Starts, goes lean, then dies

[big_bertha]

TL;DR: after startup, in open loop, runs for a few seconds increasingly lean mixture to the point of nearly or totally dying. Standalone wideband and HEGO1 reading (in tunerpro) both confirm lean condition. Target lambse in tunerpro is rich. What constants/functions/tables are involved in final commanded fuel in open loop other than FN1362 (Base Fuel Table / WOT Multiplier)? Recently got worse, looking for values I can monitor in tunerpro to help diagnose.



351w with a 95 eec-iv MAF swap. Trying to diagnose a lean condition shortly after startup, confirmed on a WBO2 gauge. After cranking, engine will run for a few seconds with nominal AF on wideband, then slowly lean out to the point of dying - occasionally surging rich a few times first. This is happening in open loop - drives fine with good AF in closed loop but recently I can't get it to keep running at all as the starting lean out has gotten worse.

Monitoring in tunerpro with decipha_211230.adx, spk_lambse_calc rises to >1.20, LAMBSE1 goes from .48 to .68 - yet the engine sputters like its lean, and wideband shows >17 A/F. Now IIRC during open loop a number of inputs and corrections are disabled, E.G. HEGO, long and short term trim (?). Are there any values/functions/corrections applied to FN1362 that could cause the actual amount of injected fuel to be metered too low, but only in open loop? The injectors aren't dialed in but the fact that it has suddenly gotten worse after the last week has me wondering if some sensor is failing.

Base fuel table and example log attached, for anyone curious.

dying_after_start1.xdl

(8.85 KiB) Downloaded 322 times

[decipha]

sounds like your fuel pump is on its way out

tunes dont just go bad

[big_bertha]

Thanks for the response. Fuel pressure is good (adjustable regulator set at 45 PSI as per injector spec). Threw some parts at it, tried some adjustments, still no dice.

Turns out it isn't going lean, I pulled the plugs and they were all carbon fouled and soaked with gasoline. I'm guessing my wideband reads lean because the mixture is so rich that there's incomplete or no combustion (explains the thick white smoke that doesn't smell like coolant). But I think I found something interesting: in the attached log plots, engine fires up for a couple seconds, then dies (twice). You can see that as RPM (orange line) drops, MAFV (blue line) actually increases. Here are the RPM and MAFV values normalized to [0, 1]:

And here are the raw values, MAFV voltage roughly in the range of [.6v, 1.1v] while RPM decays from ~1k to 0. FUELPW1 (not shown) tracks the MAFV voltage so the ECU is definitely dumping excessive fuel as if it thinks its lean:

Does this look like a bad MAF signal, or are there circumstances where it makes sense for MAFV to go up while RPM drops?

[decipha]

fuelpw is irrelevant but is supposed to increase as rpm drops

but if mafv is increasing as rpms are dropping then that means the engine is loading, if your not on the throttle then the maf is probably bad

simply unplug the maf

[big_bertha]

With clean plugs the engine would start and idle (very rich, as expected) with MAF unplugged, so I swapped out for a new MAF sensor, no improvement. I did notice something else about the signal now (see attachment), this time MAFV curve does not track the shape of the RPM curve. MAFV stays flat while RPM rises and drops sharply - would expect both to be roughly the same shape, so maybe there's a bad connection between MAF and ECU? Also, MAF+ pin at connector reads 5.8V with sensor unplugged, not sure if that's far enough out of spec to cause a problem (should be 5V as I understand).

At this point I have done the following:
New MAF sensor
New ignition module
Swapped in old but working dizzy
new IAT sensor
Swapped in old but working coil (in case the problem was weak spark)
Verified MAF connector power voltage, input voltage, and ground
Verified fuel pressure during cranking (set at ~45PSI as recommended for the trickflow injectors)


Unless I'm missing something it looks like ECT sensor and ECU are the only components I haven't tried replacing yet, but this would be an odd failure for an ECU...However if the 5.8V MAF+ voltage is the problem then that would implicate the ECU...

In attached image, orange line is RPM, blue line is IMAF, note how IMAF levels out while RPM rises and falls.

[decipha]

if its rich then lean it out

simply go to the startup enrichment table and 0 it out for now so u can see whats going on

you do have your crank pw adjusted correctly for your engine and injector size correct?

[big_bertha]

After waiting for warmer weather I tried it at 50% enrichment and 0% enrichment. Only difference is now the car does not fire up to ~1800rpm on initial start, otherwise symptoms are the same, cranks and kinda sorta tries to fire up but dies getting progressively worse as the plugs quickly flood and foul out.

The scaling gave me trouble when making injector slope adjustments in tunerpro so I gave up temporarily, but as I mentioned I drove ~60miles on this tune before this issue started, so it must be some component failure, bad contact, or bad ground, right? The MAF is matched to the 36# injectors.

I just noticed something in the logs - every single one of my logs starts with a TP_REL_CALC value of around -63, slowly rising to 0 over ~20 seconds. RATCH seems to reset to 275 with every power cycle and drop slowly. Is this a bad/out of adjustment TPS, or does it look normal? Could a negative TP_REL_CALC throw off a table lookup or cause some other undefined behavior? RATCH in blue, TP in green, TP_REL_CALC in purple, you can see for about the first 20 seconds RATCH is higher than TP, and again same pattern just before 100 sec.

[decipha]

if the plugs are fouked then its rich

did you reduce the crank pw at the temp where your fouling plugs?

sure u dont have any air leaks?