Turning the MIL back on would have just been a quick way to see that it's actually failing it visually. Whatever works.

 

Sorry for the slow turnaround on this but I finally got around to digging into this some more. As far as I can tell, the guy I hired to tune my car not only disabled the CEL for the MAF sensor, he also disabled the diagnostic routine. So the CEL doesn't come on, but that's because the PCM doesn't even bother checking the malfunction conditions, which means that it never considers the MAF as busted... so y'all were right, my car has indeed been running MAF mode this whole time.

Incidentally, the car had a vacuum leak when I got it back from the same shop that tuned it (and put in the cam, exhaust, intake manifold). I went back to talk about that as a possible root cause for my idle issues and was told it didn't actually matter because the car was tuned to not use the MAF. He was almost right, because he made most of the changes needed to switch to SD, but still entirely wrong because of this one setting.

I'm a little grumpy about this, but oh well. Nothing to do but move forward.

Thank you all for prodding me to dig deeper into this. Even if it did take me a while to do it.

 

What do you mean he disabled the diagnostic routine?

At least you have it figured out. I'd like to know the percentage of guys who get into tuning because of bad tuner experiences. I know I am one.

Good job!

 

There is a test that is ran on the maf during runtime to prove its working correctly. Sounds like that was disabled.

 

I was into tuning before this, but I wanted to get a head start by paying a professional... it didn't work out that way.

You can choose how the PCM will behave for each of the fault codes, HP Tuners calls it "error modes" here:
https://www.hptuners.com/help/vcm_ed...ngdiag_dtc.htm
Option 3, "no error reported," sounds like the right thing to do for the MAF in an SD tune, and that's what the tuner did... but it would be better described as "no error detection" instead. As in, the PCM doesn't evaluate the MAF sensor health, so it doesn't notice that the MAF signal is out of spec, so it continues using the MAF sensor.

Quoting from https://www.hpacademy.com/previous-w...uners-changes/ but with added boldfacing for the key thing:

 

Glad you sorted it, most give up. No one said any of this was easy even the pros on TV get stumped all the time its part of the game you have to rise above the BS and work methodically through every step to uncover the real troubles. Vacuum leak made this a lot harder for you, I did mention that in post 12 to check first

 

Vacuum leak is always the first thing I look for with idle issues too. I found and fixed that pretty early on, well before I got it to idle as good as I described post 5. But it's sound advice for the OP, for sure. Speaking of which, we haven't heard from the OP in a while. Hopefully I didn't hijack his thread too hard.

In a separate conversation, Darth_V8R gave me a couple more ideas to try, and I'm optimistic about getting my remaining issues resolved soon. Once I do, I'll be happy to meet up with the OP if he wants (I think we're in the same area), and see if what I've learned so far will help with his setup.

 

After another 20 iterations of experiment (flash, drive, ponder, edit, repeat), my car has gotten pretty good. Good enough that I started reducing the idle speed (initially 800, now 700) even knowing that might bring the problem back. To recap, my problem wasn't exactly idle, it was the transition from driving to idle, e.g. when lifting off the gas pedal and pushing the clutch or putting the car in neutral. Generally that causes RPM to drop to 1000 (rolling idle) and then settle to idle when the car comes to a stop, but sometimes RPM would just drop to zero instead.

When the PCM knows it needs to idle, it can idle just fine. But the transition from driving to idling goes wrong sometimes. It seems like the software that's supposed to govern idle just doesn't kick in until it's too late.

Throttle cracker and throttle follower can both slow the drop in RPM, which helps, like it gives the PCM time to switch to idle mode. If there's too much cracker or follower then RPM falls annoyingly slowly, but I think I've found the Goldilocks zone now. One other advantage of throttle follower was that if RPM drops below and, and I tap the throttle to save it, the follower kicks in to ensure that RPM declines slowly and idle takes over. Without follower (or without enough follower) I'd often have to poke the throttle repeatedly before it would idle on its own again, which is super annoying when coming to a stop (I got lots of toe-heel practice though).

Adding a lot more airflow correction when RPM drops below the idle setpoint also helps it recover on its own when RPM drops below the setpoint. I had tried adding a little more than stock (2x more), but Darth_V8R's example used a LOT more (almost 10x). But for small deviations (within 250 RPM) I only use about 50% more added airflow than stock, and than helps prevent idle speed from oscillating a couple times before settling down.

Enabling "spark smoothing" seemed to help a lot as well, with it the "stall saver" stuff (like the extra airflow mentioned above) doesn't come into play nearly as often. But I still need to try disabling it again to confirm that it really did help. Figuring out cause-and-effect is kind of hard when chasing a behavior that's hard to trigger in the first place.

 

Timing is quicker to react than the IAC for throttle cracker/follower/saver. Make sure your idle area in the high/low octane tables and idle tables match in timing. I also bump the rows (efi live) below idle so it gives it a boost due to an underspeed.

 

Mine do not match at all

 

Have you tried it? Always been a key to transition from/to idle when I've tuned cammed cars.

 

I have. It made the car very sluggish off idle. When tuning a light clutch, you need lower idle timing and more supporting airflow. I have 16-18 degrees in my idle tables and 26-28 in my main tables same cells. My return to idle is better vs having them matched.

 

It seems to me that the idea behind cracker and follower is to get the IAC opened up while driving, so that when you lift off the pedal, RPM doesn't drop as fast it would otherwise. That allows time for the adaptive routines to take over before the engine stalls.

Early on, I reduced timing quite a bit in the idle tables. It was originally 28 degrees at 800 RPM (any load) but lowering it into the high teens works much better (gives the timing compensations more room to add torque, I think). It's 20 now in the idle tables (tapering above 0.3 g/cyl) and 25 in the main tables. How much timing do you run in those tables at 800 RPM, ddnspider?

I've been wondering whether the spark smoothing is mostly for the transition between the main spark table and the idle tables, and if that's the case, then increasing the idle timing to match the main tables in the ~2000 RPM area might make smoothing irrelevant. But I haven't been able to find any info about what the spark smoothing feature really does. I just got the idea to try it based on a post on another forum where a guy said he turned it off to get RPM to fall faster. I wanted it to fall slower, so I turned it on, and the stalling was gone. I still need to try turning it off to confirm that it wasn't just coincidence though.

 

I could see where that would be sluggish, did you try the other way though? More idle timing that matches the main tables.

 

Throttle follower and cracker are 2 different things.
Follower is how fast the iac closes. Cracker is added airflow over base airflow. Just like PE is to base fuel.

 

Some really good idle tutorials on hpt and efi live.

 

Please stop giving advice, lol. Follower is not how fast the iac closes (follower decay is), it's something used to add part throttle air, following (follower, get it?) your pedal input, set to decay at a certain (tunable) rate.

Cracker is airflow added when you let off the pedal, at a certain rate, for a certain time delay. Both are IAC - based, and both are tunable.

 

Both add airflow, and both have their own decay rates ("how fast the IAC closes") so I'm not sure what you're trying to say here.

 

Yes but then you cannot run as much adaptive spark. At idle, peak tq is at 33. If you run something like 28, you can only run six degrees of adaptive spark which is not enough to catch undershoot when you take moi off the crank. I mean you CAN run more, but if you add 12 degrees to 28, you LOSE torque but the controller does not know that so it undershoots worse. So by running 18 idle timing I can use 14 degrees of adaptive spark.

With a heavy clutch and flywheel you can get away with increased base spark and less adaptive. When you cut moi, you need the more aggressive adaptive spark.

 

This is correct.