MAF Tubes and Long Term Trim
 

It is my intent to start a “what to look out for” or, more aptly, a “what bit my butt” section on the EFIDude forum but have yet to get any free time. I did get a question from someone on one of my target subjects, MAF tubes, and thought I would post my thoughts here as well.

Before anyone attacks me for being high and mighty, I do not know everything nor do I claim to be the smartest guy on the planet. I just happened to run into this problem and, given that a lot of people will soon be exposed to a couple of ways to remap/reflash their PCMs, I thought it worth sharing. I welcome additions and corrections provided they are positive and to everyone's benefit.

When you look at the MAF calibration curve you will see that it is very flat or horizontal for a long period at lower flow rates. This means the MAF system is very sensitive at low flow rates which makes sense as this is where an engine spends most of its time (in normal applications at least) and where emissions are critical. It is also the range where long term trim is generated. Long term trim is interesting as it is applied at high loads when the PCM goes open loop. Your target fuel map can be dead on at high loads, but, if you have some -10% long term fuel trim applied to those perfect target map numbers, you can get real lean real fast!

Going back to where I started, we found that ANY change from that very carefully calibrated Mazda factory MAF tube with all those Helmholtz resonator tubes and such tends to throw off the low flow MAF calibration and cause the system to build long term trim. Mazda had some very smart people working very hard on that intake system. There is a reason it weights a ton and is so complex.

I am sure there are many ways to address the problem but the one that made the most sense to me was to take a sample of cars using the exact same MAF system and run them steady state at 15, 20, 25, 30, 35 and so Kg/hr while recording long term trim. You then correct the MAF calibration on all of the systems in the exact same way and repeat the process until all of the cars are affectively trimming about zero on fairly new motors. Once this is done, a reasonable sample of similar cars should be monitored over the life of the engine to make sure that the factory long term trim logic and boosted engine aging process generates the desired results. Put differently, you need to make sure the normally aspirated way of accounting for engine aging is compatible with a boosted engine as it ages.

For what it is worth, I would suggest you check the long term trim mechanism very carefully if you make any changes to the MAF system as I can almost guarantee you that the MAF calibration will no longer be accurate. It is not that one tube is better than another; it is just that every tube has its own very special requirements for MAF calibration when installed in the target vehicle. It is fairly easy to check by simply clearing long term trim then driving at say 40 mph in forth gear or 50 mph in fifth gear for several miles. Log data during this process and you can literally watch the PCM build long term trim. If your motor is in reasonably good condition, it should not build much. If it does, correct the MAF calibration for the new intake before the trim catches you out under high load.

Again, I welcome comments, corrections or additions.

Ok let the stft vs ltft discussions begin. And the maf tube discussion also.
olddragger

never seen a ltft in the neg.. it is usually about 3.7

beers :beer:

unless you have a vacuum leak.
OD

At higher calculated loads (like those that you see on an FI application), the LTFT ends up not making that much difference.
Considering that is more sensitive below .65 and a few PSI very easily creates calcLoad values above 1.2, the forced lambda target won't move in boost more than .1 because of LTFT.

However, you can easily make adjustments to the MAF calibration table by just letting the LTFT/STFT numbers settle (especially in the range from .20 to .85 calcLoad) and then make an offset spreadsheet and transpose the numbers.
In practice, its not usually more than 3% - 4% of volts to g/sec for any given section of the calibration line.

Fact is, the stock MAF calibration will run out of head room just as it hits the PCM's max of 2.0 calcLoad, so you will have to come up with a solution before then for the overal calibration of the MAF.

Huh???

Asphinctersayswhat?

This not the appropriate forum for this discussion

this is: https://www.rx8club.com/series-i-engine-tuning-forum-63/

You know, I just can not seem to get anything right.

just pointing it out for future reference

Sorry, Team. I didn't even notice it.
Maybe Lola can get a mod to move.