Quote:
Originally Posted by Cobra #3170
Hi Tony,
The system runs in alpha N just like the 498 did, way too radical to do any self
learning, the old one worked perfectly in autox. The only reason I changed was to get traction control. I am pretty sure it can be fixed, with tuning we made major improvements just smoothing the base fuel tables and changing the TPS enrichment curve. Carbs cannot handle the g loads and stumble because of float slosh. We are getting 1.5 G cornering and 1.2 braking and accelerating on 200 TW tires now.
|
I think I understand. The "analog response" of the carb and floats can't deal with the impulse inputs on the x-cross run. Hopefully a digital oversampled system will be able to do it. To do that the system needs to be able to sample an actual run and somehow factor all of its variables into the DFI adjustment tables. The problem will be in the accel/decel transitions. I had the same thing happen to me - the tables didn't really understand what happened when you were (at the same place in the table) accelerating going up hill and decelerating going downhill during shifts. After we talked about what was happening the light in the room got real bright and he "got it" and was able to smooth the table across those spots but had to compromise it. It was better, still better than carbs, but compromised the overall performance to do it.
It certainly was an interesting experience but the PhDs in HDD servo controllers were able to explain it
