When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Have you logged LTFT values? Maybe there is something funky going on there?
Very strange that it is responding to a part of the table, and not the rest of it though.
Another check you could do is log both MAF V and g/s and see if the ECU calculation agrees with the table the AP claims to have uploaded.
I have logged LTFT, MAF V, and MAF g/s. I have not ran the car long enough in between flashes for LTFT to really learn it's self. I also never tried correlating datalogged MAF V and MAF g/s to what my table is saying it should be, I'll have to open up one of my logs tonight and look at that, that's something I didn't think to look at.
Thanks!
This is not a stand-alone ECU. We only have limited access to certain parts of the equation. There are many sub-limiters and strategies Mazda has in place to protect the engine in various situations.
There is also no shortage of people touting off ideas on a subject they still don't fully undrstand ...
That's not entirely true. It's true for us because we have a way out dated V2 AP, with only limited end user adjustability in ATR. If we had a V3 or V4 AP and AT pro/tuner edition, we'd have access to almost everything. Flash tuning has come a long ways in the last 3 years. New Subaru cobbs can so everything a TurboXS piggyback can do. New Hondata Flashpros can do everything Hondata Kpros and AEM infinities can. HPtuners for coyotes and LSs can do everything a standalone can. Obviously we aren't trying to compare a $500 Cobb to a $5000 MoTec, but most modern flashtuners are comparable to your average stand alone or piggy back.
Every time I open ATR or use Cobb I wish I ponied up for adaptronics.
Last edited by ninetysixyenko; 07-18-2015 at 06:56 PM.
Fwiw, both are needed ... and I never said it wasn't possible, but rather specifically mentioned the limitation of the thread subject as it currently exists. Woulda, coulda, shoulda is a pointless exercise. We don't have that capability so EOS ..
Talking injectors reminds me, I always had a hard start issue with the car, I took the injectors all out and sent them out to be cleaned in January, all flowed well. Upon re-installing them, I noticed I had my P1 pigtails plugged into my P2 injectors, and vise-versa. This got rid of my hard start issue, the car then sat for the next 4-5 months until the end of June. Car still has no hard start issues. All injectors should be good, so should the fuel pump (replaced when diagnosing hard starts). I would also think if it was a fuel starvation issue, lowering target lambda would not help richen it out.
This explains why my MAF needed to be cal'ed majorly last fall though, and also explains why it is currently out of whack. It does not explain why I cannot calibrate my MAF currently, or why even a stock MAF calibration is so far out of whack.
MAF g/s at idle normally shows 5.8-6.0, will the street port raise it a tad? Idle AFR is 14.6-14.8 like it should be, but it does Hunt for idle and AFR for a 2-3 seconds before it stabilizes (most likely adjusting to STFT/LTFT).
I figured 11.2 was rock bottom AFR, most other cars with factory widebands usually flat line around there. After charging the dead battery, the car would not start. It acted like a bricked ECU or a failed upload, I re-uploaded the map and it started right up. Could be coincidence.
no the proting won't change the idle g/sec, but you are at a good number if and only if you are at that number with a stock maf scale. What is your fuel trims. If you have big fuel trims from it hunting then that would show somwthing is wrong. Fuel trims are the equivalvent of having the afr being wrong.
Sounds to me like you might still have an injector wiring issue.
Last edited by logalinipoo; 07-20-2015 at 05:05 AM.
Have you logged LTFT values? Maybe there is something funky going on there?
Very strange that it is responding to a part of the table, and not the rest of it though.
Another check you could do is log both MAF V and g/s and see if the ECU calculation agrees with the table the AP claims to have uploaded.
LTFT is +11% almost everywhere on stock MAF scale.
MAF G/S measured and expected align with each other with stock MAF scale or my own 'corrected' MAF scale.
I tried scaling my Injectors today, taking 'bank 1' down to 300, from 369. I had a 12.7-12.8 throughout the whole RPM range except with 14.0-14.2 spikes when other runners open up. The problem with this is in closed loop, it is now pulling 14% fuel. At idle, it is pulling 8% and only getting a 14.4 AFR.
I also tried returning 'bank 1' to 369, and scaled 'bank 2' to 440 from 476. I had a ~15.5 AFR until 5200 and then it dropped to a 13.0 and stayed there for the most part. Both injector scalings were with stock MAF scale.
I guess I am going to try another new MAF and if that doesn't work, call it quits on the MAF scaling, keep the 'Bank 2' at 440, and then just adjust air/fuel tables according to actual AFR.
well to start with you're just going about it all wrong. you adjusted your bank 1 injector by 19 percent so of course the car is going to try to remove it and get 14% closed loop trims.
Then you adjusted bank 2 which does not come on at idle so it obviously is not the problem at all.
You say MAF measured and expected line up. how do you know how much to expect from the engine. Then if you've scaled it it will be off. There is only one place in the whole scale that should be semi known. Even that place has variation's between every engine.
well to start with you're just going about it all wrong. you adjusted your bank 1 injector by 19 percent so of course the car is going to try to remove it and get 14% closed loop trims.
Then you adjusted bank 2 which does not come on at idle so it obviously is not the problem at all.
You say MAF measured and expected line up. how do you know how much to expect from the engine. Then if you've scaled it it will be off. There is only one place in the whole scale that should be semi known. Even that place has variation's between every engine.
That is exactly why I scaled bank 1, it obviously shows the Injector scale/flow is not the issue like you suggested in your prior post; I eliminated it as a potential problem. But by scaling them clearly I can adjust AFR. If I had scaled them with a closed loop change, but no WOT AFR change, it would have proven that there was a fuel flow/pressure issue in my fuel system.
I adjusted bank 2 exactly because it does not affect idle or closed loop, to see if I could control WOT AFR with bank 2 scaling, leaving bank 1 alone. I can, and it works fine once they have opened up, but it is obviously not the proper way to do it. Not to mention it still leaves me lean before they open up, granted <4000 WOT is rarely used.
MAF measured and expected align, showing that cobb is not failing to upload the new scale, like blu3dragon mention, eliminating that as a potential problem. That doesn't change the fact that I can do a 40% global scale upwards on G/S aiflow and the car STILL says I am lean. 40%, I've tuned aftermarket turbo cars that don't need a 40% MAF scale. Stock scale and 40% global scale still shows measured and expected aligned, with NO change to AFR. To me, this says my OL tables are not correcting fuel flow according to my MAF reading. You're going to tell me that my street ported motor with pretty much stock intake diameter size piping is calling for well over 40% increased airflow from a factory tune? that I should keep scaling it?
I also logged injector duty. There was no increase in injector duty going from a stock MAF scale to a 40% increased MAF scale. There was a lot of injector duty increase when i scaled the injectors. Again, this tells me my OL tables are not correcting fuel flow according to my MAF reading.
Considering I have no idea what I am doing and I don't understand anything, what else could cause the ECU to not react to a MAF reading, other than a bad MAF, or a problematic ECU.
LTFT is +11% almost everywhere on stock MAF scale.
MAF G/S measured and expected align with each other with stock MAF scale or my own 'corrected' MAF scale.
I tried scaling my Injectors today, taking 'bank 1' down to 300, from 369. I had a 12.7-12.8 throughout the whole RPM range except with 14.0-14.2 spikes when other runners open up. The problem with this is in closed loop, it is now pulling 14% fuel. At idle, it is pulling 8% and only getting a 14.4 AFR.
I also tried returning 'bank 1' to 369, and scaled 'bank 2' to 440 from 476. I had a ~15.5 AFR until 5200 and then it dropped to a 13.0 and stayed there for the most part. Both injector scalings were with stock MAF scale.
I guess I am going to try another new MAF and if that doesn't work, call it quits on the MAF scaling, keep the 'Bank 2' at 440, and then just adjust air/fuel tables according to actual AFR.
There is definitely something strange going on. I'm not sure a new MAF will help at this point. The fact that there was a reaction to your changes in only part of the MAF scale seems more like a software/programming thing, but I guess there could be some funky hardware thing at play here.
I was actually asking about LTFT with your own 'corrected' MAF scale. Does this somehow cancel out the corrections you are making for part of the g/s range? (keep in mind for LTFT the g/s scale is split into 3 regions, each of which can have their own learned LTFT value)
It may well be time to give up on scaling the MAF for OL and go do it the "easy way" as team suggests. Scale the MAF for closed loop so it does not build much LTFT, then scale your OL fuel tables for open loop (ignoring the actual values, just change them by a % and observe the resultant AFR). I would leave your bank 2 injectors at the stock scaling value for this. I personally don't like this approach, as it implies we are missing something, but it may well be the only way to work with the current state of the tuning software
I posted my previous reply before refreshing to see this. Sounds like you have worked quite thoroughly to debug it... I am still trying to work my own scale out. It does not quite react the way I expect either, but it does seem to alter AFR based on an updated MAF scale for the higher g/s values. This is an s2 with Mazdaedit though.
Reading through your post, I wonder if there is some protection, or sanity check the stock ECU applies to the MAF scaling, or perhaps a different scale entirely used for open loop? I think it would be quite reasonable to suspect that neither cobb or Mazdaedit do everything the way you might expect them to. (I have some history with both cobb and hondata on a civic prior to this car and there have been some funky effects created by both solutions over the years. At this point though flashpro is a league and a generation ahead).
Originally Posted by ninetysixyenko
That is exactly why I scaled bank 1, it obviously shows the Injector scale/flow is not the issue like you suggested in your prior post; I eliminated it as a potential problem. But by scaling them clearly I can adjust AFR. If I had scaled them with a closed loop change, but no WOT AFR change, it would have proven that there was a fuel flow/pressure issue in my fuel system.
I adjusted bank 2 exactly because it does not affect idle or closed loop, to see if I could control WOT AFR with bank 2 scaling, leaving bank 1 alone. I can, and it works fine once they have opened up, but it is obviously not the proper way to do it. Not to mention it still leaves me lean before they open up, granted <4000 WOT is rarely used.
MAF measured and expected align, showing that cobb is not failing to upload the new scale, like blu3dragon mention, eliminating that as a potential problem. That doesn't change the fact that I can do a 40% global scale upwards on G/S aiflow and the car STILL says I am lean. 40%, I've tuned aftermarket turbo cars that don't need a 40% MAF scale. Stock scale and 40% global scale still shows measured and expected aligned, with NO change to AFR. To me, this says my OL tables are not correcting fuel flow according to my MAF reading. You're going to tell me that my street ported motor with pretty much stock intake diameter size piping is calling for well over 40% increased airflow from a factory tune? that I should keep scaling it?
I also logged injector duty. There was no increase in injector duty going from a stock MAF scale to a 40% increased MAF scale. There was a lot of injector duty increase when i scaled the injectors. Again, this tells me my OL tables are not correcting fuel flow according to my MAF reading.
Considering I have no idea what I am doing and I don't understand anything, what else could cause the ECU to not react to a MAF reading, other than a bad MAF, or a problematic ECU.
Per a previous post from another member, did it ever occur to you that you were adjusting in the wrong direction and hit the limit of what the PCM is programmed not to exceed?
As an example, try inputtting a value in the Open Loop gear/fuel map exceeding 1.00/14.7 ...
Originally Posted by logalinipoo
I concede.
Yeah, the lack of force is strong in this one ... so sad
I'm a little overwhelmed by the more in depth discussions going on here and the occasional disagreements over proper scaling methods. Would anyone be able to take a look at my cruising log and tell me what they think?
SHnifty, you're in open loop there, and you're running pretty rich.
It's easiest to look at it in closed loop. Then as you build long term fuel trims adjust them out. I would guess something is off, but I don't know what your fuel maps are set to. If your commanding it to run that rich then you look pretty good.
Thanks for the reply. About 60 miles on this tune. As far as I know, I'm running a stock map. I took the stock style cobb map and adjusted omp, fans, and catless cel. Cruising is the log I want to start with for MAF scaling right? What do I need to do to log in Closed and how long do I need to wait after flashing before it learns trims enough to adjust the tune? I swear I've spent 4 hours reading about just MAF scaling and I feel like the more I read, the less I understand where to start.
How to Scale your MAF for Flash Tuning (Cobb, Hymee)
What's your end goal? Are you extra concerned about the command equaling the actual? Or will you just be happy with a tune that produces the desired result.
Either way you should start with the idle log. Just take your time. After 60 miles you should have decent idle trims set. When you get home from a drive sit for a minute then pull a 30 sec idle log.
If your maf is in the 5.5 g/sec range and the trims are below 10. Scale the maf table from the third cell to the last cell by 0.75*ltft. Then drive around and watch your trims. After that adjust the scale in the specific maf range that is off. Tri.s set faster the larger they are. I've seen some in seconds. Anything below 5 doesn't need to be adhusted. I aim for below 3. Any more accurate and it's going to be off every time the wind blows. To log in closed loop watch your stft. As long as it's active your in closed loop.
If your g/sec is off and or you have a bigger trim then it is likely you have a problem.
On your fuel gear tables take any squares below 12.2 and bring them up to 12.2 for cat less only.
Also give your coils some more dwell. I like team's table, but I'd at least multiply your current tables by 1.45.
Then your pretty much done for a na tune. After that it's just playing around. Look up the inspection threads for a list of all codes to block. There's several that don't throw codes but the computer won't become ready if they aren't set.
Last edited by logalinipoo; 10-17-2015 at 05:02 PM.
No problem if you need more help post a log. or shoot me a PM. I like where you're going with the canbased tuning. I read a bunch of Oltman's posts, btu could never figure out how he was pulling the data. That's a little above me.
Yeah, I stayed up until 2am a few nights back reading through a thread from 2006 called "cracking the ecu". I have a decent idea of what needs to happen in the long run but I've got to get familiar with can and odb2 before I can produce anything. We will see. Trying to learn how to create a tune and flash it to the car at the same time. Ha.
I know this is an old thread, but still the topic is relevant, at least for me ;-)
I followed carbons way to scale the MAF, reading OL AFRs and comparing them with target values from the OEM OL fuel maps (using the excel plugin he linked in his posts). I than fitted a curve following the funcion y=a*(x-a)^c to extract a new MAF scaling. I do not trust the results, as they seem to be much too fare off in the lower parts of the scale..
A huge factor in the whole thing is the load, as this defines the target AFR and I'm a little confused about the values I got there. I use Mazdaedit to log and map. There is even a parameter "Commanded Air/Fuel Ratio (OBD) (AFR)" that I can log, I was unable to see the logged values in the OL fuel tables though finding the correct cell with corresponding rpm and load. Also load is a problem. There are 2 parameters I found, "Absolute Engine Load (OBD) (%)" and "Calculated Engine Load (OBD) (%)". The second one is always the higher value. Which one corresponds to the load that is used to find the target AFR?
I also made some CL logs but the fuel trims seem rather eratic to me. I have a huge LTFT in idle (about 15!). Steady cruisin I see low numbers for total fuel trim: 1500 to 3000rpm (all positiv in the range of 2). At 4000rpm the combined fuel trim ramps up above 10 again. I also did a Cl run going up slowly to 4200rpm. Again the combined fuel trims range from low negatives to about positive 20 at maximum. It is very hard to make a trend analysis out of this data. Could it be, that logging CL needs the map to settle for a longer time? I flashed a basemap and started logging more or less right away.
Can someone point me in the right direction here?
I skipped the idea of fitting a function for the new scaled MAF curve for now and adjusted values by hand in the voltage range I see from the logs...
There are 2 points where the resultion MAF curve does not look smooth. On at ~3,7-3,8V and another at ~3-3,1V this coresponds to an rpm range from about 4000-4500rpm (for 3-3,1V) and 7200-8000rpm (for 3,7-3,8V).
Could this be related to the switching of SSV/VDI?
While my knowledge is still growing, I'd suggest checking out my thread. I wrote out Kane's method of tuning from his video recordings. In my first post there is a spoiler with my plan to follow what Kane went over. It is fairly straight forward and you could probably be done within a day if you wanted. I had several vacuum leaks at idle I had to fix. Recently when I got time to do higher rpm testing I had found that my Tach was not reading anywhere close to what the ECU read in high rpm.
On the Rx8 since we have the variable intake valves and staged fuel injectors you may find it easier to climb your way up the rpm ranges by these stages.
Starting with idle tuning can help weed out potential errors like vacuum leaks. Which would throw off any tuning you do.
07-18-2015, 05:00 PM
