fuel trims causing a VERY lean condition.

[rumblebox]

A few months ago i posted about a weird stumbling issue i was having with the car. since i had gotten my VE's within 2% across the board(using LTFT's), i ruled it out as a lean condition and assumed it was an ignition problem.

well, now i know for sure that it's running lean as fuck regardless of what i do with the VE's. it goes even leaner when i plug in the MAF, to the point of being barely drivable from popping and stumbling.

i figured this out by manually commanding the AFR with the scan tool. when i set it richer than 11.5:1, the problem goes away and runs exactly like it should.... strong, no stink, no popping or stammering. it took me so long to figure out, because i had only associated bad exhaust fumes with a rich condition and blamed it on my previous exhaust leaks causing it to add fuel.

so i simply added 10% across my whole VE table(which i had previously tuned within 2% trim). it ran great for a minute until it started trimming about 10-15%, returning to the same popping and stammering shenanigans.

being a fairly inexperienced tuner, a few possibilities come to mind:

1. my injectors are a different size than advertized, causing my IFR calculations to be WAAAY off... but wouldn't the fuel trims still correct for this?

2. my engine just likes to run rich? unlikely, being N/A at about 10.2:1 comp.
LQ4, .040 gaskets, 823 heads, TR-55's

3. my MAP sensor is malfunctioning..... i say this, because it started throwing a code last week.... even though it still has consistent readings.

if you're wondering why i'm not using a wideband, i have an LC-1 and spent 3 months trying to get it to work. despite pages of help from experts on several forums, i never got it to work with EFI live using serial or analog connections. i even tested the analog output with my multimeter, and the voltage was right where it should be. so now it's sitting in a box.

[joecar]

Post:
- tune file,
- log file,
- calc_pids.txt file.


How is your LC-1 wired up (analog side, and serial side)...?

[joecar]

Look at table B3605 in the tunetool, in the upper right what units does it say...?

[joecar]

1. yes, trims will be able to correct, upto a point.

2. something is going wrong, the AFR you believe is being commanded is not...


What fuel do you use (E00, E10, E85)...?



If the trims start trending positive, this is an indication of a possible air leak.

[rumblebox]

1. yes, trims will be able to correct, upto a point.

2. something is going wrong, the AFR you believe is being commanded is not...


What fuel do you use (E00, E10, E85)...?



If the trims start trending positive, this is an indication of a possible air leak.

the intake gaskets are good. the exhaust is pretty good now. good enough to where it shouldn't add fuel from leaks

when i command it to 11:1, i assume it's actually closer to like 13-14:1

it thinks it's correcting with trims, when it's actually unnecessarily cutting fuel.

http://i.imgur.com/TDBRyWul.png

[rumblebox]

15891
15892
15893

the log file was before i added 10% to the VE's

[joecar]

Ok, if you have no wideband, then edit your calc_pids.txt as follows:

in CLC-00-305 and CLC-00-315 use Notepad.exe to replace the term {CALC.SELBEN} with {CALC.LTFTBEN}

( when you get your wideband working, you will have to undo that change )


It seems to me that the problem is that you have no wideband, and when the PCM commands other than stoich (CALC.CL becomes zero), CALC.SELBEN tries to use CALC.WO2BEN (which is pegged at 1.7 since wideband is not present), this is messing up your VE/MAF tables.

[joecar]

LC-1 serial comms:
- do you have the terminator plug all the way into the SERIAL IN jack...?
- do you have a null modem adapter (not just a plain gender bender)...?

[rumblebox]

Oh hell. I'll try that now, make a log, and get back with you.

I made a serial cable first. I'm pretty sure i had the terminator plugged in all the way. is it supposed to be difficult to press in all the way? cuz right now it's as easy as a headphone jack.

[joecar]

I had to put a dab of dielectric grease on my terminator plug and then it snapped all the way home.


With your serial cable, did you try swapping the Rx and Tx signal wires to see which way works...?


( I'm sorry if I don't remember all the details from a few months ago )

[rumblebox]

once i get the LC-1 installed into the new car (old one got smashed in a parking lot. so i bought a V6 car for $1400 and swapped everything over.), i'll check everything again and keep trying.

[rumblebox]

15894

here's a log i just recorded after correcting the calc_pid.txt file.

it's still cutting a lot of fuel and stumbling at low RPM's

[rumblebox]

Until i get it figured out, I'm just going to stay SD OL and command it to run richer. At least it'll be safer, and drivability should be better.

[gpr]

Have you tried replacing your front o2 sensors? could be running lean cause both sensors are not reading correctly. Thus the trims are close because they are reading wrong.

[rumblebox]

I switched one out to test that theory. They're good.

[swingtan]

My first impression here is that there is something going on that has not been factored in.

From reading the thread, the car runs fine in OL mode but runs lean in CL mode. This is not how the ECM is designed to work.... The ECM uses the O2 sensors to correct the fuel trims to run the ideal Lambda for the fuel used. So if the ECM is pulling fuel in CL mode, then the O2's are telling the ECM that the mixture is rich. No amount of playing with the tune settings will fix this, apart form running OL.

There is little in the log file to indicate anything, apart from the fact you you are moving the throttle way too much to be attempting to correct the fueling. When working on anything apart from dynamics, you want to work the throttle slowly. Another tip is to never use fuel trims for correcting fueling... Unless you have already set everything up and are doing some fine tuning, or have a very large amount of data (2 to 3 hours in a single log).

It's probably worth getting a new log file and doing some repeated tests to indicate exactly what the issue is. Place "notes" in the log (use the "F" key in BBL mode) to show when the issue is occurring. If the following is true...


Car is worse at idle and light throttle.
Fuel trims are consistently -ve.
Car seems better under some load (not WOT).
Fuel consumption is higher than expected.


I'd be checking the injectors to see if they are sticking / leaking. If you've tried to correct the fueling in the VE and the trims still keep pulling fuel, then something is letting in more fuel.

As has already been requested, post the tune that matches any log.

Simon.

[joecar]

I agree with Simon.

Question: could the O2 harnesses swapped between banks...?

[rumblebox]

At this point, i must have screwed up something last year when i was trying to make my wideband work, because the car ran great with the tune from my old setup. So I'm going to reinstall the whole program to default settings and try a re-flash.

[rumblebox]

Car is worse at idle and light throttle.
Fuel trims are consistently -ve.
Car seems better under some load (not WOT).
Fuel consumption is higher than expected.



Simon.

It idles fine in OL but lean-surges when it starts trimming.
It runs fine with no load but falls on its face with more than 25% throttle.
MPG's are atrocious.

[joecar]

Log shows that LTFT is consistently removing fuel as Simon said, it seem to be because VE table is too high...

apply LTFTBEN to the VE table (using a map of LTFTBEN x RPM x MAP) and see if the trims come closer to zero.

[picnic_george]

It runs great in SD, but as soon as you go back with the MAF it goes complete apesh&t?
That to me, means there is something wrong with the MAF or the scaling is off. If your MAF is off by 10% the car will run really really bad, ask me how I know lol

[joecar]

Get VE dialed in first, and then +1 do MAF.

[swingtan]

When running in SD, make sure you only fail the MAF in the tune, but leave it in place. Then when logging, ensure you log the MAF as well.

If you already have logs like this, set up a CALC PID in Scan tool to calculate MAF Air Flow from the SD tables using the following formula...



"{SAE.RPM}*{E38.APCYL_DMA}*.0667)"


If you are logging from scratch, ensure you add the GM.DYNAIR PID to the list.


Then in Scan Tool, look at the MAF measured airflow and compare it to the SD calculated air flow (CALC PID or GM.DYNAIR). This will show you if anything needs correcting.

Here's an example with measured MAF airflow and CYLAIR on the left, with their calculated equivalents on the right. (Note: this is for an E38, but the theory is the same.) In you case, SD should be the reference and you should adjust the MAF tables so that measured MAF airflow matches the calculated air flow.

15927

Note that transients don't work well in the calculations, so move the throttle slowly.

Simon.

[picnic_george]

I was just reading about your wideband problem. Are you using a null modem adapter on it?? That's all I used and my LC1 worked perfectly.
Think this is the one I got
http://www.ebay.com/itm/High-Quality-Mini-Null-Modem-DB9-Male-DB9-Male-Adapter-AD911NU-/400381200456?pt=LH_DefaultDomain_0&hash=item5d38944848


If you're using a straight through serial adapter it won't work
http://digital.ni.com/public.nsf/allkb/1EE0DD8AF67922FA86256F720071DECF

[rumblebox]

UPDATE:

Now I'm 98% sure it's a hardware problem.

I deleted everything except a few pid, tune, log, and map files and reinstalled V8 and V7.5 from scratch. Then i flashed a tune i knew worked very well before the issue started. Once it reached OT, it immediately started cutting between 15-32% of fuel. Ao now I'm replacing the MAP(and rewiring the extensions). If that doesn't work, I'll change out the O2's.

A few people mentioned making sure O2 harnesses didnt get switched, but i don't see how they possibly could have.

[swingtan]

post a current log file showing the warm up. Make sure you log ECT, STFT, LTFT, airflow and IBPW as a minimum.

[rumblebox]

UPDATE:

i loaded the tune i had from my old setup that patrick did for me in 2010.

it immediately started trimming between 20-32% fuel and popping, stammering, backfiring like before.

i ran that same tune in SD/OL, and the car runs great.

so now i assume my O2's are fried from running so dangerously lean for so long.

i'll install new O2's and post an update.

there's no MAP code with this tune either.

[rumblebox]

Here's the log with the good tune in SD/OL.

Keep in mind, this tune was for 317 heads, ported f.a.s.t. 90, 9.4:1 compression, 239/247 .624.624 114 cam. (i know...)

The current setup is L76/L99 top end, 10.2:1(.040 gaskets), ~226/231 .598.601 114 cam.

15976

[picnic_george]

Are your o2s switched to the incorrect bank? Your fuel trims will go haywaire if they are.

Why hasn't there been a wideband put on in OLSD to verify how far off the VE actually is is?

[rumblebox]

there's a 6 page thread where i couldn't get my LC-1 to work after 5 months of trying. It's sitting in a box until i have time to mess with it again.

A lot of people have mentioned swapped O2 harness, but i don't imagine how it could have possibly happened.

[picnic_george]

If your logworks are set up right and you're serial logging with a null modem adapter I'm unsure how it could be off. But I don't know sorry :(

Try switching the harnesses just to see if that's the issue. Those fuel trims are way off. When you are running OLSD, put your wideband in and see what it's showing even if it's not communicating with EFIlive. The car could be at 12:1 and running ok which is why your fuel trims are trying to pull 20%.

[swingtan]

Looking at the log, the engine is running rich in OLSD, at least as far as the NB O2's are reporting. Interestingly, there is very little change in the NB O2 voltage with throttle movements which makes me think that either the O2's are bad, or there is a wiring fault that is messing up the signal. Given both NB O2's are reporting similar signals, I'm thinking it's a wiring fault. Who did the wiring for the O2's? Is there any chance that the heater and signal wires got swapped around?

[rumblebox]

Looking at the log, the engine is running rich in OLSD, at least as far as the NB O2's are reporting. Interestingly, there is very little change in the NB O2 voltage with throttle movements which makes me think that either the O2's are bad, or there is a wiring fault that is messing up the signal. Given both NB O2's are reporting similar signals, I'm thinking it's a wiring fault. Who did the wiring for the O2's? Is there any chance that the heater and signal wires got swapped around?

Ive never modified the harness, but I'll see if i can swap harnesses before i buy 2 sensors.

[joecar]

I agree with swingtan, the NBO2 signals look wrong...

if the NBO2's were really reporting the rich level shown in the log, then you would be able to smell/see it in the exhaust.



15979

[SOMhaveit]

I see this same thing happening with one of my cars. At idle, my AFR is lean according to my wideband(s) and my trims are removing fuel (-20). B4105 (O2 Switch Point) all set to 450. Might the solution be to adjust the switch point at idle to match up with the wideband?

I'm about ready to forget about all of the MAFT, VE, & VET tuning, set my MAF values back to stock, disable VE, dial fueling in with the IFR, get the VE table close, & call it good.

[joecar]

You mean get the MAF close...?

[picnic_george]

How do you know the wideband is correct?

Your IFR and VE needs to be correct or it's going to be out of whack.

[SOMhaveit]

You mean get the MAF close...?

No. SSpdDmon had been playing with assuming the oem MAF values were correct and adjusting IFR to get AFR correct. He was just running pure MAF. I'd prefer to use both theMAF and VE table.

[SOMhaveit]

How do you know the wideband is correct?

Your IFR and VE needs to be correct or it's going to be out of whack.

I'll be making some comparo's here soon. I have a Wideband Commander, LM-2, and now i have a new TechEdge 2J9 & LSU 4.9 sensor. As far as being out of whack, GM uses the same injector in multiple vehicles and the IFR values vary. Also, I'd bet you could take a bone stock LS car and if you were to do VE tuning or MAF tuning you would get values different from factory values. Check the GM IFR values against the IFR spread sheet, they aren't using that value. Look at the injector voltage offset tables for the same injector in different models and they make changes to them. All of that tells me this is not an exact science and GM fudges the values somewhat to get where they want to go. I wish I had installed a wideband in one of my cars before I made any modifications so I could have seen just how close they are getting the parameters oem. Maybe someone else has done it.

[statesman]

Your IFR and VE needs to be correct or it's going to be out of whack.

And that's a real problem. Here's my stock Holden IFR table;

16143

I downloaded stock tunes from the same OS from a Camaro and a Corvette. They're both different to mine and they're different to each other. Okay, they might be using different size injectors, but that doesn't help me much.

I suspect that my IFR table is wrong but haven't been able to work out what to do with it. Without pulling the injectors and bench testing them, is there any way to dial in the IFR table without putting some trust in the accuracy of the MAF?

[joecar]

The Corvette and Camaro IFR tables differ by 1.5%... they're close:

1614516146

[joecar]

And that's a real problem. Here's my stock Holden IFR table;

16143

I downloaded stock tunes from the same OS from a Camaro and a Corvette. They're both different to mine and they're different to each other. Okay, they might be using different size injectors, but that doesn't help me much.

I suspect that my IFR table is wrong but haven't been able to work out what to do with it. Without pulling the injectors and bench testing them, is there any way to dial in the IFR table without putting some trust in the accuracy of the MAF?I don't know why the Holden IFR is kinked, but the VE table is built on top of that

( MANVAC 10-15 kPa corresponds to MAP 85-90 kPa)

[joecar]

Hmmm, I just looked at some Holden VE tables, and they're different in the other direction/axis also...

[joecar]

How do you know the wideband is correct?

Your IFR and VE needs to be correct or it's going to be out of whack.+1 on both.

[statesman]

I don't know why the Holden IFR is kinked, but the VE table is built on top of that

( MANVAC 10-15 kPa corresponds to MAP 85-90 kPa)

The VE is built to suit, but what happens when I go WOT?

I don't have a WBO2 sensor, and at this stage I don't really want one. I'm not looking for the ultimate performance... I just want a 'reasonable' tune. If I can get the IFR right, then PE shouldn't be too far off.

I could have the injectors bench tested, but I'm hoping for a simpler solution.

[swingtan]

The only way to get the "correct" IFR is to test the injectors and ensure you know what they flow.
The only way to then get the "correct" fueling, is to measure the exhaust AFR. Plugging in a theoretical IFR and a theoretical VE will not guarantee a good exhaust AFR.


The "simplest" solution is to get a WB O2 and dial in the VE to suit the injectors. If you don't want to buy one, see if a local shop will sell you some dyno time and use their WB. You need a known reference point for fueling, you can't guess it. Everything else in the tune then follows the fueling, spark, cranking, idle airflow.... If the fueling is off, everything is off.

Not using a WB means you will be relying on the stock mapping, which is not calibrated at all over say, 4000 RPM ( because the MAF has full control over this point ), or you use someone else's VE which may not be suited to your car. I've seen tunes downloaded off the 'Net that were completely wrong. the best examples were for European vehicles that had VE tables and fueling set for E25 fuel and then used normal Unleaded. The guys who loaded these tunes never thought about the differences in fuel.

If I was really, really stuck, and I mean "I had to replace all the injectors on the side of the road and they were nothing like the original flow rate", then I'd run CL and aim for a NB O2 reading of 950mV for PE. But I'd only do that if I had no other choice.

For the sake of $200 - $300, a WB O2 is good value....

[joecar]

Does Holden use a variable speed fuel pump...? I wonder how that effects the FPR and IFR.

[swingtan]

I'm pretty sure only a few Utes got a variable speed pump.

[joecar]

How does it interact with the FPR...? Is the FPR still able to meter a constant pressure (assuming unreferenced)...?


Unfortunately I don't have access to any Holden Service Manuals.

[swingtan]

I think the FPR is still the final control and in reality, the pressure at the injectors remains constant. I can't remember the rational for the variable pump but it's no longer used AFAIK.

[joecar]

Ok, that makes sense, thanks Simon.

[statesman]

You need a known reference point for fueling, you can't guess it. Everything else in the tune then follows the fueling, spark, cranking, idle airflow.... If the fueling is off, everything is off.


That's why I wanted a simple solution to sort out my IFR table first.

I found a simple solution... I looked for the part number on the injector, then searched that part number online. My injector is 12561462, which has a flow rate of 24.74lb @ 3bar. I can now use the formula which joecar posted on the thread titled "Calculating Injector Flow rate" to create a new IFR table. Okay, my injectors and fuel pump are old and worn which means my actual flow might be a little different to what's calculated... but at least I'll have a table that's fairly close.

[statesman]

The "simplest" solution is to get a WB O2 and dial in the VE to suit the injectors. If you don't want to buy one, see if a local shop will sell you some dyno time and use their WB.

That is a really good idea.



For the sake of $200 - $300, a WB O2 is good value....

The car is a daily drive, not a race car. I'll only tune the car once or twice a year. Once I get everything dialled in, the WB would just sit on a shelf collecting dust. It's not so much about the cost as it is about it being a waste of money for me. I much prefer your suggestion of buying some dyno time and using their WB O2.

[swingtan]

Fuel flow is a combination of injector orifice size, fuel pressure and fuel type. All 3 factors determine the maximum theoretical fuel delivery. It gives no indication of how much fuel is delivered during short pulses, but lets leave that for now. Assuming your injectors flow exactly what the spec sheet says (which is very unlikely), you still have no way of telling how efficient your engine is at pumping air, IE. how accurate the VE table is. Without knowing this, knowing the injector flow rate will still not give you an efficient fueling table.

You can use the LTFT's and the STFT's to dial in the part throttle areas, but as soon as you go into PE mode, you have no idea what the fueling is like.

Check out what a local shop might want for some dyno time and then decide what is best for you. It might be better to get a WB, use it for what you need and then sell it second hand to recover some costs. To be honest, you could probably just guess some figures and the car will feel fine to drive. But you'll never know if you are in a dangerous AFR mixture or if you just way off optimal and you may as well run figures. The WB will tell you when you are optimal, without it you simply don't know. Also remember that if the fueling is not optimal, you can't get spark timing optimal. You set the fueling, then move on to the spark timing and that will never work if the base fueling is off.