch boost consistency is a 3 dimension timing table for the launch control timing retard. I do this with other ecu brands like autronic. The higher the boost goes, you can remove some of the retard. Done correctly the boost will go where you want it and no more, no matter ho long you are on the throttle.

[GintsK]

If you use vems to control boost, you can adjust 1.3bar when your throttle is 87.5% open(boost scaling vs tps).

What firmware you use? Witch strategy of boost control?

[GintsK]

Those are some exciting changes from the new software.

Gints, the fix for launch boost consistency is a 3 dimension timing table for the launch control timing retard. I do this with other ecu brands like autronic. The higher the boost goes, you can remove some of the retard. Done correctly the boost will go where you want it and no more, no matter ho long you are on the throttle.

I have experience with fixed angle only. And it was more predictable.
2D or even two point table similar desired idle rpm should be enough for simple control.

[MWfire]

If you use vems to control boost, you can adjust 1.3bar when your throttle is 87.5% open(boost scaling vs tps).

What firmware you use? Witch strategy of boost control?

1.0.73 pid controler

[GintsK]

Anybody else? Can you share experience?

My new idea is switch to other configuration (last firmwares). And tune this specially for Launch.
But this feature have not clear documentation

I have tuned two cars witch both pretends on third place in our dragrace cup 3..5L class (Eclipse 2L turbo G1 and old Audi 90 with 20VT). It is great success, because two first cars have about 10x more budget. Good launch is key to stable times. Now it can drift even 0.3-0.4s

Gints

[GintsK]

bump

[tom.k]

bump

What about reducing max retard angle and reducing enrichment? It should work better, but may cause that more time will be needed to build boost.