Total Fuel Trim Calculations - What's Your Formula?

Scott from Claremont Manager Posted   Latest   Edited  
Discussion
Driveability
Emissions

As per Randy‘s request, I'm starting a thread on Total Fuel Trim and how it is calculated. I'd like to get multiple perspectives on this by hearing your thoughts.

It seems fair that if you were going to calculate total fuel trim you would simply add STFT and LTFT together, right? I posted the other day about a vehicle I was road testing as a follow up on a repair we performed. The graphed data presented (TFT), didn't line up with the 3D table (TP% x RPM) (Note: Table data is averaged). According to TIS, LTFT + STFT = Smoothed Fuel Trim.

So where are you on Total Fuel Trim?

+5

Chris from Commack

 

Diagnostician
 

I'm kinda with you on this. This is how I've based much of my diagnostic knowledge on the fact (myth?). I would love to learn how to interpret it differently.

0

Geoff from Lahaina

 

Diagnostician
 

My immediate thought is "semantics". If every single instructor calls it "total" and now an article you found in TIS calls it "smoothed" I'm not worrying about it.

0

Hans from Salt Lake City

 

Diagnostician
 

I've see it where it wants you to add the two together. But I did notice Randy said additive and multiplicative, which happens to be my favorite type of fuel trim (VW/Audi uses it). It gives me three numbers to work with. Current o2 sensor reading (standard STFT) and a LTFT at idle and part throttle. It gives me a lot better insight to the fueling instead of just one averaged LTFT number.

0

Albin from Leavenworth

 

Diagnostician
 

Total, means just that, the sum of all the fuel trims, which is long term and short term, but how about rear oxygen sensor trims? Not all manufacturers list this, but over the years, I have seen some rather heathy trim come from the rear sensors.

+3

Brin from Melbourne

 

Diagnostician
 

Albin, the way that you describe it is the way that I've come to understand it. STFT+LTFT+Rear FT. 

What terms and how many terms you'll see depends on what vehicle you're working on, what model year, whether you are using a aftermarket or O.E. scan tool, or whether you are monitoring global or enhanced data.

Regardless of what data there is, modern vehicle fuel control strategies use all sensors for fuel control, correction and adaption.

It feels like the wild west at times when trying to make sense out of what each manufacturer uses for pids or how they may change from one year to the next. Some of the changes are driven from Government regulations while others can be viewed as a simple difference in language. When we're working on several makes, we have to be careful of making assumptions that the vehicle we are working on has the same strategies that we are used to working with. We can take advantage of the somewhat standardization that we get with global but we also have to realize that the translation that is needed to make a system that works differently then how the global side displays the data, can make things more difficult for us. 

+3

Albin from Leavenworth

 

Diagnostician
 

Bryan, You make some great points. When working on a fuel trim issue, many times the generic side of OBD II will get the job done, and some times that is the place to go, although as you pointed out, this changes from year to year, model to model. This is one reason why "a quick look at fuel trims" can head a person down a rabbit hole. From time to time, it is a MUST that we study the data on our scan tools.

Several years ago, I remember a 1996 Saab 9000 with a fuel trim issue. The trims were commanding +49% total trim. This vehicle is a speed density turbocharged gas engine. Looking back, an exhaust gas analyzer would have nailed this problem real quick. The problem ended up being a very small air leak at the gasket on the rear oxygen sensor. I found the problem with a smoke machine.

This is a rather old vehicle, but my point is, the scan tool data (Tech 2) had nothing about fuel trim from the rear oxygen sensor. Generic data was and is also totally useless on this vehicle. Things are better today in issues like this, although many times we still do not know what was in the minds of the engineers that designed the diagnostic systems we get to work with.

+1

Tim from San Diego

 

Curriculum Developer
 

I've been formulating a similar question that is related. Does Rear Fuel Trim get added to Total Fuel Trim? If we count all three it might add up! I assume that each manufacturer actually does this, but that they vary in whether or not the values are added (or multiplied) in. I think this is what Albin was referring to, and it's something I'm eager to 'catch' the next time I see a downstream sensor fault or exhaust leak cause a fueling error. 

0

Brin from Melbourne

 

Diagnostician
 

Tim, I would have to assume that all manufacturers use fuel control strategies that use all feedback sensors for fuel control, adaption and correction. How they react to different kinds of faults depends on each manufacture's strategy for that vehicle. One single manufacturer can change the strategy from one year to the next in a way that the affect that that same fault, would cause a different result. This often is the result of changes in emission regulation standards. As diagnostic technicians, we need to study as often and hard as we can but we also need to just realize that we can't make decisions based off past experiences alone and we have to look at all of the puzzle pieces during each diagnostic scenario. 

+1

Randy from Denver

   

Analyst
   

So it says "Fuel trim is related to the feedback compensation value, not to the basic basic injection time." What does that mean?

If you track Toyota enhanced PID for TFT (total fuel trim) it doesn't track with ST + LT especially when a fault occurs. Even if you add rear fuel trim, it still doesn't track so how do you explain the way Toyota handles TFT?

Toyota did a presentation on their FFV last week and how they submitted to CARB their request for handling fuel trim DTCs. As you know, I am not allowed to directly quote anything from an SAE presentation so I won't give names and details. Sorry. I will say this gave me an idea, I decided to ask engineers I knew how their vehicles handled ST and LT. I asked 4 different engineers, I got 4 different answers so I think I can safely say they handle it differently.

Several years ago it was proposed to do away with fuel trim requirements in generic because, the engineers said, they don't do feedback that way any more and I agree.

Here's the math to get you started, it's what I used last week when asking my question.

Pretend the BPW is calculated to be 10 ms (just a simple example).

LTFT shows 20%. For now, STFT is 0 or if you like, don't include it for now. What is the final PW? Is it 12 ms? ((10 * .20)+10).

Now let's say STFT is 20% and LTFT is still 20%. What's the PW now? Is it 14 ms? BPW+LT+ST. Or is it 14.4 ms? BPW+LTST. Sorry about the errors but every time I try to add a () or a * together it doesn't display right so you will have to interpret what I am trying to show.

What about Bosch fuel control? It uses additive and multiplicative numbers in fuel trim. How do you calculate that TFT?

Now let's include rear O2 control, if it simply adds or subtracts fuel by percent, like most here seem to think, how does that factor into the formula above?

Anyone that's really studied rear O2 fuel control knows they don't use percentage for most modern systems (some do). So if they are looking at frequency or switch points, how the heck would that be displayed in percent and how would TFT help you?

Catalyst modeling is also included in fuel control and some use it after BPW is calculated so that is a feedback portion and is a percentage for some. Where is that in TFT?

+2

Scott from Claremont

 

Manager
 

Q1: “What does that mean?”

A1: It Depends® 😀

From my experience, injector on-time does not transfer linearly with the mass of fuel delivered. Once the BPW is established, there are other modifiers such as time (small PW), voltage, temperature, etc) applied (and more due to rear FT) that affect the final value.

In the service environment, we as the service technician look at fuel trim as an error in the desired cylinder AFR. Was the error caused by the wrong cylinder air measurement or is the fuel delivery system flawed, etc? We were taught that the % is the mass amount of fuel under or over the commanded EQ ratio and the ECM is adjusting the amount of fuel mass.

Next week I’ll do some logging of ST, LT, Enhanced TFT on a late model Toyota and report back with my observations. I started writing this up yesterday and since then, you dropped in a couple more useful papers to digest, thanks for that. Although I don’t work on vehicles all that much anymore, developing a deeper understanding of what’s really going on intrigues me.

Looking at this from a calibrators perspective, the cylinder air prediction or VE needs to be solved for. I’ve performed modeling on speed density with and without mass air systems and when doing so, we use a calibrated wide band sensor data piped into the scantool so that a math function, Base Efficiency Numerator (BEN factor) value could be constructed and used to apply a correction to the VE table or “Virtual VE table”. These numbers allow one to back into the desired AFR equaling actual AFR. We typically apply filters to disregard events such as rapid throttle changes, DFCO, etc. you can also establish a threshold of samples needed in order to start recording and averaging the data in each load cell. I’ve set up many of these system on race cars & hot rods and when you have everything numerically correct, engine performance is very nice. 

Now back to the subject.

One could also create a BEN from LTFT + STFT averaged ({SAE​.​LTFT}+{SAE​.​STFT})/200+1. The graph on my RAV4 shows the following in one operating condition (not averaged):

STFT: -2.344LTFT: 13.28TFT: 10.94

Running the BEN formula I listed above, I arrive at the following:

…= 1.0546 or 5.5% TFT or essentially ½ of TFT. I have no idea if this is a valid assumption to make but we did end up disregarding it (RAV4 case) as a service issue and put the vehicle back into service.

In regards to the proposal of going away from reporting FT, there are some (probably many now) using Neural Networks to model VE because of the complexity of varying valve timing, intake runner control etc, presents. These systems work way different than the basic tables (RPM vs. MAP) of yesterday and fuel trim numbers may not make any sense at all in these cases.

Reading through the Mazda paper you referenced in your reply to Robby, you can see what they did to achieve high converter efficiency under all operating conditions. Essentially, if we were to try and drill down on TFT and include post trim numbers as a technician, we would need to know a whole lot more than just trim as cat temp, flow rate, catalyst aging, and OSC all play a role in determining what is actually happening. This has been a great exercise thus far. I’m learning quite a bit and will continue as I digest more on my multiple passes through these SAE papers and of course listening into your wisdom. I (and I know many others in the repair community) deeply appreciate your knowledge and participation here and elsewhere. 

My understanding at this point is that the technician needs to how much weighting he or she needs to apply to fuel trim numbers when evaluating a problem. 

One more note I picked up where the Mazda paper states that the catalyst monitor test is “very easy” when tightly controlling POQ. Looking at the dates of these papers should inspire one to think what the engineers are up to today...

+2

Scott from Claremont

 

Manager
 

Randy,

Here are a few screen caps from a 17 Tacoma with the 3.5L Direct Injection V6. 

Idle FT Data

Cruise FT Data

Cruise TFT Data Graph w/ Front Lambda Sensors

Cruise STFT w/ Front Lambda Sensors

I honestly don't believe that I learned anything from looking at this data and I haven't had time to dig into SI for TFT explanations but if you'd like to drop a few clues that would be nice. Thanks

0

Randy from Denver

 

Analyst
 

A coworker gave me a great line today and I plan on using it so consider it copyright protected.😀

“When the football player crossed the goal line, did he score a safety or a touchdown?” That‘s an 8 point swing depending on the direction you are going. You switched directions on me. This Toyota is very different than the Toyota that started this discussion. You will notice the TFT pid has changed as well. Hmmmmm, I am really debating on what to do here. I’ll do this, I will have to do some digging before I can talk about this system so I will ask the group to jump in here and take over. I’m quite confident the fuel trim instructors on this network can explain the fuel control system on this vehicle. I will jump in if I can.

0

Scott from Claremont

 

Manager
 

But I did use “multiple perspectives” in my second sentence...

I‘ll see what I can do to bring in more input this subject.

0

Albin from Leavenworth

 

Diagnostician
 

I have been following this discussion closely and to say the least, I am close to overwhelmed with the strategies of these late model vehicles. After looking at scan data from several late model vehicles, and looking closely at the captures you just put up, I can see no way to reverse engineer these systems. I think Randy said that way back early in this discussion.

0

Jim from Frederick

 

Curriculum Developer
 

Scott,

Do you have access to this vehicle easily? If so, I would like to see the the A/F learning values both standard and dual and did you graph the TFT with the LTFT? 

The only thing I will say here is that the approach to this vehicle would be different based on design and available PIDs and Active Tests.

0

Scott from Claremont

 

Manager
 

Hi Jim, 

I plan on taking a deeper dive on this vehicle as soon as I’m able. I’ll update you accordingly. 

Thanks!

+1

Scott from Claremont

 

Manager
 

I just finished reading SAE Paper … (thanks Randy) which is titled “Development of PZEV Exhaust Emission Control System

I’ve re-read several sections multiple times and I learned quite a bit. One big takeaway for me (related to this discussion) was the sub-O2 sensor construction mods to deal with the abundance of Hydrogen exiting the close-coupled catalyst (not engine out) when there was a cylinder to cylinder AF balance. They didn’t go into depth on how the sub-feedback control worked but they had incredible results in NOx reduction. 

For anyone interested in learning more about what Toyota accomplished with this and all of the tweaks performed, Go to SAE​.​org and purchase the paper, it’s well worth the ~$19!

+3

Randy from Denver

 

Analyst
 

That is a great one, especially about the sub O2 hydrogen control. I always think of Jon Riggle when I read that part.:)

When you want to go deeper, look for sliding mode controler and more specifically, PRediction and Identification type Sliding Mode Control (PRISM). Make sure you understand POQ as it relates to OSC and then you are off into the land of understanding modern fuel control. If your head doesn't explode.

+1

Scott from Claremont

 

Manager
 

Yeah I bet Riggle gets a kick out of that. When he brought up that topic years ago I had no idea when he was asking about!

As for your other suggestion, I’m on it and thanks!

+1

Paul from Shelby Township

 

Engineer
 

Hi Scott,

Current fuel control system are pretty complicated. As an example, Ford uses feedback from the upstream UEGO sensor (A/F Sensor) and a PI controller to determine the feedback trim needed to achieve the desired A/F ratio in the engine exhaust. This is called the inner control loop. A "PI" controller has a proportional term (short term trim) and an integral term (long term trim). The short term corrections are learned into the long term corrections so you don't have to learn as much every time you start the engine.

There is also feedback from the downstream HEGO sensor (switching sensor). There is a proportional controller that perturbs the A/F ratio going into the catalyst and determines the target A/F ratio into the catalyst to achieve best catalyst efficiency. (You get better conversion efficiency if you oscillate around stoichiometry, not just sit there slightly rich or slightly lean.) Then there is an integral controller that that is used to learn a correction to the upstream UEGO stoich point. This system is called the outer loop. 

So can you added short term and long term fuel trim terms together to get total fuel trim? Yes, you can do that to get an rough idea of how far off the inner loop is. This however, does not account for the entire fuel control system which may may be biasing the outer loop rich or lean for catalyst efficiency or for UEGO sensor aging and perturbing it at the same time. Trying to turn fuel trims into a fuel pulsewidth would be really challenging.....

Note that there are all kind of other noise factors that affect fuel trims like ethanol concentration that will show up directly into the fuel trim numbers.

It's pretty hard to reverse engineer what is going on with the fuel control system. The controls have changed over the years as have the calibration methods. Things like PFI/DI systems make inner loop fuel trim learning even more complicated as you are trying to attribute fuel trim corrections to either the PFI or the DI system.

I hope this sheds a little light on the subject.

+3

Randy from Denver

 

Analyst
 

Paul,

Can you answer Rob's post, why did you show rear O2 fuel trim that way on those systems? That's not what is really happening.

Do you agree with me, we can't be using fuel trim the way we did 20 years ago?

As I have already said, we see these systems fail IM240 with no DTCs and they are extremely difficult to diagnose. The last one saw a 20% increase in FE after repair. I have another one that took 18 tests over multiple months before it was fixed (I was not involved). I'm still working on a diagnostic approach so any input is welcome.

+1

Paul from Shelby Township

 

Engineer
 

Hi Randy,

Are you asking about the switcher and follower graphs below? The switcher is the one I described in my post. The outer loop is modulating the A/F through the catalyst. I'm not sure what model year the follower is but as I said, the rear fuel trim control strategy has changed over the years. In the 2000s, the idea was to try to control the A/F into the catalyst using the inner loop control and then use the rear O2 sensor feedback to optimize the A/F out of the catalyst to maximize catalyst efficiency. Differences in upstream and downstream sensor readings were used to indicate a shift in the upstream sensor and was used to adjust the inner loop control. The rear fuel trim target bias was learned in very controlled "high confidence" zones (which were typically moderate cruise conditions, never at idle). Proportional control was used sometimes as well. 

It's hard to classify fuel control systems. The upstream sensor changed from a HEGO to a UEGO in about 2008 MY which fundamentally changed the controls. The controls kept changing over the last 35 years in order to get better catalyst efficiency and each calibration group could calibrate the same control system in different ways. Rear fuel trim has gone from a small adjustment in upstream setpoint to a major driver of A/F ratio. 

Note that things like exhaust leaks can affect the rear O2 reading and cause control system and emission issues. In other words, don't just replace the rear O2 when you get a DTC. 

Note that when you use rear fuel trim, the gains you need for optimum control are affected by catalyst oxygen storage. Aftermarket cats affect those gains.

As far as fuel trims being different after a KAM clear, I mentioned that ethanol learning is a source of variability in fuel trims, especially for control systems that are FFV capable. They are designed to relearn fuel trim after a KAM clear or after refueling. Clearing KAM can clear out such a correction,

If you have a specific question, I may or may not be able to answer it but it always helps to know the vehicle, engine and model year.

+4

Scott from Claremont

   

Manager
   

Hi Paul,

I believe he's referring to Robby's post within this thread where he's listed the YMME.

0

Randy from Denver

 

Analyst
 

We will probably have to discuss this over a beverage someday but we are on the same page. I am not disagreeing with anything you said, I was trying to help the readers with my rear fuel trim question. I have said there are multiple iterations of this control and, as you said, it is impossible to reverse engineer with all the underlying inputs to modern fuel control.

Between Rob and my uploads, we have given 3 different fuel system controls. Here is a Fourth. This is a 2014 Ford Escape with a GDI turbo 2.0l. There are no problems, no DTCs, mass emission results during this run are very clean (I have the results at work). 

I don’t believe you can use fuel trim in the same way we did years ago. Specifically, the rear o2 fuel trim as shown. I am saying Ford is not increasing and decreasing ipw based exactly on the rear o2 fuel trim percentage nor is the bpw increased/decreased based solely on the sum of ltft stft rearft on the vehicle in this response. The rear o2ft as displayed is not directly correlated to a feedback of the rear o2.

Do you agree? 

+2

Paul from Shelby Township

 

Engineer
 

Hi Randy,

Yes, I agree. Looking at STFT and LTFT is useful but adding in rear fuel trim is not. Rear fuel trim is not directly feeding back to the injector pulsewidth.

+2

Robby from New Market

 

Mobile Technician
 

Hi Paul!

"Note that when you use rear fuel trim, the gains you need for optimum control are affected by catalyst oxygen storage. Aftermarket cats affect those gains."

Is this an example of what you're talking about?

That is from a 2014 Ford F550 6.8L V10 that had a weird surge under load and setting a P0420. Come to find out, the Catalytic Converter had stopped up and the customer gutted the cat. RO2FT was switching from +8% to -6.5% and the EQRAT PIDs were swinging from about 0.9 to 1.06. I didn't record the EQRAT desired PIDs for some reason. After clearing KAM, those wide swings in EQRAT were no longer present and the surge was gone. This was only a temporary fix though. After several miles of driving, the surge would return along with those wide swings in EQRAT. Was that a response to no catalyst oxygen storage?

"As far as fuel trims being different after a KAM clear, I mentioned that ethanol learning is a source of variability in fuel trims, especially for control systems that are FFV capable. They are designed to relearn fuel trim after a KAM clear or after refueling. Clearing KAM can clear out such a correction"

On the 07 Expedition that had different trims before and after KAM reset, it is not a FF vehicle and there was no data showing FF%. Would the PCM calculate FF in the background?

0

Paul from Shelby Township

   

Engineer
   

Hi Robby,

Your V10 data looks like what I am commenting about. There is no oxygen storage in the outer loop because the catalyst is gone so the control system gains are too high and it's bouncing off the stops.

If it's a FFV, the software tries to infer the percentage of ethanol after a KAM clear of a refuel event. If it's not an FFV, the software assumes a 10% ethanol content. Normal adaptive fuel learning may have learned something different for a variety of different reasons that can affect AFR. It takes a 2 or 3 FTPs to learn a reasonable LTFT correction. 

+3

Scott from Claremont

 

Manager
 

Thanks Paul, I appreciate your chiming in here. Yes I can see from what you wrote and the number of SAE papers Randy has referenced that these systems have come a long way and no longer work exactly how they did years ago.

This definitely shed some light on the situation. Thank you.

+2

Scott from Claremont

 

Manager
 

Thank you for the extremely enlightening descriptions here Paul, I appreciate your insight!

0

Robby from New Market

 

Mobile Technician
 

Hiya Scott, I'm a little late to the party, but would like to share a few thoughts. 

As far as adding LTFT and STFT, generally speaking, I feel it's pretty accurate to do that. I do it all the time when diagnosing fuel system issues. Plus, there are several documents from the OE (like you shared) that state to do that. However, there are times when that does not seem accurate. BTW, when I say "Total FTs", I'm simply adding LTFT and STFT. I'm not referring to Rear Fuel Trims - I'll get to that in a minute.

To show you what I'm talking about, I did some playing around with fuel trims on my 2007 Ford Expedition. The initial total FTs were:

  • Bank 1: 7.5% - 3.5%
  • Bank 2: 8.7% - 4.7%

Initial data

I then created a vacuum leak (by pulling a vacuum hose) and allowed LTFT to adjust. Total FTs changed to:

  • Bank 1: 45.6% - 41.3%
  • Bank 2: 46.7% - 42.1%

Vacuum leak

Next, I fixed the vacuum leak and allowed the engine to run a few minutes. LTFT corrected back down some, but were higher than the initial.

  • Bank 1: 15.6% - 12.1% 
  • Bank 2: 16.7% - 13.4%

Vacuum leak fixed - before KAM reset

Last, I cleared KAM and allowed the engine to idle a few minutes.

  • Bank 1: 7.2% - 2.8%
  • Bank 2: 8.1% - 3.9%

After KAM reset

It may be a little hard to follow what I did here, but there are a couple of things I want to point out:

  • After the vacuum leak was fixed (before resetting KAM), total FTs were about 8% higher than the initial readings. 
  • After resetting KAM, total FTs returned very close to their initial value. 

So, why difference in total FTs between the before & after KAM resets? All of that captures were taken within just a few minutes of each other and the only thing I did was remove a vacuum line and put it back on. 

Rear Fuel Trims

First off, consider STFT for a second. When the front O2 indicates rich, you'll have a decreasing STFT. When lean, you'll have an increasing STFT. I think we're all failure with that. 

I've noticed 2 different strategies for rear fuel trims on Fords.

I'll call this one the follower. I call it that because the RO2FT follows the rear O2 signal. This is backwards from how STFT responds to the front O2. 

I'll call this one the switcher. I call it that because RO2FT is switching up and down (like a digital signal). Notice how RO2FT follows EQRAT desired. 

The names I've given to each strategy are by no means official. I made those up because I didn't know what else to call them. In either case, RO2FT appears to be trying to keep the rear O2 signal at a certain setpoint. 

+3

Randy from Denver

   

Analyst
   

Robby!

Where to start.................

You say you use it, then go on to prove it wrong so that begs the question: Do you know when to use it and when not? What is the true total fuel trim for your Expedition? I assume both before and after had exhaust gas readings of lambda 1.000?

The examples you have given use different fuel control strategies and each must be discussed in isolation. Let's get to the last one since I never find anyone that understands it. 

For Scott, (bolds are mine)

SAE … (Mazda)Measurement of Oxygen Storage Capacity of Three-Way Catalyst and Optimization of A/F Perturbation Control to Its Characteristics

"ABSTRACT In order to study alternate methods of Air Fuel ratio (A/F) perturbation for maximizing three-way catalyst conversion efficiency, two methods for measuring the Oxygen Storage Capacity (OSC) of Catalyst were developed on an engine test bench. The first is to measure just the break-through Perturbing Oxygen Quantity (POQ, which is defined as the product of A/F amplitude, perturbation period and gas flow), and the second is to measure the response delay of the rear A/F sensor, which has been improved to be very similar to the former. Then, the OSC values of many catalysts were investigated with different perturbation parameters. The results show that OSC would not be affected by amplitude, period of perturbation and gas flow, and that the best conversion efficiency is obtained when the value of POQ is about 1/2 of the value for OSC. These results suggest that the best way to control perturbation is to keep POQ at 1/2 of OSC by setting perturbation parameters. Then optimized perturbation control algorithm has been developed and satisfactory conversion efficiency is realized. Furthermore, the perturbation control is appeared to make the catalyst monitor system more confident."

I've attached files showing a steady state at 35 mph on a dyne from a similar vehicle to yours. This condition is not isolated to this speed, it does it most of the time (idle, cruise etc). LTFT is not included because it is irrelevant. Pay attention to frequency and the EQ ratio v. STFT. Is STFT a feedback or part of the feed-forward device?

For extra credit;

SAE … Implementation of Air-Fuel Ratio Feed-Forward Controller Considering Heat Transfer at Intake System to SI Engine

ABSTRACT For further development of the thermal efficiency of SI engines, the robust control of the air-fuel ratio (A/F) fluctuation is one of the most important technologies, because the A/F is maintained at the theoretical constant value, which causes the increase of the catalytic conversion efficiency and the reduction of pollutant emission. We developed the robust controller of the A/F, which is the method to change the fuel injection rate by using the feed-forward (FF) controller considering the heat transfer at the intake system. The FF controller was verified under transient driving conditions for a single cylinder, and the A/F fluctuations were reduced at approximately 84%.

+3

Robby from New Market

 

Mobile Technician
 

Hi Randy! Yes - where to start. This stuff is so hard to get across in captures and text. 

The example from my Expedition is just a classic/textbook case of a vacuum leak. High trims at idle, lower trims off idle - high LTFT and high STFT. I'll will use total trims knowing it's not totally accurate, but it gives the info I need to get a direction. From what I've seen, the inaccuracies really show up when you have high LTFT and STFT is 0 or below (in cases like my Expedition example where you fixed a fuel system fault but did not reset fuel trims). In other words +20% LTFT and -25% STFT really isn't -5% total FTs. You could also say +20% LTFT and +25% STFT really isn't +45% total FTs, but it's close enough. It would be hard to tell just from scan data whether that's accurate or not - you know there's a fuel system fault.

 "I've attached files showing a steady state at 35 mph on a dyne from a similar vehicle to yours. This condition is not isolated to this speed, it does it most of the time (idle, cruise etc). LTFT is not included because it is irrelevant. Pay attention to frequency and the EQ ratio v. STFT. Is STFT a feedback or part of the feed-forward device?"

I have no idea how to answer that question 😊. The only thing I can add is, you mentioned frequency. In this capture, I noticed RO2FT switching frequency changing while driving. Does that go along with what you see in your captures? 

I'll be looking into those SAE documents in the mean time. 

+1

Randy from Denver

 

Analyst
 

You can answer my question, look again. Is EQ reacting to STFT or is STFT reacting to EQ? The key to understanding this is 1/2 POQ.

BTW, I work with hundreds of technicians and trainers and you have been the only other person (non-engineer) to ever send me files displaying this system even though it is fairly common now. I will look closer at your last capture later, I have to go now and I see an anomaly I need to verify before giving the wrong answer.

+4

Robby from New Market

 

Mobile Technician
 

EQ is reacting to STFT. You're capture doesn't show EQ desired, but I feel that STFT is reacting to the desired EQ. Take a look at this capture from a 2013 Ford Fusion 2.5L:

RO2FT seems to be driving EQRAT_DSD. This shift in desired EQ is driving STFT which, in turn, causes the shifts the in actual EQ. Here's the same capture where I've copied the STFT trace and overlay-ed it with the EQRAT_DSD. 

+1

Randy from Denver

 

Analyst
 

Your rear o2 fuel trim pid is a happy coincidence, not reality. It’s messing with your brain. Move over to generic and ignore rear o2 ft. It appears to me your pid refresh rate is too slow as well.

Having said that, you are so close, so close. I’m not being a jerk, it’s just that explaining the rest won’t be clear unless I upload more images and explain each one and I am exhausted right now and have a very busy week or three coming up so I may have to bail on this conversation for awhile. It’s been fun!

So what is total fuel trim on these vehicles? 😀

+3

Robby from New Market

 

Mobile Technician
 

"Your rear o2 fuel trim pid is a happy coincidence, not reality. It’s messing with your brain. Move over to generic and ignore rear o2 ft. It appears to me your pid refresh rate is too slow as well."

I'm still trying to wrap my head all that. I do have some data from Mode 1, but it's much slower than the OE IDS data.

"Having said that, you are so close, so close. I’m not being a jerk, it’s just that explaining the rest won’t be clear unless I upload more images and explain each one and I am exhausted right now and have a very busy week or three coming up so I may have to bail on this conversation for awhile. It’s been fun!"

"So what is total fuel trim on these vehicles? 😀"

I still don't know the correct answer to that 😜. In a reply to Paul B, you said:

"I don’t believe you can use fuel trim in the same way we did years ago."

I see what you're saying. FTs are still very useful for gross errors. However, when you see a shift in FTs, it could be the result of a fuel system error or a normal response to a fuel control strategy.

0

Randy from Denver

 

Analyst
 

Here are 2 different vehicles, both broken, both are 2007 Mazda CX7 with 2.3L. No DTCs, failed IM240 for CO. The second one had an increase of 20% on measured fuel economy after repair. I don't have that data for the first one.

Using taught fuel trim procedures, diagnose these 2 vehicles. The fix was different for each one.

The ones titles 35 mph and 58 mph are from the first vehicle. The one titled idle is the second one.

+2

Jim from Frederick

 

Curriculum Developer
 

Since I do not see VSS or APP or TPS I have questions. The captures look the same for the drive trace portion, only the Gas bench values change a little. 1. Where were the speed ranges in this capture? 2. What caused to bump to OL 3/4 through the capture? Decel? and finally, if this is steady state as you indicate by saying 35 mph and 50 mph why is load so variable? Or is this actually a trace during the IM240 test not your own recording for information gathering.......

0

Randy from Denver

 

Analyst
 

It's an IM240 trace. Make sure you open in new tab so you see the whole image. Yes on decel fuel cut OL.

0

Randy from Denver

 

Analyst
 

I had to check my notes and papers but the sawtooth pattern is used during warm up of the catalyst so if Ford is following this principle then this capture was taken when the cat temp model prediction was low. I can't say if Ford does this though. There are 7 variations of the pattern depending on multiple factors.

Pretty simple, Huh?

+1

Steve from Irvine

 

Instructor
 

Hi Randy, 

My understanding is EQ reacts to the final O2 based off the combustion event if this is a torque based fuel strategy. 

As explained to me by an Engineer, if the post Cat O2 voltage is where they want it to be, then fuel correction is near 0. This is manufacturer specific and could vary by manufacturer of engine management. This explains why there is a difference between the fuel cell or load strategy, but if the post cat O2 is where the engineers want it be than all is ok.

0

Paul from Shelby Township

 

Engineer
 

Hi Steve,

In current control systems, there is a target rear O2 voltage that can change at different speeds and loads to produce the desired emissions. There is also a modulation around that point to get better catalyst efficiency. There is proportional feedback to maintain the target rear O2 voltage set point and there is an integral term that is used to adjust the upstream UEGO stoich point to compensate for and aging or other effects. So basically you are correct for the target rear O2 voltage, but there is a lot of other stuff going on as well. 

+1

Jim from Frederick

 

Curriculum Developer
 

STFT surely looks to be ahead of the game to me. If I understand the point you are trying to make...... the common teaching of fuel control is that STFT is mainly a response to a S1 sensor. Here your captures show an episode of "trading places" as an indicator that times have changed. This is no longer dad's Olds

+1