Understanding Fuel Trims & Using Them To Detect Leaks

[ms3blackmica]

A little tip for anyone trying to diagnose where they could be leaking (other than boost leak testing)..There is something very important to remember about an engine that operates in Vacuum as well as boost and that is the difference between the two.

Vac. Pressure is defined as the difference between Atmospheric Pressure and Absolute Pressure denoted as:

Vac. Pr. = (Atm. Pr.) - (Abs. Pr.)

Boost pressure is defined as the amount of pressure above Atmospheric Pressure.

So your trims have shifted very far and you are concerned about a leak. How do you diagnose?

Remember that a (-) fuel trim denotes that the ECU is removing fuel from the stream and a (+) fuel trim indicates that fuel is being added to the stream. Now lets think about the operation of the engine during heavy vacuum (i.e. idle) and under boost (i.e. WOT). The ECU reads airflow via the MAF sensor which is the first portion of the intake tract. The turbo inlet side undergoes a suction pressure. If you are pulling in un-metered air due to your TIP not being on fully, you will see very (+) fuel trims due to the fact that the ECU did not measure that air and now needs to add fuel to compensate. Very (+) fuel trims can also occur if you have a vacuum leak (for example a cracked intake manifold or gasket) due to unmetered air being sucked in. If you have a leak from the turbo outlet up to the throttle body, you will see very (-) fuel trims due to the fact that the turbo discharges air and will be leaked out to the atmosphere and not taken in by the engine. The ECU will have to pull fuel out due to metered air being lost.

At WOT, the same analysis exists but with 2 exceptions. 1 being that there are no trims at WOT so you will have to look at AFR. If unmetered air is being sucked in, you can expect leaner AFR's. If metered air is escaping, you can expect richer AFR's. The 2nd exception is that at WOT, boost pressure exists but vacuum does not. Therefore, any vacuum leaks from the intake manifold will now become a boost leak and air will have escaped causing richer AFR's.



***Notes: Looking for Richer/Leaner AFR's at WOT may be difficult in a 2nd Gen MS3 due to a newer logic that can correct AFR's at WOT. Diagnose your intake tract very well.

I hope this helps clear up some of the confusion I have been reading from new members.



As always, post any questions and I will be more than happy to assist.

[ASHMS3]

Another good thread man.

[TiGraySpeed6]

Excellent post!
Distilled down to the basics and in clear language

This post will be most helpful for a lot of folks

[Lex]

Here's another way to diagnose leaks and where they come from:

Prepare the car and map for diagnostics

1. Start with a known good MAF curve for your intake, make sure the MAF sensor is clean, make sure the O-ring on the MAF sensor is not broken or damaged. Also make sure you have no exhaust leaks - these throw off the WBO2 sensor!

Set your LTFT learning zones to the following using ATR:
A:5.70, B:12, C:20, D:30, E:40 F:150g/s

Remember that for this to work you need to start with a good MAF calibration so grab one from an OTS map or from the Stratified section here for the larger intakes:

MAF calibrations for bigger intakes

Also remember that Gen2 cars have WOT fuel trims that adjust for leaks or a poor MAF calibration and we can't log these. They are not LTFTs or STFTs but they can add or take away ~12% under WOT. This means that on a Gen2 your fueling/leak/MAF deviation has to be more than 12% off for it to show up as a deviation from the fuel targets.

2. Put in pump fuel and turn off any meth or auxiliary fueling.

3. Drive the car for 40-60 miles.

4. Look at your part throttle LTFTs and your WOT targets compared to what the tables are calling for. LTFTs that are +/- 10% or lower and WOT targets that are no more than 5% off probably just mean the MAF table needs a tweak and may not indicate a leak.

LEAK SCENARIOS:

If you find that you are running very rich under boost but under vacuum your LTFTs are in check (no more than +/- 10%) then you may have a boost leak between the throttle body and compressor outlet.

If you find that you are running very rich under boost and are also running lean (LTFTs greater than 10%) under high vacuum such as the idle region, then you may have a leak between the throttle body and engine head (intake manifold gasket leak?).

If you find that you are running lean under boost and LTFTs are very positive (+10%) then you either have the wrong MAF cal or you have a leak between the MAF sensor and turbo compressor where the turbo is drawing in air around the MAF sensor.

So now what - you suspect that you have a leak. Time to do a leak test:

Build a tester
How-To NATOR A Boost Leak Tester for $15
Perform test
http://www.mazdaspeedforum.org/forum/foru...st-leak-89918/

[MazdaBoy2.3]

@ms3blackmica; @Lex;

Thanks for the info. Us noobs appreciate it!

[Lex]

Ok updated my post with some more info.

[ms3blackmica]

Thank you to Lex for adding that very important information.

Much appreciated.

Noobs should read the OP and Lex's post in detail.

[shaneski]

Originally Posted by Lex Ok updated my post with some more info.

Love reading your posts, you put things in a scientific way but make it simple to understand.

[mrQQ]

hi,

correct me if I'm wrong, but.. at IDLE, under vacuum, no matter where the leak is, it will be positive trims. Even if the leak is after the turbo, it still is under vacuum, and will suck air in, and not blow it out.

Correct?

[ms3blackmica]

Originally Posted by mrQQ hi, correct me if I'm wrong, but.. at IDLE, under vacuum, no matter where the leak is, it will be positive trims. Even if the leak is after the turbo, it still is under vacuum, and will suck air in, and not blow it out. Correct?

Not True.

Because turbochargers act essentially as a positive displacement pump, a leak between the turbo exit and throttle body (at idle with throttle plate closed to preset idle opening) will result in air escaping from the system.

If the leak is subjected to strong vacuum pressure such as, for example: brake boost vacuum hose, manifold gasket leak, BPV vacuum line popping off, then your trims will be positive due to said air being unmetered.

Because the intake/tip is subjected to suction pressure from the compressor wheel, any air sucked in that does not pass through the MAF sensor will also result in positive trims due to the fact that it was again, unmetered. This can happen if your TIP is not on properly and the compressor sucks in air from the engine compartment and not the filter.

[mrQQ]

I'm sorry, but, uhm, still the question stands - what kind of leak would cause NEGATIVE trims at idle?

Because turbochargers act essentially as a positive displacement pump, a leak between the turbo exit and throttle body (at idle with throttle plate closed to preset idle opening) will result in air escaping from the system.

yes, but at IDLE it would still be SUCKING not BLOWING, or not?

[ms3blackmica]

Originally Posted by mrQQ I'm sorry, but, uhm, still the question stands - what kind of leak would cause NEGATIVE trims at idle?

At idle, losing air from the system would cause negative trims. Where this can happen is anywhere from the turbo's outlet up to the throttle body.

Some examples of such leaks are:

-Couplers on hot or cold side boost tubes slipping off.
-Hose clamps on hot or cold side boost tubes not tight.
-Cracked weld on intercooler.
-Boost tube torn.
-etc...

All these issues would allow for air to escape from the system and cause a negative trim. In plain english, the ECU measured this mass of air but then some air was lost requiring the need to remove some fuel from the initial calculation.

A leak after the throttle body cannot cause a negative trim at idle since the intake manifold is subjected to high vacuum pressure causing air to be sucked in.

As Lex said, a properly dialed in MAF curve will show this properly. If your MAF curve is off and you have very positive trims, air escaping may not actually throw your trims negative but actually just less positive. Always perform a boost leak test if you have any suspicions.

yes, but at IDLE it would still be SUCKING not BLOWING, or not?

Again, it depends on where the leak is occuring.

The turbo's compressor inlet will always be "sucking" air in. The turbo's outlet will always be "blowing" air out.

[mrQQ]

so you're saying, that even at idle the turbo is making enough pressure for it to actually be above atmospheric? this doesn't make sense, because idle always reads as less than atmospheric pressure, so any hole in the intake should still be drawing air in...

[TiGraySpeed6]

mrQQ, you're not taking into account the throttle plate and it's effects on the air flow.

TPP at idle creates positive pressure between the turbo and TP, and negative pressure in the IM where the MAP sensor resides. Consequently air can be pushed out of loose boost tubes etc, creating negative trims as metered air is lost.

After the throttle body air would again be sucked in and create positive trims.

You have three regions of intake volume, if you will, two that can create positive trims which are located 1) before the turbo and 2) after the throttle plate. The one region that can create negative trims is a positive pressure area between the turbo and the throttle plate.

Pre-turbo = + trims, air added or sucked in
turbo to TP = - trims, air lost or pushed out
post TP = + trims, air added or sucked in


Does that help?


.

[mrQQ]

Hm, true, didn't think about that. I guess I just didn't think that turbo at idle loads can cause enough of pressure..

Well, thanks!

[NJSPEED3]

good stuff guys.

[DSIT995]

[MENTION @ms3blackmica;

Usually when I perform a leak test I do it off the turbo inlet and I have always pulled off the intake.. Today I got some different fittings and decided to do a leak test right off the intake and found my MAF sensor was actually leaking air. I was only trying to get to 10psi and it was a very slow leak but that would seem to be problematic?

Trims seem ok, thats usually how ill check for leaks initially, I have however noticed a slight jump in STFT's right when I blip the throttle(+15) and then its back to normal -/+ 10 and LTFT's are all spot on -/+4

Anyone else have a leaky MAF sensor?

[DSIT995]

Here is a screenshot of what Im talking about.. I also just installed a MD stg2 OCC so Im not sure if that is a factor too.. Doesnt appear to be leaking..


And also Im hitting my AFR targets at WOT

[mrQQ]

Does everyone get +25 STFT when pumping the brakes?

[BrandonRKD]

So the only thing that can affect fuel trims is unmetered air entering or leaving the system? I have -7 ltft while cruising highway speed (under vacuum). I've done boost leak tests and have found nothing. I had a coupler blow off my fmic setup and this started happening. I did boost leak tests after replacing the coupler.

Also my car seems to have trouble starting.

[Agent_Orange]

Originally Posted by mrQQ Does everyone get +25 STFT when pumping the brakes?

Originally Posted by ms3blackmica If the leak is subjected to strong vacuum pressure such as, for example: brake boost vacuum hose, manifold gasket leak, BPV vacuum line popping off, then your trims will be positive due to said air being unmetered.

Reading > You?

Originally Posted by BrandonRKD So the only thing that can affect fuel trims is unmetered air entering or leaving the system? I have -7 ltft while cruising highway speed (under vacuum). I've done boost leak tests and have found nothing. I had a coupler blow off my fmic setup and this started happening. I did boost leak tests after replacing the coupler.

A poorly calibrated MAF curve will affect trims as well.

[TiGraySpeed6]

Lots of things can & will have an effect in the fuel trims- ambient temps, humidity, how you hold your tongue while driving etc. If you go see a significant change then you want to have a good look around- for example, if you trend -7 in that particular range, and it suddenly changes to -20 for no particular reason, then you'd want to poke around for a reason.

I consider anything within +/-8 to be acceptable, tho closer to zero is good.

[OverBoosted]

Reviving an oldie here but has anyone one had trims at idle go from extremely positive to extremely negative from one drive to the next. For instance my stft's at idle were +20 in the morning then at lunch were -20. the LTFT's slowly move back and forth depending on if its + or - at idle. All the breakpoints above idle are rock solid at +/- 3.

[alextheguineapig]

Originally Posted by OverBoosted Reviving an oldie here but has anyone one had trims at idle go from extremely positive to extremely negative from one drive to the next. For instance my stft's at idle were +20 in the morning then at lunch were -20. the LTFT's slowly move back and forth depending on if its + or - at idle. All the breakpoints above idle are rock solid at +/- 3.

I never had trims that extreme, but I would see fluctuations at idle from +7 to -7. I had a leak from the MAF o ring, due to the ring being destroyed. I replaced it and adjusted the position of my filter on my intake. I still get -4 at idle now, but I think I need a new MAF cal. I have a FMIC as well and tightening the clamps helped some what.

[Vansquish]

How long has the tune been on the car?

[OverBoosted]

tune was flashed a few days ago so about 100 miles. It was started doing it on the last map as well after it had been on for a lot longer. It all started after my FMIC install. I have done boost leak test twice and found nothing. I did have to take off the intake and attach to the TIP to do the test. I will check the MAF o-ring this evening. It just doesn't make sense to me why it goes + to -

[Vansquish]

Is the car fully warmed up when you're noticing this? It could be that the trims are fluctuating at idle because you're looking at them at different times in the warm-up cycle.

[OverBoosted]

Seems like it doesn't matter. Yesterday I went to lunch and the car was fully warmed up and the idle STFT was pretty solid at -20 then once home for the evening with car fully warm the idle trims were reading +20. I think there might be a leak somewhere but like I said I couldn't find it leak testing. I might try cleaning the MAF and checking for leaks around it. I need to alter my leak tester to attach to the cobb maf section.

Edit - also threw code p2188 this morning. Rich at idle.

[btstarcher]

So what could cause trims that begin very lean at WOT and then end up slightly rich?

[MD1032]

Fuel trims aren't "lean" or "rich", they're either positive or negative. AFR remains constant...that's how the system is designed to work. And I can't tell what you're seeing because I don't know what you're trying to say. So, please specify, are they positive, or negative, and are you looking at STFT or LTFT?

[btstarcher]

Sorry, the car was running rich at WOT. I was targeting 11 AFR, and until around 3800 RPM AFR was between 11.4 and 12. Then it stayed right at 11 until 5700 RPM, where it gradually tapered down to 10.5. LTFT and STFT were fine. It was only happening at WOT. I lubricated my BPV and it's fine now; guess it was sticking.

[MD1032]

OK. Glad that solved it. I don’t know why that would solve it unless you were partially or fully VTA, but whatever. I forgot to mention an exception to what I stated above. These cars are a little weird in that they won’t aggressively compensate for gross AFR deviations at WOT. If your MAF cal is off, for example, or you have a leak, the AFR’s will be off at WOT. I actually read on here somewhere that Gen2’s actually have a global LTFT at WOT that the user can’t see.

[btstarcher]

You know, I also had my valvecover VTA....and I finally connected it to the intake again. So now nothing is VTA.

[Rspeed3]

Originally Posted by OverBoosted Seems like it doesn't matter. Yesterday I went to lunch and the car was fully warmed up and the idle STFT was pretty solid at -20 then once home for the evening with car fully warm the idle trims were reading +20. I think there might be a leak somewhere but like I said I couldn't find it leak testing. I might try cleaning the MAF and checking for leaks around it. I need to alter my leak tester to attach to the cobb maf section. Edit - also threw code p2188 this morning. Rich at idle.

Sorry for quoting such an old post but did you ever figure this out? Im throwing the same code... P2188 and seeing -20 ltfs and -10 stfs at idle when i checked this evening after a drive.

[Vansquish]

Originally Posted by Rspeed3 Sorry for quoting such an old post but did you ever figure this out? Im throwing the same code... P2188 and seeing -20 ltfs and -10 stfs at idle when i checked this evening after a drive.

Have you changed any hard parts or fiddled with any clamps recently?

[Rspeed3]

Originally Posted by Vansquish Have you changed any hard parts or fiddled with any clamps recently?

I pulled the intercooler to do a compression test but I was throwing the code before that. I actually cleaned the maf and replaced the purge valve the same time as the compression test and cleared the code. Threw the code again within 2 days.

I got the p2188 a few months back as well, cleaned the maf then and hadn't seen the code again until a couple weeks ago.

[OverBoosted]

Originally Posted by Rspeed3 Sorry for quoting such an old post but did you ever figure this out? Im throwing the same code... P2188 and seeing -20 ltfs and -10 stfs at idle when i checked this evening after a drive.

Never did figure out what was causing the issues. It went away for a few months and popped back on when I got to ENM 15' and I just cleared the code. I have since traded the car so no idea. I still assume I had a leak somewhere or maybe the purge valve.

[mrQQ]

It won't help you, but I never figured it either. Everyone is clueless and just offers to change one part or another. I've change them all, even rebuilt the engine

[btstarcher]

Originally Posted by mrQQ It won't help you, but I never figured it either. Everyone is clueless and just offers to change one part or another. I've change them all, even rebuilt the engine

Just wondering, was the check valve on the brake booster line functioning?

Also have a question about something I can't seem to get my mind around. What would happen to fuel trims if the recirc hose was leaking? And/or possibly the BPV? My AFR at idle is quite rich, even with very negative STFT (LTFT hasn't settled yet, so it's still at 0). However under normal driving, targets seem to be met without much correction; at least the trims aren't pegged at +/-25. AFR also seemed to be rich at WOT but I didn't stay there long. Car is broken; the core support/shroud and radiator are held in place with zip ties and cords and duct tape. I drove it almost 10 miles from a friend's house to mine after repairing the injector harness. So the fuel rail, IM, and throttle body were all off. The problem with my fuel trims presented itself before that repair though, and right after it came back from the shop that fixed my block (helicoil for the oil feed line).

[speedfreak44]

Mine do the same from time to time. I am thinking it may just be idling related or temp or both. When cruising im usually around 1.5 max but at times of idling i can hit stft of 25 or -25. Very weird and only seems to happen in stop and go traffic with long stop periods.