ECMlink Shootout scramble help! Randomly going lean and dying

[trogador2005]

This is a shootout scramble ongoing troubleshoot. Would love some more input to try and speed this process up!

Here's what we've got:
2g Spyder GS. AWD and turbo engine swap from a 95 talon tsi

We have the 95 ECU running link v3 on SD
Omni 4 bar
Gm iat
Innovate wideband
Stock O2 for closed loop

Fic 1000 high impedance
Walbro 190
Stock fpr
T28
Delta regrind cams
We're actually using a 1g cas wired in like a 6 bolt in a 2g. (It was cause adjustable cam gears were used and didn't have the spot for the trigger plate since it's a 95 engine)

We used the 4g64 harness and added knock and changed how the fuel pump is wired (different for GS spiders) anyway.


The problem: never getting a CEL
The car was running perfectly for a couple months . Autocross events, wot, ide, cruise, no problems at all. then all of the sudden while sitting in traffic the car would bog and go lean, feels like fuel dropping out and back and out and back, it'll snap out of it for a second and go right back and die. Then it won't restart. When it goes lean the car fully fuel cuts, like the old hitting the brick wall. But it's not in boost, it's 5mph in a parking lot or 30mph on the street. Let it sit for a while and it'll be fine again for a random amount of time it seems and then happen all over again.

Nothing changed in the tune, as far as link goes, it's not calling for any fuel changes. IAC is behaving normally, O2 sensor is functional. Map and iat are reading normal.

What we've tried:
Replacing the fuel pump (ran fine til it didn't cause it sat while we swapped the pump)
Testing injectors (applying 12v to the individual injectors to hear them energize)
Swapping out for (no change)
Lock in open loop (no change)
Swapping ECUs (no change)

What we're going to try next:
Testing the wiring to the injectors
Testing wiring to the CAS

Thanks for helping!

 

[1990TSIAWDTALON]

Fuel filter itself? Just didn't see it in the list.

 

[trogador2005]

Fuel filter itself? Just didn't see it in the list.

Thanks! We do plan to check that today, I'm also going to put a fuel gauge post filter and pre rail to see if anything is going on there.

 

[CanadianTalonTC]

Got a log while it does it?

 

[trogador2005]

Forgot to hit capture on any of them last night. I'll get a fresh one tonight!

 

[trogador2005]

Update:

new fuel filter (no change)
Can't even start the car now, no injector setting changes are affecting the car starting

Checked continuity between all injector wires. Even turned key on and grounded the trigger wires on the ECU harness and they all fire.

got the car to start and idle with adding a lot of deadtime. I have a few logs. it seemed to be ok for about 2 minutes, then just nasty idle and bucking and lean.

 

[trogador2005]

Forgot to mention the fuel pressure is rock solid.

For kicks we swapped out the PTU, with no change. It does not feel like spark at all but it was easy enough to swap.

 

[steve]

To start, since you have V3 full, make sure you have the latest firmware. It looks like yours is missing some options.

There are some wild AFRatioEst swings in your Cam drive idle log and the WB spends a lot of time pinned rich 7.3:1.

Do you have a spare MAP sensor?

 

[trogador2005]

To start, since you have V3 full, make sure you have the latest firmware. It looks like yours is missing some options.

There are some wild AFRatioEst swings in your Cam drive idle log and the WB spends a lot of time pinned rich 7.3:1.

Do you have a spare MAP sensor?

I will definitely update the firmware!

I do see that now yeah. I will look for a spare map sensor, I think I do have one.

What drives AFRatioEst? Is it mostly input from the map sensor?

Thanks is for you help with this.

 

[steve]

I will definitely update the firmware!

I do see that now yeah. I will look for a spare map sensor, I think I do have one.

What drives AFRatioEst? Is it mostly input from the map sensor?

I'm far from an expert but this is a good place to start.

ecmlink101initialchecks [ECMTuning - wiki]

www.ecmtuning.com

I think on a SD car MAP sensor, IAT sensor and VE table combine to give air mass from which ECU computes fuel based on what stoichiometric ratio for the fuel and adjusts using the target fuel map and any enrichment. So it's the estimated AFR from the computed airflow and injector pulsewidth. I would hope it's close to both target and WB output most of the time.

 

[CanadianTalonTC]

Check the connection to your omni 4 bar or swap it out. At each instance where it goes mental your OMNI is reading all of the sudden -24" at idle from your normal -9 to 10. During these spikes it's going to compensate by dropping fuel flow as it thinks there is WAY less 02 molecules being pulled in. Based on the idle switch and inj duty it does not look like you are coasting during these periods and it also did it at 0kmh near the end of the log.

Your VE table is also abrupt and is calling for the same airflow across many cells in your idle and cruise area as it bounces around there. I don't think this is your main issue, but it probably is not helping.

 

[trogador2005]

Thanks guys! We'll be swapping that out and will report back!

 

[Stapl3]

I agree completely with Joel's post a couple up.

The SD table is doing you zero favors. At idle the cell used is a value of 62, with it sometimes using 48 below it, but from what I remember with LinkTools, you can assume 4 cells are being used at any given time at varying percentages of use. So to the left of cell value 62 and 48 you have a value of 35. It's basically a mess and the ECU is all over the place juggling that tiny window of idle. Your AFRatioEst is a disaster because of it. You need to load in a 'stock' SD table.

And as said above, check the MAP connection.

 

[1990TSIAWDTALON]

With speed density the VE table will read all 4 boxes and try to average them. A smooth transition to any other cell is what is needed and I haven't seen the logs yet. Track your VE table with your log and smooth any major changes around the boxes you are in.

If you were in Closed Loop, you can use SDAdjust to get you close to what the car was wanting by RIGHT clicking on the log and do SDAjust/CombinedFT. I am doing the same ATM (but it's too dam hot for more than ONE drive). Save your log BEFORE you make any changes so you can revert back if the car hates on you.

The settings I think I mentioned are what I have to use with FIC 2150s and they also require a lot of negative dead time, in some instances.

 

[Stapl3]

The settings I think I mentioned are what I have to use with FIC 2150s and they also require a lot of negative dead time, in some ininstances.

Tom and Dave made a testing station and concluded this deadtime for 2150s:

7v -> 2295
9v -> 1421
12v -> 939
14v -> 749
16v -> 555
19v -> 495

That is the all end all values for InjBatteryAdj DA. Retain that shape on the graph and then only do a global adjustment via the fuel tab.

 

[jed344]

Tom and Dave made a testing station and concluded this deadtime for 2150s:

7v -> 2295
9v -> 1421
12v -> 939
14v -> 749
16v -> 555
19v -> 495

That is the all end all values for InjBatteryAdj DA. Retain that shape on the graph and then only do a global adjustment via the fuel tab.

So I do not have a sheet for 2150cc in front of me right now. I did quick bring up some logs from cars I have tuned and the dead time values there are very close. I have sometimes had to modify them but usually not more then 50 either way.

The global fuel looks strange. I quick checked 3 2150cc cars I tuned recently and they all are around -58% to -62% on e85 to get VE to 100%.

The SD table does not look great. I am surprised that the global you are using then was able to hit target air fuel at around 100 VE. That would usually mean to me a fuel supply issue. The air fuel ratio estimate I honestly do not know how it calculates that when in closed loop. Once the car is in open look aka wide open it just goes to the number you have in direct access as the target. Those tables are all factory on your car.

Makes it easy to dial in a SD table as you can look at the wide band % off from target then do a quick math equation and change the values and hit dead on next time at wide open. You do have a LOT of dead time though as the dead time number is being added on top of the InjBatteryAdj in DA.

 

[Stapl3]

I don't know how AFRatioEST calculates either, but I know it's very damn accurate to the ECU's final expectation. It also doesn't care about deadtime/global fuel adjustment.

It's ironic how the ECU is mostly interested in the airflow values yet that's the most molested table - or mafcomp sliders. OPs car is using 62,48, and 35 at the same time. Wew. The entire SD table should be butter smooth based off 'stock' and tapered to a value of 100 at peak VE.

 

[jed344]

I don't know how AFRatioEST calculates either, but I know it's very damn accurate to the ECU's final expectation. It also doesn't care about deadtime/global fuel adjustment.

It's ironic how the ECU is mostly interested in the airflow values yet that's the most molested table - or mafcomp sliders. OPs car is using 62,48, and 35 at the same time. Wew. The entire SD table should be butter smooth based off 'stock' and tapered to a value of 100 at peak VE.

Yes, and when injector global is correct it will hit target AF at 100 VE. He is at 103 VE peak i saw so close and if it is hitting the stock air fuel target then i would assume there is a fuel delivery issue as i have never seen a 2150cc car on e85 need that rich of a global with good fuel supply. Many times i have cleaned up low throttle driving issues due to table having massive jumps. Sure to technically hit the target air fuel those cells maybe need that but air does not work that way as it comes into the motor. You smooth them out and its much happier.

 

[Stapl3]

Yes, and when injector global is correct it will hit target AF at 100 VE. He is at 103 VE peak i saw so close and if it is hitting the stock air fuel target then i would assume there is a fuel delivery issue as i have never seen a 2150cc car on e85 need that rich of a global with good fuel supply. Many times i have cleaned up low throttle driving issues due to table having massive jumps. Sure to technically hit the target air fuel those cells maybe need that but air does not work that way as it comes into the motor. You smooth them out and its much happier.

Is 4500rpm acceptable as peak VE on a rather stock car? I always worked off of 5500.

Not sure what I'm missing but OP is using 1000cc injectors, I'm not even sure where the 2150 talk came up.

 

[jed344]

Is 4500rpm acceptable as peak VE on a rather stock car? I always worked off of 5500.

Not sure what I'm missing but OP is using 1000cc injectors, I'm not even sure where the 2150 talk came up.

Oh my mistake, I thought it was 2150, ignore my comments on injector and global settings. Yes 5-5500 is the normal range for it.

I have seen some have a strange thing where it peaks VE much sooner even though it is not real. Kevin at Six Sigma had a write up about it on his website back in the day. I also messed with the tune after he did where VE was peaking to early and ended up with same results. Think he came to conclusion it had to do with some cars fuel set ups causing fuel pulses or something.

The cars that do it also end up with the lb/min being calculated too low. As I recall that car only logged around 42lb/min yet made almost 550whp on the dyno. Normally when VE peaks as it should then the lb/min has been very close to me vs whp.

 

[trogador2005]

I am happy to report that the new map sensor fixed the issue! Thank so much for your input everyone!

A full retune and it's ready for the shootout!