Confused about MAF

[wcj]

Can someone help clarify my question about MAF changes for me?

Ive logged average MAF airflow (g/s) vs RPM as well as CALC.SELEBEN vs RPM. My logged average MAF aiflow g/s is considerably different than my calculated MAF table after factoring in my CALC.SELEBEN. Ive eliminated TP% changes and temps below 178* on both by way of the map filters.

What is the main difference in what Im looking at?

Why are the 2 different?

Im just trying to get a better understanding for each.

[joecar]

How are you calculating MAF (from VE...?)...? Show me your calculation.


Which PCM OS is this...?

[wcj]

Os - 12221588

1294412943

[joecar]

You didn't log GM.VETABLE_DMA...

the CALC.MAFT calculation starts with VETABLE_DMA, applies the SELBEN correction, and calculates a MAF table from that.

[joecar]

Posts some pics of:
- your VE B0101 table,
- your VE BEN map,
- your CALC.MAFT map,

for each map, make sure to checkmark the Units box on each of the Data/Row/Col tabs.

[joecar]

The uncorrected calculated MAF is defined as "{GM.VETABLE_DMA.VE}/{CALC.DAT.K}*{SAE.RPM}*{SAE.MAP.kPa}/15"...

the CALC.MAFT calculation is the SELBEN correction of that...


is CALC.MAFT what you're using as your MAF calculation...? If so, then you didn't have VETABLE_DMA in your log.

[joecar]

Something majorly wrong with your calc_pids.txt... I'm fixing it...



so on your analog voltage input AD2 you have the pin shorted to ground, is that correct...?

[joecar]

Os - 12221588

DYNAIRTMP_DMA is defined for that OS... I see it in your log.

What about VETABLE_DMA, does it show up when you validate pids...?


What year/model/vehicle is this in...?

[wcj]

You'll have to forgive me for a bit. Been a long day at work so it will take me some time to shift my brain to efi.

I completely forgot I switched OS so Im sure my pids reflects that. I have a 2003 Corvette Coupe.

AD2 is not shorted. This is what my analog outputs look like.

12948

DYNAIRTMP_DMA is logged and is a valid for me.

VETABLE_DMA is a valid pid. I just relogged my drive home with it enabled, but I havent built any maps with it.

MAPS-
12949
12950

VE Table-
12951

Im also attaching the log from my drive home.

12952

[joecar]

No worries :cheers:


Your calc_pids.txt had some serious problems (calculating out-of-sequence, interfering with predefined pids, other stuff) which I fixed...


VETABLE_DMA is used for the Calc.MAFT calculation.


Ok, your AD2 input is not connected to ground (for voltage offset correction to AD1), I'll edit out the offset correction...


Try the attached calc_pids.txt file, copy to the User Configuration folder, and then restart the scantool.

[joecar]

Also look at post #29 here: Summary-Notes (http://forum.efilive.com/showthread.php?14188-Summary-Notes)

it explains the difference between Calc.VET and Calc.MAFT

( note how "calculated" and "corrected" are used )

[joecar]

So your LM1 is programmed like an LC1 (0V = AFR 7.35 and 5V = AFR 22.35)...?

Edit: I see that it is. :cheers:

[wcj]

So your LM1 is programmed like an LC1 (0V = AFR 7.35 and 5V = AFR 22.35)...?


Yeah, that's the best I could find for the analog LM-1.

Btw, thanks for looking at all this for me.

[joecar]

if you're doing Calc.VET (which corrects MAF and calculates VE from it):

[ running with VE disabled via B0120 (altho the transient filter removes VE influence) ]


corrected MAF:
http://img221.imageshack.us/img221/22/log0028calcvetcorrected.png


calculated VE:
http://img801.imageshack.us/img801/456/log0028calcvetcalculate.png

[wcj]

So since my MAF is not disabled, the data from the posted Calc.MAFT would be horribly skewed due the the MAF, correct?

[joecar]

if you're doing Calc.MAFT (which corrects VE and calculates MAF from it):

[ running with MAF disabled via C2901,3 ]

corrected VE:
http://img59.imageshack.us/img59/7985/log0028calcmaftcorrecte.png


calculated MAF:
http://img826.imageshack.us/img826/8556/log0028calcmaftcalculat.png

[joecar]

During steady state, the MAF is the source of airmass.

During transients, above B0120 the MAF is the source, and below B0120 the VE is the source.


So during Calc.VET, if you filter out transients below 4000 rpm (stock value of B0120), you are filtering out the VE contribution... but you shouldn't rely on this (how can you be 100% certain that you filtered out all the VE contribution). It is safer to edit B0120 to disable the VE, but see post #20 below.


During Calc.MAFT, there is no way to filter out the MAF contribution because when you filter out transients you're removing the VE contribution and keeping the MAF contribution. It is necessary to edit C2901,3 to disable the MAF.

[joecar]

Easy way to remember all of this:


whatever thing is being calculated (the name of the method) has to be disabled...

and the other thing has to be corrected so it has to be enabled.

[joecar]

Calc.MAFT: corrected VE (enabled) -> calculated MAF (disabled)

Calc.VET: corrected MAF (enabled) -> calculated VE (disabled)

[joecar]

So since my MAF is not disabled, the data from the posted Calc.MAFT would be horribly skewed due the the MAF, correct?If the MAF is enabled it is contributing to the airmass computation rather than the VE... the correction BEN factor is the error fraction in the source of the airmass (MAF if it is enabled).


Since Calc.MAFT is using corrected VE to calculate the MAF, it misses its point.


Of course, if the VE and MAF are already sufficiently close then you won't see any significant difference.

[wcj]

Ok so my B0120 is set to 400. I guess I can set it to 0, but Im figuring 400 is probably low enough and like you said, the rest can be filtered out.

I wanted to work with Calc.VET because I know my MAF is a lot closer than my VE. I know my VE table is all jacked up so I wanto to get that right before I work on any other part.

[joecar]

400 is good because that is below the operational range (it only sees 400 rpm during starting).


Ok, Calc.VET is a good decision (it corrects the existing MAF and calculates a new VE from it).

[wcj]

Ok. I relogged my drive to work this morning using the calcpid you fixed. This is what I got.

MAF 'corrected'
12959

VE table 'calculated'
12960

12962

[joecar]

The MAF correction looks pretty good, everything is very close to 1.0.


The VE calculation looks to be lower than what it should... can you post a pic of your previous/original B0101 table.

did you enter the correct displacement (in scantool go Edit->Log File Info);

did you apply the transient filter and the low cell count button...?

[wcj]

The VE calculation looks to be lower than what it should.

Agreed.

I updated the log file info with 5669 for displacement. Applied map temp/throttle filters and Im using low cell count of '10'. I could increase the low cell count to 25 or even 50, but I dont think it will move the 'calculated' VE much even if I did that.

12964

Here are the mods for the car.

Fast 102 w/ LS2 tb
Hand ported 61cc 799 heads (285 cfm@600)
228/236 .588/.601 @ 114
Delphi 42lb injectors
ARH LT Headers w/cats
SLP Underdrive pulley
Halltech Venom CAI

This is my tune after making VE adjustments from this morning's log.
12963

[wcj]

I'm wondering if I should now switch to calcMAFT and just update the corrected VE table.

[joecar]

After looking at your B0101 VE table (post #25), I think the Calc.VET map (post #23) is actually ok.


Take some more logs, let's see if we can get consistent repeatable results... then we can try Calc.MAFT after that.


BTW: thanks for posting your results :cheers:

[wcj]

Thanks again for looking through this with me.

Alright, here are the results from this morning's drive to work. Same filters/low cell count.

Corrected MAF shows a little rich compared to yesterday but it doesnt too bad. Most of the CalcVE is consistent with the CalcVE from yesterday.

12974

12973

12975

[joecar]

So Log_0031 is from CALCVET_0002.tun...?

Looks like it's repeatable/consistent (I'll analyze it deeper later today)... take another log later in the day and lets compare it to Log_0031.

[wcj]

Yes. It is from that tun. I will relog my drive home for us to compare.

[wcj]

Maps and log from my drive home. Same filters applied. Same 0002 tune. It looks pretty consistent.

12978

12979

12980

[joecar]

Yes looks consistent...

Now log some high load (WOT blasts)... but be safe about it;

also pay attention to see if you hear knock.

[wcj]

Good part about my MAF is it was WOT tuned on a dyno so Im pretty accurate there. Knock shouldnt be an issue. I just need to find a safe/secluded place to do it unless I want to spend $100 to rent a dyno for an hour.

Does it matter what gear my WOT pulls are in? Can I do 5+ 2nd gear WOT pulls to fill the needed cells?

[joecar]

Gearing changes the load on the engine... use the gear(s) you most often would use for a WOT blast.

[wcj]

Couple of 2nd and 3rd gear wot pulls.

12985

12986

12987

[joecar]

MAF upper end looks like it diverged...

[wcj]

It does look all kinds of rich.

[joecar]

I'll analyze the log more closely later today (sorry yesterday was too busy).

[joecar]

Ok, this is what I see regarding the MAF correction map in post #35 compared to the MAF map in post #31:
- post #35 is close to 1.00 for below 6300 Hz which agrees with post #31 which goes up to 6300 Hz;
- post #35 drops below 1.00 for above 6300 Hz which means the MAF isn't yet correct in the higher range;

so I would go ahead and apply the MAF map from post #35.


I also see that the VE calculation map in post #35 iv very close to the VE map in post #31;

so I would appy the VE map from post #35.


After applying both VE and MAF, take another log that includes both part throttle and wide open throttle and lets compare.

[joecar]

Higher gears (numerically lower ratios) place a higher load on the engine (high cylinder airmass) compared to lower gears (numerically higher ratios)...

as long as you eliminate cells with low hit counts and apply the transient filter, the logs taken in different gears should produce results which are consistent agree with the already corrected/calculated tables...


I've seen a few cases behaving very nicely and consistently... but I have also seen a few cases diverging which we couldn't explain.


Something we can try: after taking a log that includes 2nd and 3rd gear pulls, edit the transient filter to also filter by gear (this is tricky if you have a manual transmission, we would have to construct a pid for the ratio of RPM to VSS, or something similar), so then we could see the correction/calculation for each gear and see which portion of the tables they hit.

[wcj]

Yeah, I have a m6.

I'm suspicious of how far off my 6300+ MAF AF is reading (10-15%). The car was tuned by a local tuner not more than 3 weeks ago on his dyno so either my WBO2 is reading off or his is. I do have cats and his dyno used a tailpipe WB so his 'corrections' may be why it's running rich.

I stopped by AZ this morning and picked up a new WBO2 for my LM1 ($50 SWEET!) so I will relog using the new WBO2 to make sure Im still getting the same corrections from the new sensor. Once I confirm the previous logs using my sensor was not skewed due to incorrect sensor readings, I'll do some more high load/rpm logging and see what we get.

Side question... Im putting 450 rwhp and currently have the Deliph 42lb (3 bar) injectors. At 4 bar, these are closer to 51lb. Im only seeing about 60% DC, am I running too much injector and can that affect my ability to get the tune dialed in?

[joecar]

When you added those injectors did you recalculate the B4001 IFR table...?

I'll have to look later today at your tune file to see the IFR...

it sould be ok because your VE calculation is spot on.

Injectors that are too big will mostly affect idle since its more difficult to get them flow only a small amount of fuel.

[joecar]

Yes, the LSU4 is a good price :cheers:

[wcj]

Yep, I adjusted the IFR using data calculated from the injector worksheet. Also made some adjustments to the pulse width voltage adjuster based on the worksheet as well.

These are the injectors I have.

http://www.racetronix.biz/itemdesc.asp?ic=01D030x&eq=&Tp=

I calculated the flowrate at 43.77 (3 bar) based on this:


Dynamically flowed and grouped to within 1% tolerance using Racetronix custom injector flow bench when purchased in packs.
All tests run with n-Heptane @ 20oC (+/-1oC), 3.0 Bar (+/-0.01 Bar), 14VDC (+/-0.05VDC), saturated driver with RC cutoff.

Injector flow rates are based on testing with n-Heptane test solvent (0.6855 g/ml). U.S. EPA emission test gasoline 40CFR86.113 (0.735g/ml) flows in most common injector designs at about a 3 to 4% higher rate than n-Heptane. A ratio of 1.035:1 can be used when converting the flow specifications above.

[joecar]

Ok, the IFR looks correct based on your injectors.

[joecar]

Do you have the original tune file that the PCM came with from GM...?

[wcj]

Do you have the original tune file that the PCM came with from GM...?

This is the tune that was on the car when I got it.

13018

It was tuned by my tuner for a Halltech CAI, Headers, 160* stat before I bought the car.

Point of reference, this is the original OS. I have since swapped to a different OS for additional logging capabilities.

The file I used for the swap was 13019

[wcj]

Alright. I feel good about the previous logs from the old WB. Logs with the new WB shows consistency.

13028

13029

So I guess I can go ahead and make some adjustments to my 6300+ MAF tables and do some more 2nd/3rd gear WOT logs.

[joecar]

Yes, I agree.

[wcj]

I made some manual adjustments to take some fuel out 6300+ and smooth the MAF table based on the previous WOT log.


13032

13031

[joecar]

The after looks fairly good.

[wcj]

This was my drive home. It's a lot closer.

13038

13039

13040

[wcj]

I know it's probably not necessary to keep posting up the logs and screenshots but I may as well keep doing it until my MAF is dialed in and I get a good usable reference for the VE table just in case someone new (like me) to this is looking for reference to see what it looks like from one adjustment to another.

I did increase my MAF from 2500-6000 by .75% just to tilt it on the rich side of 1.0.

Just as a note, my PE calls for 12.64:1. After getting the 6300+ MAF table more in line with that, the car 'feels' like it pulls easier.

13043

13044

13045

[joecar]

After a few iterations the VE and MAF have to converge as the new maps are applied (i.e. stay the same).

Yes, when it seems to pull better that is a sign that the AFR is getting close to what the engine likes.

[wcj]

Just to update, Im on tune edit (6) and Im now slowly chipping away making small adjustments to the 6500+ cells of my MAF table. I think it is interesting to see how far off my MAD table was after I got the car tuned.

Original
13062

Tune (edit 6)
13063

This is a pic of a 3rd-4th short shifted WOT burst this afternoon. The AFR is just about where I want it through this range.
13064



MAF Freq
Edit 6
Original
% Change


1500
3.007813
3.007813
100.00%


1625
3.648438
3.648438
100.00%


1750
4.34375
4.34375
100.00%


1875
5.0625
5.0625
100.00%


2000
5.898438
5.898438
100.00%


2125
6.773438
6.773438
100.00%


2250
7.757813
7.757813
100.00%


2375
9.507813
8.828125
92.85%


2500
9.890625
9.992188
101.03%


2625
11.039063
11.289063
102.26%


2750
11.867188
11.4375
96.38%


2875
12.828125
11.742188
91.53%


3000
13.789063
12.296875
89.18%


3125
14.296875
12.84375
89.84%


3250
14.953125
13.398438
89.60%


3375
15.914063
13.960938
87.73%


3500
17.179688
14.5
84.40%


3625
18.375
15.617188
84.99%


3750
20.210938
17.09375
84.58%


3875
21.710938
18.523438
85.32%


4000
23.351563
20.078125
85.98%


4125
25.40625
21.828125
85.92%


4250
27.570313
23.890625
86.65%


4375
29.164063
25.734375
88.24%


4500
30.84375
27.601563
89.49%


4625
33.34375
29.4375
88.28%


4750
35.914063
31.671875
88.19%


4875
38.601563
35.921875
93.06%


5000
41.414063
40.226563
97.13%


5125
44.359375
44.3125
99.89%


5250
47.46875
48.070313
101.27%


5375
50.8125
50.789063
99.95%


5500
54.460938
54.21875
99.56%


5625
58.46875
59.289063
101.40%


5750
62.835938
65.0625
103.54%


5875
67.28125
70.367188
104.59%


6000
71.679688
74.476563
103.90%


6125
76
80.679688
106.16%


6250
80.242188
85.945313
107.11%


6375
84.4375
91.007813
107.78%


6500
88.921875
95.757813
107.69%


6625
93.523438
100.296875
107.24%


6750
98.359375
106.71875
108.50%


6875
103.523438
113.382813
109.52%


7000
109.085938
121.414063
111.30%


7125
115.039063
129.023438
112.16%


7250
121.359375
136.375
112.37%


7375
128.039063
142.453125
111.26%


7500
135.078125
149.132813
110.40%


7625
142.5
155.109375
108.85%


7750
150.265625
161.429688
107.43%


7875
158.3125
168.820313
106.64%


8000
166.578125
177.875
106.78%


8125
175.015625
184.796875
105.59%


8250
183.59375
190.851563
103.95%


8375
192.28125
202.523438
105.33%


8500
201.03125
213.828125
106.37%


8625
210.234375
224.6875
106.87%


8750
219.6875
235.296875
107.11%


8875
229.421875
243.1875
106.00%


9000
239.875
255.039063
106.32%


9125
250.492188
267.140625
106.65%


9250
261.4375
281.375
107.63%


9375
272.820313
294.53125
107.96%


9500
284.773438
308.015625
108.16%


9625
297.40625
322.054688
108.29%


9750
310.804688
335.515625
107.95%


9875
325.0625
348.09375
107.09%


10000
340.242188
359.851563
105.76%


10125
356.375
373.523438
104.81%


10250
373.453125
386.679688
103.54%


10375
391.429688
400.429688
102.30%


10500
410.234375
415.710938
101.33%


10625
429.726563
431.382813
100.39%


10750
449.796875
447.484375
99.49%


10875
470.289063
460.796875
97.98%


11000
491.070313
490.734375
99.93%


11125
511.992188
506.734375
98.97%


11250
511.992188
511.992188
100.00%


11375
511.992188
511.992188
100.00%


11500
511.992188
511.992188
100.00%


11625
511.992188
511.992188
100.00%


11750
511.992188
511.992188
100.00%


11875
511.992188
511.992188
100.00%


12000
511.992188
511.992188
100.00%

[joecar]

The shape of the MAF(6) table looks correct (polynomial, consistenty increasing slope, no jerks/jolts).

BEN(6) is looking very good (dead on 1.00) during both 3rd and 4th gear pulls (engine is loaded differently in each gear).

Good job :cheers:

[wcj]

Joecar - Thanks a million. Wouldnt have made it this far without your help.

It makes so much more sense when someone is able to confirm the changes. Plus having my pid file fixed didnt hurt either :)

Now that Im fairly squared away with the MAF and VE, I need to work on the occasional stalling I get when I come to a stop. Doesnt happen all the time, but it is definitely more frequent with the AC on. I'll look over the idle tutorial and give that a go.

[relentless addiction]

can this be done w/o having factory narrowbands installed just using my lc-1

[joecar]

can this be done w/o having factory narrowbands installed just using my lc-1Yes.