Has anybody attacked the fueling issues at idle turbo cars with airmeters have after throttle lift? This is probably the cause for 99% of everybody's problems and I know it is on my car. Here is the order of events:

1) Low-medium throttle lightly pressurizes the intake system and has the turbo spinning
2) Driver lifts throttle at low rpm
3) Intake system depressurizes over the first 1-second or so after lifting the throttle and creates a huge, false airflow due to backflows through the airmeter
4) Car misfires and ends up EXTREMELY rich, with as much as 4ms of injector on time at closed-throttle on my car
5) Short-term fuel trim starts to yank out a bunch of fuel to keep it from being pegged rich over the next 1-2 seconds
6) Airmeter signal settles back to normal and now the STFT has the car extremely lean, so it takes another 2-3 seconds for the STFT to recover
7) About 6 seconds after leaning into the throttle at 1.5k, then lifting throttle and stabbing the clutch it is FINALLY back to running closed loop at stoich again and ~1.8ms injector time on my car

The easy conceptual fix to this is a simple airflow limit based on throttle position. If we could set an airmeter limit at what would generate an equivalent flow for 2.0ms injector time at idle speeds this would completely fix the problem. It could also really some of the wacky fueling seen during the same behavior described in 1-4 above but when at a roll out of the throttle condition at 2000rpm or higher in-gear.

So my question is, has anybody looked into this? I haven't the slightest clue how difficult this would be to incorporate from a ecu code point of view. I know on the old DSM's they had an 'airmeter reset' function where the ECU would set out a fixed signal after a throttle lift. Based on logs in my EVO, it appears mitsu did away with this function, but potentially without another fix.

If these cars use something other than the airmeter to determine fueling at idle then this point is completely moot. My understanding is they use the airmeter exclusively for fueling everywhere, so somebody please correct me if that is a misinterpretation.

Thanks,
Kevin

 

Subscribed. Sounds like the same problem I'm having.

 

yea i have the same problem and so it frets me.

 

I thought the injectors are dead until the RPMs drop to a value just above idle?
I have noticed the back flow of the pressurized air through the air filter when lifting off under boost.

 

I wouldn't think that a KV MAF would register any airflow going backwards through the MAF because I see no way to generate a vortex in reverse airflow. Or are you talking about a vehicle with a hotwire conversion? Have you documented through logs that airflow is actually increasing on throttle lift?

 

I dont have airflow logs showing this, but I have seen the load logs peg out to over 120 under the described conditions.

 

Recirc or vta?

 

recirc mrfred, normal setup.
I have seen this on the Evo 6-7-8 and 9, though I think its worse on the 8-9.
I PM'd you some time ago, thinking it may have been some un-described table.
But the OP description matches my experience exactly. Nice write-up BTW OP.

 

Thanks Merlin.

I can't find the exact datalog described above, but this one gets the point across:



This is with the BOV recirculated, it would be even worse with the BOV vented to atmosphere.

It took it forever to recover to a 'clean' idle state in this log after the throttle blip. Injector pulse width only spiked up about an extra 1.5ms, so about 3.0ms compared to 1.5ms. As this is approximately 7.5:1 AFR, of course it misfires. For what its worth, I've tuned cars with closed-throttle maps. This is a map of injector pulse width vs rpm at 0% throttle, and it usually works out to be pretty darn flat injector on-time wise to make it run clean. Any car that is Alpha-N is throttle position based for the fueling and that works fine too.

Kevin

 

I'm trying to understand where this could be coming from. From what I can recall of my own vehicle experience, the only time I had significant idle issues was when my throttle butterfly shaft seals were toast. It would recover to idle very poorly on a regular basis, but I never logged it. After I replaced the seals, it idled beautifully again. I don't mean to sound like a broken record, from a theoretical perspective, I don't see how a KV maf is going to produce an airflow signal when air is flowing through it in reverse because there is nothing to produce a vortex. I suppose one possibility is that the air coming through in reverse is not smooth flow. This could be exacerbated by positioning/orientation of the recirc dump into the intake pipe if a non-stock intake pipe is being used.

If you're going to throw up some logs, I'd like to have those same items in the log above and also include load and ISCV step position.

 

my evo 9 did it with a haltech and motec ecu...not running maf

 

I think I'm also having the same problem. More noticeable when cold outside

 

I've said it before, but it really seems like there is secondary fuel control for idle conditions.

I have closed loop disabled and when I had the MAF it still did odd stuff like that.

Even on speed density, I get a very noticeable jump in AFR when there is no apparent change in ANYTHING else other then AFR and IPW. Load is the same, no change in RPM, nothing. My fuel map is smooth as are my VE maps. It is just like somebody threw a "run like ****" switch and the car jumps to 20:1 AFRs for no apparent reason.

 

I agree that it could be an ISCV issue as well.