Fuel trim for driveability

Albin from Leavenworth Diagnostician Posted   Latest   Edited  

Many years ago I was introduced to graphing scan tools and to the use of fuel trim data to get a direction on driveability problems. I have also taught fuel trim classes many times over the years, but I still see people struggle for a direction when it comes to engines not performing as they were designed to perform. Take for instance a vehicle with a poor power complaint, or even a misfire at idle, under load or at cruise.

Whether there is a check engine light glowing or not, don't get too worried about not being able to gather a ton of information with your scan tool. Fuel trim data is way cool, since it tells you what the PCM is having to do to keep the air fuel ratio where the PCM wants it to be. All this trim information comes from the oxygen sensors, both front and rear, and is real time data that depends on the engine load, speed and throttle opening to be able to snoop out many problems. Looking back over my learning experiences, Mac Vandenbrink and; Jim Linder were the two guys that got me started testing things dynamicaly. Any time you use fuel trim to analyze a problem, you are testing for the problem dynamicaly, with the vehicle in its normal working conditions, and the ultimate goal is to get the vehicle to tell you what and where the problem is. Stop and ponder the value in getting the vehicle to tell YOU where the problem is, without you haveing to get dirty, sweaty and skin your hands up, while poking and prodding around looking for problems.

The files I have uploaded are from a Subaru, I believe it was a 1999 or so vehicle. Yes, this vehicle is a little on the old side, but the basics of fuel trim are the same, whether it be a 1983 Chevrolet or a 2018 Subaru. The trims are still telling you what the PCM is having to do to keep the air fuel ratio where the PCM wants it to be. Yes, today there a few different ways the PCM is doing it job, but the basics are all the same.

Many times, the easiest way to explain a concept is with an example. This example is the 99 Subaru vehicle, which is using a MAF sensor in the air intake duct. Some of these vehicles would use either a MAF, or a MAP, so be aware of how the air/fuel management system, since going down the wrong trail can be time wasted and embarasing. In this case, there was no check engine light, only a complaint of "the engine speed will surge at times when sitting at a stop light, or when pulling out of the garage", as said by the lady that owned and operated this vehicle.

When I started on the diagnostic problem, my first test would be fuel trim data. You might ask why? Well, I wanted to see how the fuel control was working, since we are dealing with a power complaint, in this case, there was a little too much power when the power was not being asked for. (the gas pedal was not being pushed).

When selecting a scan tool, I want a tool where I can select a custom pid list to capture the data. This does two things, it makes the data rate quicker for the scan tool, and it makes it easier and quicker to put the captured data into a format that can be easily understood. Data that is not easy to understand is of little value to me. I'm a visual kind of guy, so I want to see pictures. The pids I need are; TPS, RPM, long term and short term trim for all banks, oxygen sensor voltage for all sensors and the load pid. With these few pids, problems can be found such as; low fuel pressure, restricted exhaust, cam timing problems, poor operating oxygen sensors, the cause of a misfiring cylinder, misreporting MAF & MAP sensors and the list could go on from there.

These data captures were taken several years ago, and I did not include the rear sensor voltage, or the loop pid, but they will still work to explain the concept of the FRTD (flat rate test drive). To get the vehicle to divulge the needed information and wave the red flag to show you where the problem is, you need to put the vehicle in the correct working conditions. These conditions are; start with the engine at idle, accelerate to wide open throttle through at least one shift point, then cruise for a short distance. This will allow the data for a good VE test, by using either the load pid, or a VE calculator, and will allow the system to operate at a cruise condition and then a deceleration. Here is the example data. process​.​filestackapi​.​com/resize=h:1000…

In this data, the trims are not looking good, the oxygen sensor voltage isn't looking all that good, so my question is why? Is the problem a faulty oxygen sensor, low fuel pressure/volume, an engine breathing problem, or a MAF that is misreporting? By examining the captured data, all these questions can be answered .

Lets take a closer look at the data, and tear it apart a little to see what is is saying. process​.​filestackapi​.​com/resize=h:1000… Starting with the idle, I noticed the idle speed is close to 1000 RPM. What can cause this? Is it the IAC stuck open, or a fuel/air ratio problem? Take a look at the oxygen sensor voltage. I noticed it is above .9V at idle. Why? Is the oxygen sensor not reporting properly, or is the air/fuel ratio actually rich? Yes, I know, a quick look with an exhaust gas analyzer would answer these questions, but why spend the time with that, these answers can be had while still sitting in the front seat of the car. Take a look at the MAF volume. Many years ago I learned a rule of thumb, most engines will inhale 1 G/S of air @ 500 RPM. This engine is idling at 1000 RPM, it is a 2.5L engine, so I would expect to see about 5 G/S of air showing on the MAF pid, this one is showing 12.9 G/S of air. Is that enough fuel to cause a rich air/fuel mixture, as shown on the oxygen sensor?

Don't jump to conclusions quite yet, there is more. Go on to the light throttle portion of the data. process​.​filestackapi​.​com/resize=h:1000… Some things that jump out to me are; why the wide trim swings from 0 to -29%, how about the erratic oxygen sensor voltage swings? Lets move on to the WOT VE portion of the FRTD.

process​.​filestackapi​.​com/resize=h:1000… I only want to look at three pids on this portion of the capture, the load pid, the TPS position and the oxygen sensor voltage. I want to use the TPS to verify the throttle blade is wide open. I want to see the oxygen sensor voltage flat and at .9V or above and I want to see what the load pid is showing. Now a caution on the load pid, be aware of the pid, is it a calculated load calculation, or is it an absolute load calculation? I will write more on these later. In this case, Subaru uses the calculated load calculation, which takes most of its information from the MAF calculation.

With all the data collected and analyzed, I first looked at the high engine idle. The idle at 1000 rpm asks the question why? In this case, the engine is not running in fuel control at a very rich mixture. How rich? I don't know, since anything much over a Lambda =1 will show high voltage on the oxygen sensor, and since this is not in fuel control, (closed loop) the trims are of no help. The rich air fuel ratio is causing the high idle, lots of fuel and air.

On to the off idle data, the system goes to closed loop, (I can tell because the oxygen sensor voltage changes, and the fuel trims change) when they do change, they drop to -29, which is telling me there is too much fuel for the air that is being reported going through the engine. (If the MAF reports more air than is being pumped through the engine, there will be too much fuel) When it comes to the wide open throttle portion of the data, I can see the oxygen sensor voltage is high and flat, which tells me there is plenty of fuel being supplied to the engine to support its habit, but the load pid is a tad bit on the low side. The reason I say this is from experience. This family of Subaru vehicles will show 99 -100% on the load pid at wide open throttle and at a transmission shift point. This one only shows 90%. Yes, its only 10% away from 100%, but it is still a red flag.

So the red flags are, high MAF flow at idle, fuel trim swing from 0 to -29, both in long term and short term, and the load pid off by 10%. I have my 3 different arrows in the target, its time to pull the trigger on a new MAF.

The data shown in process​.​filestackapi​.​com/resize=h:1000… shows the problem is fixed. The nice thing about this diag on this vehicle, I didn't get my hands dirty, I made a quick accurate analysis of the problem and the total repair was very profitable.


Brandon from Reading



Thanks Albin, I’ve learned so Much from both you and Mac over the years


Jim from Southampton


Mobile Technician

Thanks Albin, As ALWAYS...…..GOOD STUFF. It was also beat into my head Dynamic ALWAYS over Static testing.


Thomas from Manassas Park



Great stuff Albin! I've not had the privilege of sitting in a Mac Vandenbrink class, but have watched an old TST video where he was teaching how to look at secondary ignition wave forms to find driveability issues or internal engine faults. Listening to how he could look at a wave form and see "X" is to high or to low or to short or to squiggly as a result of "X" was like " how the in the...... does he know that?"


Keith from Collinsville



Always a good insight Albin! Your thought process of thinking through your collected data is the part that most inexperienced technicians skip, or simply a skill they have yet to develop. I've always said only 2 qualities are needed to make a great diagnostician: critical thinking skills, and the drive to learn.


Robert from Ballston



Thanks Albin. As always an eye opening case study

Fuel trim is indeed a great first place to start in any drive ability diagnosis. Whatever happens in the engine will effect fuel trim. 

The one thing that has tripped me up though is to be sure The pcm is in closed loop fuel control. I’ve seen some instances where a misfire is occurring or a sensor is skewed enough that the fuel control enters open loop. My experience is that in open loop the pcm is controlling the fuel with previous data that the pcm considers reliable before the fault.

I am not sure if that is always the case or just my limited experience? 


Albin from Leavenworth



Robert, don't get stuck on one strategy when it comes to open loop. Most times, what happens in open loop is up to the manufacturer, and can change from platform to platform & model year to model year. Last week I was on a mobile call to a 2002 E450 with the V10 engine. The complaint was low power. It had already had two fuel injectors installed to fix the problem, but to no avail. With no DTC's stored, my first shot was to check the fuel trims. The engine was running in open loop drive, and the trims both long term and short term were pegged at 0 on both banks. This is normal for many Ford vehicles to go to open loop drive when a misfire happens.

This E450 was misfiring on all 5 cylinders on bank 2. In this case, I used my IDS to do a relative compression test to verify compression, Since I couldn't use the fuel trim to determine if the problem was fuel or spark, I then used my scope to current ramp the ignition coils. I found all 8 coils on bank 2 were misfiring, which was caused by all the plug wells filled with coolant, from a leak at the thermostat housing. There are times when I actually have to work to get to the bottom of a problem. :)


Robert from Ballston



I agree I do see different fuel trim and loop control strategies across year make models. 

Your Ford case is a prime example. I usually find if a Ford pid indicates FMEM yes then the fuel trims can not be used as it is running in open loop fail mode. 

My understanding is that if it is in ol/flt that would also indicate fuel trims can not be used as there is a fault that is affecting the engine control module’s ability to properly monitor and control fuel control. The Ford you mention running in ol/drv and the fuel trims being at 0 is a prime example. At least Ford moves the trims to 0 when it is not reporting actual fuel trims. I’ve had other vehicles running in ol/drv that are not using fuel trims for correction but are reporting a trim correction number. That is where it becomes confusing for me. 

In that case if I am not sure if it is actually using trims for fuel control I will try to skew the mixture (unplug a vacuum line or som similar) to see if the trim actually corrects. 

In spite of all the anomalies fuel trim pids are probably some of the most useful information supplied by the engine controller. 


Jim from Southampton


Mobile Technician

GREAT POINT, Albin. I was taught and have been teaching what I call the 85 / 15% factor. About 85% of what happens on vehicles , Happens on all vehicles (like Fuel Trims), but the 15% is what THAT PARTICULAR manufacturer does (again, like fuel trims).

So we all need to know the 85% of technology strategies inside out and backwards, But we need to know where to get the 15% for that particular vehicle we are working on right now.

This simple rule has helped and saved me thousands of hours in the diagnostic process. A LOT of the same things happen in most vehicles, fuel must be delivered to the combustion chamber, the Primary must make a Secondary, The cylinders must inhale and exhale, these are all the 85%. We all know there are A LOT of different ways the manufacture can and does achieve these events, That is the 15% that if I'm not working on only one brand of vehicle, I don't need to fill my mind with how that manufacturer fulfills this task, Only need to know where to fing that 15% information ( Which BTW, one of the ways would be right here at The Diagnostic Network).