Need help with understanding MAF adjustments.

[redhardsupra]

VetPet, do an experiment: get all your stuff working flawless in SD, and then reenable your MAF with stock calibration. then come back and tell me how well this pinnacle of GM engineering working for you

if you want more verification, calculate it with my method from DynAir (histogram or my old spreadsheet), or the old HumpinSS' method by AFR% that was stolen, 'scuse me, popularized, by SoundEngineer.
I bet you you're gonna come with two pretty similar results, and they're not gonna be too close to the stock config.

remember that airflow is one of the hardest things to measure, very finicky and a permanently moving target, and this is the best we've come up with to deal with it, and it doesn't look bad. it doesn't reflect on the performance of the car, it just says what's the frequency of electric blips it takes to keep the heated element in the MAF at steady temp. it's very arbitrary, so i feel zero guilt about mucking with it, as long as i do it in well educated, repeatable, and verifiable manner.

try it, it doesn't hurt, and you can always go back if you don't like it.

[Dirk Diggler]

[quote=VetPet]
Dirk, I'm referencing a book by Charles O. Probst, Corvette Fuel Injection. I'm sure you've probably heard of it. If you haven't had a chance to read it I would certainly recommend it. I think the GM engineers working on the Corvette projects are a pretty bright group of people and wouldn't throw on parts without some thought. The EFI program doesn't recommend changing the MAF table unless you change to a larger MAF as well.

I'm not saying that adjusting the MAF table is totally incorrect. I'd just like to know what's really causing the mixture to go lean once you put the MAF back on. Is it really the MAF calibration or is the MAF signal being modified by some other sensor input that's causing the problem. It just doesn't seem right to try and recalibrate a measurement device unless you know for sure that it isn't right to begin with.

My take on it is the hardcoded air values vs frequency get thrown outta wack if you look at your stock klid you will notice the lids neck is a little oval shaped. Aftermarket lids are completely round. I am assuming the oval shape helps direct the air more to the center of the resistors in the maf. When you swap to an aftermarket lid and you now get more air that travels to the edges of the maf these frequencies vs are no longer valid IMHO. Also I feel what causes it to go lean is the pcm is still using these hardcoded values to figure out fueling for the same hz and now they are mismatched esp when you cam the car and are getting a lot of reversion (reading air twice amongst other things). I can prolly bet my right arm that my WB is an accurate external device (after all that is what we are paying for) so what better way to calibrate a HARDCODED sensor with a device that is able to measure exactly what we need to dial the maf back in. Question with all the electroincs in our vehicles do you think a one stop shop harcoded value will take into account all of the tolerances of the electrical components in the vehicle. Any grounding feedback, radio noise from the ignition sytem and everything else that can contribute to dirtying that signal. I am pretty sure when the calibrated it they calibrated it in a controlled environment where this wasnt an issue. Also I feel to avoid waranty work the best way not to cook a motor under ANY condition is to run it a little richer. Ie come up with a calibration and then add 10% for good luck. Has anyone logged a bone stock car where commanded = actual via the ve or them maf?

If you TRIMS are derived from your o2 sensors closing the loop, fueling is based on the primary air metering device (MAF) and your trims are + who is to blame and how do you fix it?

PS GM spent millions of dollars in development for the LS1 to stick a damn 10 bolt behind it. You think I am gonna trust them with calibrating a sensor for me.

PPS If these MAF sensors are as acurate as we believe there would be no reason to buy EFILive, WB's and all the other fancy smancy tools we have. It wouldve all been taken care of at the calibration plant one time and one time only unless you spent 650 to change your rear end gears

There's one either item that I found interesting in reading the above book. I was under the impression that if your LTFT's are positive that they affect fueling under WOT. If I"ve read correctly, once you're in WOT open loop operation, the LTFT's are out of the picture and have no effect. Your WOT AFR would be dictated by the commanded AFR based on your PE table values and MAF sensor input. I'll have to look to find the page reference for this.

I am not a LTFT tuner I havent been for some time. I dont like narrow bands and how headers affect them and the likes. I will stick with my FJO unit. I dont know what all the pcm does to come up with LTFT values but I do know that after dialing in my ve table with my WB and then calibrating the maf with my WB the LTFT's are perfect. I have shortened my tuning experience since i have elimintated the need to wait for the car to relearn the trim values and can vurn more gs tuning instead of driving 40 pointless miles to get them to settle in again.


My tuning method was developed by hours of logging cause and effect open loop and closed I didnt follow any of the stickies but helped pioneeer some of them you just dont see my name in them but I have proof read most of them. You can do searches on Tech under my HumpinSS handle and read all the way back in 04 when i first started messing around with this SD/MAF stuff and my findings and methods. Marcin (RedhardSupra) will tell you as him, I, Black02SS, txhorns, TAQuickness and a few others hashed most of this stuff out. People like wait4me and NoGo also spoke of some of this stuff even before that but it never caught on and they were chastized for their beliefs. I was doing ve tuning with edit (before hpt, right BINK ) using excel and access to derive all the information you are getting today via maps/histograms. Could you imagine trying to average and change a ve table with 65k lines of data in excel filtering out FTC 21 and then when i got smarter with excel using formulas to derive the AFR error/BEN then trying to apply it. LOL talk about a long freaking time. The inconsistencies in the maf is what got me to unplug it because i once believed it was a perfectly calibrated sensor it bothered me that my trims would go -10 when unpluged and then I would plug it back up and everything would go +10 so i set out to find why. Now that I/we have figured most of it out I am just to lazy to put one back on the car and stick with full time OL SD

[joecar]

More info: http://www.ls1tech.com/forums/showthread.php?t=363392

[VetPet]

VetPet, do an experiment: get all your stuff working flawless in SD, and then reenable your MAF with stock calibration. then come back and tell me how well this pinnacle of GM engineering working for you

if you want more verification, calculate it with my method from DynAir (histogram or my old spreadsheet), or the old HumpinSS' method by AFR% that was stolen, 'scuse me, popularized, by SoundEngineer.
I bet you you're gonna come with two pretty similar results, and they're not gonna be too close to the stock config.

remember that airflow is one of the hardest things to measure, very finicky and a permanently moving target, and this is the best we've come up with to deal with it, and it doesn't look bad. it doesn't reflect on the performance of the car, it just says what's the frequency of electric blips it takes to keep the heated element in the MAF at steady temp. it's very arbitrary, so i feel zero guilt about mucking with it, as long as i do it in well educated, repeatable, and verifiable manner.

try it, it doesn't hurt, and you can always go back if you don't like it.

I plan to do just that in the springtime. I'm fully prepared for tuning now having the equipment and a lot more knowledge, thanks in no small way to this site and in particular people like yourself, humpinSS, dirkdiggler, joecar, topless texan, jpfilla and more. It is because of your efforts and great programs from EFILIve that tuning has become a lot easier. I'm certainly looking forward to tuning and will continue to look for help from you guys. I hope that I can make a useful contribution soon.

[TAQuickness]

More info: http://www.ls1tech.com/forums/showthread.php?t=363392

Beat me to it Joe... FYI that thread starts off very well with MAF fuctionality and how engine mods effect performance of the MAF and ends with splitting hairs.


To summarize that thread, any modification to the engine that effects airflow will have consequense on the MAF. Some mods have greater consequence than others.

[dfe1]

Let's bat this around--

Assuming that the stock MAF table is correct-- correct meaning that if we put a MAF on a flow bench and its output was the same as the bench's-- there's no need to alter the table. If the mass says the engine is using 300 grams/second of air, then that's what it's using. Now, if we look at commanded versus actual AFR and find that it's off by quite a bit, why do we want to correct it by changing the MAF calibration? In so doing, we're fudging the MAF readings so the PCM will alter fuel flow appropriately. Isn't altering the VE or IFR tables a more realistic way to correct AFRs? If commanded versus actual AFR is off, that's more a function of VE and IFR values than anything else.

If the MAF output is not correct, how do we know, and how do we know how far off it is, unless we test it? The engineers at GM are pretty sharp, and they have to calibrate engines that meet some pretty demanding emissions and fuel economy standards. They're also under the gun to do things cheaply. My guess is that if they could meet these requirements without using a fairly expensive sensor, they would. In turn, that indicates that they consider the self-adjusting aspects of a MAF to be essential, which further indicates it must be accurate. If changes are made that significantly alter the way air flow into the sensor, then some tuning of the MAF table may be warranted.

The IFR table presents a similar situation. If the stock numbers are correct, then altering them is another way of fudging. This isn't meant to imply that fudging is bad-- somethimes there's no other choice-- but it should be called what it is so you can maintain proper perspective on the tuning process.

On the other hand, I've seen many cases in the past, where the injector constant (L98 and LT1 engines) and actual injector size didn't match. The injector constant in the factory calibration had been fudged. I haven't had a chance to flow test my injectors yet to see if actual matches IFR values, so I can't comment beyond that yet.

I've also noticed that when operating in MAF mode, changing VE values definitely alters LTFTs. That means the VE table is part of normal MAF mode. (If you multiply the VE table by 50-60%, the engine will barely start, if at all.) I've also found that simply switching to speed density- no other changes-- really whacks the LTFT values. That tells me that the PCM alters fuel flow when it detects that the MAF has failed. Don't forget that speed density is only enabled if the PCM thinks the MAF has failed. We don't know what other calibration changes the PCM makes when running in "sensor failure mode".

Using LTFTs to tune when in closed loop isn't a bad way to go. The narrow band sensors are surprisingly accurate as long as AFR is stochiometric. They are a complete waste of time at other AFRs. I once dyno'd a car with a wide band, changed AFRs by a full ration (as indicated by the wide band) and the car's narrow band sensors indicated a change of only 25 millivolts. Obviously, the best approach is to use a wide band for all tuning, but if you don't have one, you can get really close with a narrow band sensor (except as noted).

Comments?

[TAQuickness]

Let's bat this around--

Assuming that the stock MAF table is correct-- correct meaning that if we put a MAF on a flow bench and its output was the same as the bench's-- there's no need to alter the table. If the mass says the engine is using 300 grams/second of air, then that's what it's using. Now, we look at commanded versus actual AFR and find that it's off by quite a bit. Why do we want to correct AFR by changing the MAF calibration? In so doing, we're fudging the MAF readings so the PCM will alter fuel flow appropriately. Isn't altering the VE or IFR tables a more realistic way to correct AFRs? If commanded versus actual AFR is off, that's more a function of VE and IFR values than anything else.

Very good point you bring up. Flow on a bench is not the same as flow on an engine. Flow on the bench is uni-directional, on the motor, it's bi-directional. There is quite a bit of detail on this topic in the link Joe posted.

The IFR table presents a similar situation. If the stock numbers are correct, then altering them is another way of fudging. This isn't meant to imply that fudging is bad-- somethimes there's no other choice-- but it should be called what it is so you can maintain proper perspective on the tuning process.

I agree, it's just fugding at this point.

On the other hand, I've seen many cases in the past, where the injector constant (L98 and LT1 engines) and actual injector size didn't match. The injector constant in the factory calibration had been fudged. I haven't had a chance to flow test my injectors yet to see if actual matches IFR values, so I can't comment beyond that yet.

I'm sure GM had their reasons, and it doesn't necessarily make it right.

I've also noticed that when operating in MAF mode, changing VE values definitely alters LTFTs. That means the VE table is part of normal MAF mode. (If you multiply the VE table by 50-60%, the engine will barely start, if at all.) I've also found that simply switching to speed density- no other changes-- really whacks the LTFT values. That tells me that the PCM alters fuel flow when it detects that the MAF has failed. Don't forget that speed density is only enabled if the PCM thinks the MAF has failed. We don't know what other calibration changes the PCM makes when running in "sensor failure mode".

One of the other changes is the PCM reverts to the low octane spark table. Adjusting your spark has a major effect on your actual AFR.

Comments?

No matter your prefered tuning method, you have to start from somewhere. In the case of LSx engines, I find the IFR table is the easiest starting point. If the injectors are stock, leave it alone. If you upgraded injectors, scale the table accordingly. By starting with the IFR table, you know have a single point to reference the rest of your tune to.

Next would be to get the idle in-line. Since the IFR table is now a constant, you can adjust the idle settings preparing you for the rest of your tune.

Next, get your VE table in-line. You now have your IFR and Idle as a constant, your VE should fall into place with relative ease. You must have a VE table in line to base the rest of your tuning on.

Then, if so desired, dial in the MAF. With the IFR, Idle and VE now constants, you can base the MAF table on them.

Does this mean the tune is correct, not necessarily, BUT, your tune will be correct to itself.

No matter what your tuning method, so long as you are consistant, and achive consistant results, then it's good enough for government work.

[Dirk Diggler]

Let's bat this around--

Assuming that the stock MAF table is correct-- correct meaning that if we put a MAF on a flow bench and its output was the same as the bench's-- there's no need to alter the table. If the mass says the engine is using 300 grams/second of air, then that's what it's using. Now, we look at commanded versus actual AFR and find that it's off by quite a bit. Why do we want to correct AFR by changing the MAF calibration? In so doing, we're fudging the MAF readings so the PCM will alter fuel flow appropriately. Isn't altering the VE or IFR tables a more realistic way to correct AFRs? If commanded versus actual AFR is off, that's more a function of VE and IFR values than anything else.

First of the bench will only measure a straight airflow not the ever changing conditions of the engine and you surely cant account for an revisions even with the stock cam. Why would you alter the ifr the ifr table is a representation of fueling hardware ie the injectors. Lets not forget our widebands are reading AIR and Fuel it is outputing to us the ratio between the two. So yes it totatally plausible to adjust the maf to get you AIR:FUEL ratio in line. The ve table is only used for sanity checking and to my understanding only used really at idle and unstable maps. Ive tried altering the ve table in maf mode at WOT and PT driving conditons and it takes large ve movements to see changes in LTRIMS or AFR something on the order of 4:1.

The IFR table presents a similar situation. If the stock numbers are correct, then altering them is another way of fudging. This isn't meant to imply that fudging is bad-- somethimes there's no other choice-- but it should be called what it is so you can maintain proper perspective on the tuning process.

I trust that at 58 psi an 0 vaccum that my injector flow X lbs/hour. Ive changed injectors and kept it along the same line as the stock GM flow rates but with representation for my injectors so I can assume these are correct. On all the cars i have tuned the ifr table was changed ONLY if the customer had changed his injectors. If they havent I havent ever ran into a situation where i had to fudge that table. I set it and forget it. I tune the car modeling the engine for its air requirements inthe Ve and MAF tables.

I've also noticed that when operating in MAF mode, changing VE values definitely alters LTFTs. That means the VE table is part of normal MAF mode. (If you multiply the VE table by 50-60%, the engine will barely start, if at all.)

At idle the map sensor works faster than the maf and the reason it wont start is because the ve table is being used for idle conditions. Pretty much belwo 4k the pcm is using both tables to figure out fueling so yes you are correct by saying the car wont start. You have actually flooded the engine by telling the PCM the engine is 50% more efficient at 00 rpm than it really is

I've also found that simply switching to speed density- no other changes-- really whacks the LTFT values. That tells me that the PCM alters fuel flow when it detects that the MAF has failed. Don't forget that speed density is only enabled if the PCM thinks the MAF has failed. We don't know what other calibration changes the PCM makes when running in "sensor failure mode".

Comments?

I wouldnt necesarily say the pcm alters fuel. It uses the ve table which may or may not be dialed in for fueling (defualt ve units are in air just like the maf table). With those hardcoded values the calibration engineer coded into those table the PCM thinks the enigne is pumping x amout of air for the given rpm and kpa and decides on fueling. Can this be wrong hell yes that is why we have these tools and are on a neverending quest to dial them in.

Will anyone answer my question

If the car is dialed in in SD mode via ve, the IFR table presumably correct and you plug the maf in and the car goes lean who is to blame?

[TAQuickness]

I'll take a stab at it...

If the car is dialed in in SD mode via ve, the IFR table presumably correct and you plug the maf in and the car goes lean who is to blame?

You are. You shouldn't have plugged the MAF back in.

[redhardsupra]

TAquickness, contrary to the popular opinion, MAF is not the source of all evil
i seriously have problems being able to tell the difference between both modes, if i disable the SES light for SD operation. The only time where i can tell the difference is at the track (the one with turns) 'cause SD is a bit quicker to react to minute changes in throttle, which you can only feel while pushing it through a corner and trying to precisely feed it gas.
so in the vain of 'guns don't kill people--people kill people':
MAFs don't fuck up your tunes--people who don't tune MAFs fuck up their own tunes