MadDog

I am trying to figure out exactly how LTFT is going to effect my GReddy turbo tune. The only means I have of looking at LTFT, like most of you, I assume, is the CANScan. Scan tools will output an instantaneous LTFT value. But, have you ever noticed how that will jump around sometimes? I did some poking around. It turns out that LTFT is a matrix of cells based on load and RPM or is a function of air mass. (I previsouly thought it was a single number that applied to all conditions) This information comes from the Ford OBDII theory manual. So, I am not sure, but I think the LTFT that the scan tool gives you is the instantaneuous LTFT cell that the ECU is using at that moment. I guess it could also be some average over the range. Does anybody know for sure what that single number represents?

Anyhow, the manual is pretty clear that the LTFT is only learned and used in closed-loop mode. The GReddy solution was to trick the car into going open-loop under boost. So, is the LTFT really a concern for us?

I thought it was, since my LTFT is at -15% after a couple of days of driving. But, now I'm not sure. Now I think that the LTFT at idle is -15% and I have no idea what the LTFT cell value is when I am under boost. If things are working as GReddy thinks, then the ECU doesn't even use LTFT under boost.

I know people, like me, have seen the scan tool report LTFT getting more and more negative. But, I can only get a scan at idle, while I am sitting still. Does anyone have information on LTFT that the tool reports other than at idle? Does anyone have actual AFR data under boost that shows a drift over time?

-MD

zoomzoom_8

I don't understand the ltft at all but I can assure you something is going on with the stock ecu and the emanage. i ran my turbo for a month with no problems then the 8 started jumping and acting insane, a/fs was terribly lean. Reset the ecu and everything was fine. i now have to do this about once or twice a week cause it gets progressivly faster.

MadDog

I did a little research on the way home from work. Using my Can Scan

Idle:
"closed loop"
ECT=206 deg.
STFT=1%
LTFT=-15%

driving NOT under boost:
"closed loop"
ECT=206 deg
STFT=4%
LTFT=-11%

dring UNDER BOOST:
"open loop driving"
ECT=82.6 deg
STFT=0%
LTFT=-4%

So, it looks like the GReddy dongle does put the car in open loop mode by masking the true ECT. It also looks like LTFT is variable. I have no idea what parameters dictate the LTFT being used at any given time. But, its a little reassuring that the LTFT under boost is a lot lower than the LTFT I was getting while at idle! I also don't know if its actually using/updating the -4% LTFT under boost.

I also logged some boosted runs using the e01. I am running between 11 and 10 AFR and very close to 10.5 AFR over almost the entire RPM range.

I guess I'll try to keep an eye on the LTFT under boost as a measure of the health of my setup. So far, the AFRs are nice and conservative while I build confidence in my installation.

two rotors

I checked the service manual and found a couple of DTCs,071 and 072 which use the LTFT and STFT values as triggers when the control system is in feed back or closed loop mode.The normal range of values of STFT are -4 to +4 % and LTFT are -12.5 to +12.5%.Values outside these ranges would trigger a DTC. I get the impression that these trim values vary constantly in closed loop depending on a bunch of factors,particularly O2 sensor outputs.At this point I am unsure what the significance is.

Rasputin

The LTFT is only learned in close loop but if it works like the Adpative Fuel in Ford strategies (reference table FN1325A and fuel table LTMBn), it is used in open loop mode too. Indeed, where the system cannot learn in close loop, the cells in the matrix refer to other cells that are in the close loop area (negative value in the reference table).

The ECU "knows" (the ECU was actually told by an engineer ) what AFR is expected for a given injector pulse width for a nominal engine. Any deviation is due to hardware variation and/or ageing. The system therefore stores a correction value for each closed loop operating conditions (the Ford table has 8 x 10 cells). As I said above, in open loop conditions, corresponding cells refer to other cells in the closed loop area and therefore use the correction value (LTFT) from that cell in the closed loop area.

Apparently some people have problem in open loop mode with the Greddy kit. It would be intersting to find out which cells in the closed loop area are refered too (by LTFT value comparison) when running in open loop. In these closed loop conditions that are refered too, one must make sure the fuel pusle width is as expected by the ECU (or within 5-10%).
To do so, no boost (or hardly any boost) should be applied in these closed loop conditions. I don't know if there is a closed loop boost control in the Greddy E-manage system that allows the user to specify which boost (or absolute) pressure he wants at specific operatng points. Can someone tell me?

Hope it helps.

Fabrice

Rasputin

I realised I made a mistake in my development of ideas above. Thinking about it, it might well give some hints of an explanation for the troubles with LTFT.

Fabrice

r0tor

iTrader: (1)

talking to some friends in the business, I came to the same conclusion as Fabrice did awhile back

the LTFT is actually similar to a greddy fuel map in the fact that its not one value but a ~8 x 10 map that trims the normal fuel map for each rpm/load condition. I also learned as Fabrice mentioned that the cells only adjust in closed loop conditions - but the cells that trim the open loop areas and actually linked to the cells in the closed loop area. One exception to this rul that I've heard is that if it detects detonation in the open loop areas it will richen them.

I've said this once before, but I believe people should be letting their LFTF's stabalize and then tune it on a dyno rather then resetting the trims and tuning the maps. The characteristics of the intake are completely different even with no boost and the computer is always going to try to correct for this - which will then effect the open loop areas your trying to tune. Restting the trims is just going to be a never ending battle.

MadDog

Since the LTFT is changing so dramatically at idle (closed-loop), I figured that it might have a lot to do with the airflow characteristics of that rediculous intake that comes with the GReddy.

Now, if we only knew the look-up parameters for the LTFT. It would be real nice to know where the LTFT that is used under boost (open loop) is coming from. It shouldn't be learning under boost.

r0tor

iTrader: (1)

I've always wondered what would happen with the turbo cars if you would adjust the greddy map to simulate the LTFT in the closed loop areas... theoretically it should make the LTFT's head towards zero

Rasputin

Exactly! You've found the solution (I think anyway). Indeed, if you make sure that, using your e-manage, you obtain an AFR very close to stoich (lambda = 1) for all close loop operating points, the adaptive fuel will think all is OK and won't try to correct (much).
It also means you can run any boost pressure you like in closed loop provided it allows you to run a stoichiometric AFR.

Fabrice

MadDog

The ECU will already seek the right AFR for closed-loop. Yes, that will be through the LTFT and STFT. What we really need to understand is where the cell value of LTFT that is used during open-loop fuel control comes from. It isn't necessarily a 1:1 with the closed-loop LTFT that would be used for the same look-up parameters.

Has anyone else logged the LTFT under open-loop and compared it to closed-loop?

-MD

Rasputin

Yes the ECU will seek the right AFR for closed loop (stoich). But it won't get without massive correction by the adaptive fuel strategy if the e-manage is not mapped in such a way that you get a stoich AFR.
Example :
Say at 6000 RPM, WOT, the reference cell for LTFT is, say, 4500 RPM, 0.4 load.
Let's see what we get at 4500 RPM, 0.4 load. The PCM expects in normal conditions to get a stoich AFR (lambda = 1) for a fuel pulsewitdh of "t".
If, after fitting the turbo kit and the e-manage, your lambda continuously remain richer than stoich (lambda < 1) because of the way you programmed the e-manage, the system will start to lean it out using the adaptive fuel strategy. Depnding on how much you're away from a stoich AFR, you might get the the LTFT limit and trigger a CEL. Otherwise, the system will take away fuel until it reaches a stoich AFR, using the lean-out LTFT. Unfortunately, the same lean-out LTFT will be used at 6000 RPM, WOT!
SO, if you make sure you program the e-manage to get a stoich AFR in closed loop conditions everywhere, you won't get into trouble. The other way to work, is to try to find out which closed loop cells are refered to in open loop by comparing values and/or airmass (as it is the factor used by calibration engineers to find the correct reference cells).
Capice?

Fabrice

r0tor

iTrader: (1)

Yes the ECU will seek stoich in closed-loop through the LTFT.

What we are saying is to use the e-manage to try and zero out the LTFT in the closed-loop area - if you can accomplish this in the closed-loop areas you will be fine in the open-loop conditions.

For example your idle is at -15. If you take fuel out with the greddy map at idle, the ECU will temporarily be missing stoich with the current LTFTs and will have to adjust over a drive cycle. If you get the greddy adjustment correct you should see the ECU adjust to ~0 for the LTFT. Then repeat this process for say 3000 rpms and 30% throttle, and get a 0 LTFT there.

If you can devote 2 or 3 columns in the greddy map to do this, you should be able to zero out the LTFT in the closed-loop areas, which means open-loop areas will zero out, which means your greddy map will not only function the way greddy tuned it but the ECU will have no reason to fight you anymore and the tune will stay.

MadDog

I understand what you guys are saying. I guess my point is that you really have no proof that getting the closed-loop LTFTs back into line will correct the open-loop LTFT. It sounds reasonable, but we really don't know what its using. Besides, the GReddy map is using boost vs RPM. That's not directly correlatable to what the ECU uses to look up the LTFT values. You'll never get the emanage map to work off of exactly the same parameters that the ECU is using.

Another tidbit of information for you. My LTFT under boost hasn't changed at all since yesterday, and that's with a lot of boosted driving. Its still at -4%. My AFRs under boost are right where GReddy wanted them 10.5 and that's across the board. So far, I have not had any problems with the LTFT. It seems to ne that GReddy might be right. Since the ECU is not learning in open-loop, its not going to fight me under boost. I'm not concerned about the close-loop LTFTs. They could be +6000% and i don't care. So far, I haven't detected any LTFT changes that are going to hurt me.

Rasputin

We have no proof that would stand in court, for sure. But I'm a engine management calibration engineer who left Ford 5 years ago (my last assignment was on a joint Mazda-Ford program). And it would not make sense not to use a refered LTFT in open loop. However, the Denso PCM could work along different lines. I don't know for sure.

If you have no problem, it means that your e-manage is properly calibrated to start with. Good for you. But others seem to have to reset every 2 or 3 days!
Regarding the fact the e-manage and PCM maps input don't match, I totally understand, but these different input should corrolate somehow. Not exactly but enough to keep it in control.
And one should be concerned about closed loop LTFT, since if they reach their limit, it will trigger a CEL apparently.

And thanks for the info on how your system works. And also for starting this interesting thread.

Cheers,

Fabrice

r0tor

iTrader: (1)

as far as my info, my ex-roomate was on the team that setup the 300c Hemi with cylinder deactivation... bit more complicated then the 8

MadDog

Fabrice,
good information! Since you are well qualified, here is a question for you. Apparently one of the look-up parameters used by some ECUs is airflow. When I am under boost, the airflow is much larger than it ever would be under closed-loop conditions. How would the ECU determine the right LTFT value to use under these conditions? It seems that it would never be able to learn the right value for this kind of flow rate since it would never see them during closed-loop operation.

You say that others are having to reset their LTFT every 2-3 days. I agree with one subtle difference. Others are resetting their LTFT every 2-3 days. I am not sure the have to reset every 2-3 days. At first, I thought that I would have to be resetting all the time, too. I was checking the LTFT that my scan tool reported while I sat in the garage at idle. I really didn't understand that the value the scan tool reported was the instantaneous LTFT and that the LTFT was really a look-up table. I was alarmed that I was at -10% to -15% after 2 days of driving. Then I started reading about LTFT on the Ford and Toyota websites and I understood a little more. I then tried to capture the LTFT that was being used under boost. I was relieved that it was only -4% and appears to have stabilized there.

I was hoping some other GReddy boosted owners would do the same thing and report what they saw under boost. If I hadn't, I would be under the impression that I had to reset a few times a week as well as I would only be aware of the LTFT at idle, which is quite high. RIght now, I think things are stable and performing quite well. I wonder what the real story is with the other GReddy boosted 8's. I guess that I put myself in a good position to start with since I installed a WB02 and have it displayed next to the boost. That way, I didn't freak-out when I saw -15% LTFT (at idle).

Rasputin

Now, I had a look at the Greddy E-manage specs, and it does not at all work the way I thought. It does not modify injection pulse width outputs based only on a map, but actually lies to the Mazda PCM by feeding it with altered sensor signals (MAFS?). I wonder how this actually interacts with the adaptive fuel. Me must think a bit.

Fabrice

MadDog

Well, that's almost right. The emanage can intercept the MAF signal and lie to the ECU, it is not using that function with the turbo setup. It is intercepting the injector pulses after the ECU and lengthening them before sending them to the injectors. That's why you can't take away fuel using the emanage with the stock setup from GReddy.

Rasputin

It's how I initially expected the system to work. It's a relief.

Thanks,

Fabrice

MadDog

Anybody else record their LTFT under boost?

MadDog

Come on! Someone has to have some concrete evidence of LTFT drift to discuss! Guys have been complaining about the stock ECU screwing with their LTFT, but I am starting to think that's totally bogus.

After 2 weeks of turbo-fueled ecstasy, my LTFT under boost hasn't budged from -4%!

Can anyone else please check their LTFT under boost? I think the kit is performing just as advertised. My car is not learning under boost and is not messing with the AFR under boost in any way.

rkostolni

Well I don't have a canscan to check my LTFT, but I definitely notice that after 2 days of driving the power starts going down and I have to disconnect the battery overnight. Also, oddly I notice a quick 5 minute disconnect doesn't do the job, it has to sit for a few hours to restore the hp. I should try to do a gtech run and quantify this loss. I don't know what the stock ecu changes, maybe its timing, maybe its AFR, but it changes something.

MadDog

What you describe must be something other than the ECU adjusting LTFT. Since the turbo is running richer under boost (target AFR for GReddy was 10.5) , the stock ECU would try to lean it out if it weren't in open loop mode. You would notice a significant INCREASE in power, followed by some nasty sounds from under the hood!

rkostolni

I don't know what the stock ecu is doing, but its not a small difference, its clearly noticeable. Many other people I've spoken to notice something similar. Unplug your battery overnight and see if your car feels faster in the morning. It is learning something.