More tuning
Well instead of faffing around with cameras, here's a proper tinkering update. If you've been following along, earlier this year I got the SoS Link installed, running, and pretty drivable by tuning from the o2 correction values. The next step was to tune tip in / accel fuel, that is: how quickly does it blend when you stab the throttle?
Generally throttle response was ok, but a little on the doughy side. It was worst when going from vacuum to boost at lowish RPM.
I’m not a super advanced tuner and have usually tuned tip in by feel. The basic gist being: if it bogs, too rich; if it hesitates / skips, it’s too lean; if it feels good, it’s probably fine and I should spend time on something else.
The car bogged, so it was a little too rich. The wideband would show a brief lean spike, go way rich (high 9s - low 10s), then even out to target AFRs (11.5 in boost). This little half throttle stab does a pretty good job illustrating:
Based on experience with other ECUs, this would tell me the tip in fueling was correcting too much and / or too late. The weird thing, though, was that out of the box, tip in fueling was basically off. It hardly ever added any correction to the final fuel value, even with very large TPS movements. WTF?
Turns out the Link ECU has a “wall wetting” feature. When set correctly, you may not need any / much tip in correction. It’s only a single parameter: percent 0-100. The SoS basemap had it set to 10%, and tip-in correction turned way down. Cool, but the lack of wall wetting timing control made me nervous: if I turn it down, will the lean spike be worse? Why can’t I just have it wet those walls earlier?
I decided to experiment with it. What happens if I do turn it down?
But before messing with it, I figured I’d turn off O2 correction. Up until this point I’d been tuning with O2 correction ON (closed loop). O2 correction values in boost were showing that it was pulling about 1% fuel, indicating the fuel tables were pretty close.
I did a little driving and a few pulls at different boost levels with the O2 correction off. Fuel in vacuum was good enough, yay:
Fuel in boost was rich, though, especially at lower RPM. Maybe tuning from O2 correction values in boost isnt so accurate...
And, strangely, the rear bank was even richer in boost (it’s leaner in vacuum):
After several passes pulling fuel and messing with the VTEC crossover to smooth things out (the Link uses 2 fuel tables), I got it to a pretty good place. A chart of a couple quick stabs, it’s a lot better:
And here are the logged AFRs cautiously working up from partial throttle 4psi to full boost (~8.6). There are some weird cells in there caused by transitions (e.g. letting off the throttle between shifts). But other than that, it doesn’t get a whole lot better than this, as far as hitting AFR targets goes:
Guess what? Tip-in feels really good now, legit throttle response is back. I didn’t even change the wall wetting, it was probably doing its job. What was happening? O2 correction disables on large TPS changes, so it’d run open loop for a bit, go rich because of the tables, then the O2 correction would catch up and lean it out to the target.
There is still sometimes a small lean spike, but it’s not consistent enough to mess with, and not causing a skip, so all good. Maybe I’ll play with the wall wetting in the future out of curiosity, but it feels real good now.
Next stuff
I’ve been moving pretty slow on car projects this year. The next logical step is to take it for a proper dyno session. I have some new bad ideas, though, they will need to happen before the dyno. It’s looking like end of the year or early next before it goes to the dyno.
In the meantime, I got a brand new OEM carpet from Mita, and a few new interior bits to go along with it. Those should go in over the next couple months.
: "Ight imma head out"
