Extensively. That's really a subject for a completely different discussion - the point of this discussion is to make it stay running once it fires. But since all my threads eventually turn into into free-for-alls here's what I've learned.Originally Posted by Terminal_Crazy
I had the same problem after my injector swap, before the stroker kit. It would flood when restarting warm. I spent a lot of time fiddling back then and got it "pretty good" restarting warm. Then when I looked at these tables again a couple weeks ago I realized I'd scaled the prime pulse tables in the wrong direction (made them bigger) and then compensated by removing fuel from the cranking VE tables. Totally wrong way to go about it, but it worked.
Look at the prime pulse width tables - this is a raw pulsewidth that does not reference the injector constant. The ecu fires the injectors for this period in batch mode (I believe all 8 at once) when it sees the first low res pulse from the opti. This is to re-establish wall film in the intake ports. So you'll want to scale this table by the relative difference between your stock injector constant and your current injector size. I started with about 25% (42 / 24 = 1.75) or multiplying by 0.75, but I also factored in the additional displacement so ended up around 0.82 as the multiplier. The warm and hot cells of this table will need to be massaged by hand in hex because the stock values are like 0x02, 0x03 and 0x04 and there is very limited resolution for removing an accurate amount of fueling. You'll have to settle for "close".
Then move on to the cranking VE table and use this to add / subtract fuel based on how it behaves in those first few firing events. The fueling calculations behind this table do reference the injector constant so you shouldn't need to factor that in. Just changes in pumping efficiency caused by cam & displacement.
Lastly, the Crank AFR vs Low Res Pulse vs Coolant Temp can be tweaked. This is the target AFR the calculations are using with the cranking VE table for. Notice the first two cranking pulses are much leaner than the rest. I assume this is because the first and second prime pulses aren't factored into the fueling and wall film modeling. I've no idea however if this then adds to the prime pulse or what.
Mine is far from perfect, but I'm not trying to get it perfect yet. This can be a slow & tedious process b/c you have to let it cool down to test.
That's a question kur4o might be able to answer - he mentioned it a few posts back as being set to 144 but not the address for the table / constant. But long story short the ecu uses IAC for coarse idle speed control, and spark timing for fine control. So it will move the IAC as the ecu sees the need (there's a RPM Variation pid you can monitor / graph in eehack).
My WB controller pulls the analog pin to ground until the warmup timer has expired (30 or 60 seconds). So it reads 0 (in units of lambda).
Once it starts reporting AFR it's all over the place. Sometimes showing rich of the narrowbands in OL and then usually 20% lean in closed loop. I'm really not paying it much attention because at cold start, or in any other instances where you have unburned fuel getting in the exhaust the effect is lean output from the O2. For example one of the Banish tuning guides I talked to you about says at cold start you should see the widebands showing stoich because the unburned fuel isn't seen by the sensor. Only the concentration of oxygen in comparison to other gases. Bear in mind his wideband probably costs more than my car, so I doubt we'll be able to use this method to favorable results with a $200 instrument.
I didn't get in a hurry to pull my intake off to try that, and I don't intend to unless nothing else gives results. I'd like to see some quasi-scientific data that compels me towards that before I jump out of that airplane. I already had a plan to fix my IAC bypass before I decided to drill it. Doing the same with the port feeds is a whole different animal.
Reply With Quote
Bookmarks