On my way to work this morning I got my average up to 57.5mpg before getting off the highway. Roads were wet, outside temperature around 31F, target speed was 62mph. Really a huge improvement. I think I'm narrowing in on my engine tune.
A few days ago I scrapped my ignition timing maps and started from scratch again. Here's what I did:
First, I went back to an article I read months ago:
Fuel Economy Tuning - Tech Review - Honda Tuning Magazine
This applies to a Honda K20, not a K24, but one can make some extrapolations.
Superstreetonline used 3000rpm as their RPM test point. Since the K24 motor is almost identical, just 20% longer stroke, it's not a stretch to say that it should have similar ignition timing at 20% lower RPM (2500).
For my engine, at 2500rpm, at 90% load, ignition happens 27 degrees before the piston reaches the top of its stroke. Any sooner and cylinder pressure gets too high, and there's detonation (knock). So 27 degrees is as advanced as my timing can go on the 2500rpm/90% load cell on the timing table.
Superstreetonline found for their K20 engine that ~43.5 degrees provided the best fuel economy - extracted the most energy from a given amount of fuel, all else being equal. Stock ignition timing is 47 degrees for that engine.
I figure I'm running a little lean at 15.8:1 right now, so I picked 45 degrees as the intermediate value between those two. Leaner burn requires more ignition advance because it slows down the flame in the cylinder (something actually NOT good for efficiency). If I go leaner still, I'll add some more timing, but it will require very careful control of conditions to tell if I've added the right amount.
Hondata provides this chart as a rough reference:
Basically, what it shows is that MBT (timing that provides the best fuel economy) plateaus from 0-300mbar MAP, then drops roughly linearly to wide open throttle. At around 600mbar (~60% load) ignition starts to become knock limited, meaning you can't actually add enough timing to get best efficiency using pump gas - meaning you can find the perfect timing in this area by simply adding timing until you get engine knock. I'm less concerned about getting this area perfect since the engine spends so little time here.
So, I set my ignition timing to 45 degrees for 0-400mbar / 0-40% load at 2500rpm (going to 400mbar due to some EGR effect from cam advance pushing the table to the right). I then interpolated it to the 90% mark where timing was 27 degrees. Later I'll go back and advance timing until I find knock and get timing "perfect" for 60-70% load, but for now it's close enough.
VIOLA! the 2500 degree ignition timing row is, in theory, close to perfect.
It's good to have peak pressure happen when the engine is at a certain point after TDC (generally ~15 degrees). At a given cylinder pressure, it takes a roughly fixed amount of time for combustion to happen. This means that as RPM goes up, ignition timing needs to happen sooner, because the piston is moving faster while the flame is moving at the same speed. For peak pressure to happen at the same degrees after the piston reaches the top, combustion needs to start earlier.
My WOT ignition timing is fairly close to knock threshold, meaning I can look at it and extrapolate how many degrees have to be added or subtracted as RPM goes up and down. So, since 2250rpm knocks at 90% load at just above 25 degrees, and 2500rpm knocks at just above 27 degrees, that means ignition timing at 2250rpm in the 0-40% load area should be 43 degrees. The entire row loses 2 degrees timing to "correct" for RPM. Repeat this for all rows and I have my ignition table... assuming my assumptions were correct.
~
So, took it for a drive and BAM,
8mpg improvement. I still haven't quantified what cam angle gives the best fuel economy but might see as much as a 5% difference once I get that nailed down too.