Volume III, Issue 1, Page 10

Just When You Get Cocky About This Whole Car Thing...

’ve been working on my car’s fuel injection these past few weeks. This is no fun. After spending years making floors, welding control arms, and routing brake lines, I’m in the last stages of tuning.  Last summer was about springs and shocks; things you can touch and mess with.

Now I’ve got a much bigger challenge: deciphering the arcane language of electronic fuel injection.  We’re not talking about easy stuff like fuel tables where you see if the cell is rich or lean and change the number accordingly. There you’ve got an O2 sensor that tells you what the engine is doing at that load, rpm, and throttle opening. Hell, the hard part about the main fuel table is getting the engine to use the cell so you can read the output from the sensor.

Nope, I’m talking about the most difficult part of custom EFI tuning, stuff car companies spend millions (dollars, man hours, etc.) developing: cold start. The average number of programs assigned to the cold start syndrome is 725—that’s just for the cold start thing, nothing else in the spectrum.

Cold start: the part where you get in your daily driver, turn the key, and no matter what the conditions, the engine starts within a second or so.  I’ve been working on the bloody cold start on this damned car for 6 weeks, and after all that tuning, it isn’t much better than when I started working on it.  For one thing, you only get one chance a day to try.  Once the engine starts, the behavior changes until the next time it sits overnight.  My setup has sensor and parameter logging, so I set up the log, get the car started, let it run a bit, snag the log file, and go for a drive.  I guess that’s the part that keeps me coming back: at least I get to take the car out for some fun.

All the fun evaporates when I get back and start looking at the log file.  So far, all the experiments I’ve tried have only seemed to make matters worse.  The engine takes 6 seconds to start the first time, dies, and then starts for good on the second try after cranking for 3 seconds or so.  I’m convinced my starter is not long for this world.  There are cranking fuel tables, IAC (idle air control) motor settings, crank pulse-to-enrichment tables, coolant enrichment tables, cranking spark tables, RPM to spark tables, and more I’m sure, all having a bearing on this puzzle called cold start. 

When I view the logs I can see where the engine rpm is at cranking speed, and finally after about 6 seconds the engine start sequence begins, the engine starts, runs for a second or two, then it dies.  Six seconds for it to actually start is a LOONNG time.  Then, with no real difference except that the engine ran briefly, with the next crank the engine starts right up, going into fast idle, then slowly decaying to the programmed idle speed. From then on the car is fine, and I can drive it just like a new car. No further warm-up is needed to go cruising.  Just get in and go.

So why won’t it do it the first time?  When will the fun return?  At this point, I don’t know.  I may never know.  I do know one thing: I’m no EFI expert. But I’ll keep trying.  


Here's What's New!