twinfreak wrote: ↑Tue Aug 11, 2020 9:26 am
Great work!
For the injection data: you use the Ducati Throttles?
Maybe you can read out a Ducati ECU for the base injection data set up?
Twinfreak
The Ducati uses the Alpha-N system, I'm trying to use speed-density and see Alpha-N as a last resort (on a roadbike, on a racebike, yep, no problem)
For a simple explanation as to what Alpha-N and speed-density are;
Alpha-N is basically the simplest (and crudest) way to determine how much fuel to inject, it assumes a theoretical link between throttle position (alpha) and rpm (N) and you simply tell the ECU "at ****rpm and **% throttle opening open the fuel injector for **ms"
Advantage: very, very simple, no more sensors needed than the bare minimum, no restrictions in the intake, doesn't need a stable vacuum to operate.
Disadvantage: cannot deal with changing load conditions very well especially during cruising phase (1st gear vs 6th gear, uphill vs downhill)
Speed-density bases the fuel injected on rpm & the correlation between air temperature and manifold vacuum to calculate how much air should be going into the engine (and as a result how much fuel is needed)
Advantage: quite simple (a map sensor breaking is very, very, very rare), can deal with changing load conditions very well, no intake restrictions.
Disadvantage: it needs a proper resolution in the vacuum created in the engine, so fewer cylinders and 1 throttle per cylinder can cause severe issues in the available resolution in which you are able to tune.
To explain that a bit; normal 4cyl car engine, 1 throttle plate with large capacity behind it, normal camshaft: let's say Coasting at 2500rpm 40Kpa, cruising at 2500rpm 60KPa, slow acceleratinon at 2500rpm 75KPa, full throttle at 2500rpm 95KPa, so a 50KPa resolution and the rpm range to tune with.
Now, let's move it to a VTR; 2 cyl engine, 2 throttle plates very close to the valves, lots of valve overlap: Coasting at 4000rpm 60KPa, cruising at 4000rpm 75KPa, slow acceleration at 4000rpm 90KPa, full throttle at 4000rpm 98KPa, much smaller resolution (just above idle it's even worse)
As you can imagine this makes tuning far more complicated, but there are a few tricks around it, it's just a matter of trial & error to see what works best, if you can get speed-density to work it is the preferred option on the street
Although a well tuned Alpha-N setup that has a properly adjusted closed loop lambda adjustment (ie: continous adjustment of injected fuel based on feedback from the 02/Lambda sensor) will work very well.
Then there's also the option of using a MAF sensor (Mass air flow) that will simply tell the ECU exactly how much air is flowing into the engine rather than calculating or guesstimating it.
Provided you have the data which voltage corresponds to which airflow and there's no air leaks behind the sensor.
If you build a proper airbox, make sure there's a laminar flow going into the sensor this will work with any setup.
Advantage: very smooth running, practically tunes itself (when I changed my trackday car from speed-density to MAF it took 3-4 15min runs on the road to get it pretty much perfect and it ran even smoother in the lower rev range, really an eye opener)
Disadvantage: it's a (minor) intake restriction and the sensors do go bad, especially if you use an oiled filter.
[edit]
Well, so much for the simple explanation....
Ok, simpler version then;
Alpha-N -> computer
assumes amount of air going into system, any mods will upset it immediately
Speed-density -> computer
calculates amount of air going into system, to some extent self tuning when performing mods
MAF -> computer
knows amount of air going into system, essentially self tuning as long as you stay within the max flow the sensor can take
[/edit]