CALC.PID for simultaneous MAF & VE Calibration - need sanity check

[Gelf VXR]

I decided to create a calc pid that will let me calibrate MAF and VE table in one log using wide band, stft & ltfts. It must be more accurate to calibrate MAF and VE table from the same log data to get them to match, IMHO.

I need to get the math right in my head, here goes

Log all the usual pids asscociated with MAF, VE and trims.

Maps are MAP vs RPM as used in VE table, except for MAF cal.

Create map for BEN1 (GM.AIRPERSEC is PCM g/sec value) for refference

Create a pid for the map (BEN1 * GM.AIRPERSEC)/SAE.MAF gives BEN M correction for MAF.

Create a pid for the map (BEN1 * GM.AIRPERSEC)/GM.DYNAIR_LS2) gives BEN D correction for VE table.

How ever fuel trims have not been factored in...

Create a pid for the map (STFT1+LTFT1) = % TRIM

The math

For an example, (and we see that the GM.AIRPERSEC = GM.DYNAIR (The higher value of MAF & VE CALC, it could be vise versa))

BEN1 for GM.AIRPERSEC = 1.04

BEN M for SAE. MAF = 1.02 MAF is reading 2% lean

BEN D for GM.DYNAIR = 1.04 VE TABLE is reading 4% lean

How ever, for the example BEN D for VE table = 1.04, the combined fuel trims of STFT & LTFT is say 1%, this is where i need a sanity check with the calc.pid

The 1% fuel trims have been added and calculated for the BEN1 for GM.AIRPERSEC = 1.04, I need to still compensate for the 1% fuel trim, ie the BEN D should be 1.04 + 0.01 = 1.05 for the VE table calibration, and BEN M should be 1.02 + 0.01 = 1.03 for the MAF calibration.


My calc.pids

*CLC-00-900
cylair 0.0 1.5 .2 "{SAE.MAF.gps}*15/{SAE.RPM}"# ================================================== ============================
# File details
# ------------
#
# This section defines various details about the file format.

*FILE

#Parameter Value Description
#---------- ---------------- ---------------------------------------------------
VERSION 7.1.1 File version
DECSEP . Decimal separator used in this file



# ================================================== ============================
# Units
# -------------------
# See sae_generic.txt for more information on the *UNITS section

*UNITS

#Code System Abbr Description
#-------- ---------- -------- -------------------------------------------------------------

None None "" "No units"


# ================================================== ============================
# Add slot definitions here
# --------------------------------
# See sae_generic.txt for more information on "SLOT" formats
#
#Units Low High Fmt Expression
#------------ ------------- ------------- ---- --------------------------------------------------------------
*CLC-00-001
% -40 40 .2 "{SAE.SHRTFT1}+{SAE.LONGFT1}"
*CLC-00-002
% -40 40 .2 "{SAE.SHRTFT2}+{SAE.LONGFT2}"
*CLC-00-003
% -40 40 .2 "({SAE.SHRTFT1}+{SAE.LONGFT1}+{SAE.SHRTFT2}+{SAE.L ONGFT2})/2"
*CLC-00-004(
factor 0 2 .2 "(((({GM.AIRPERSEC}*{CALC.BEN1_B})/{SAE.MAF})*100)+{CALC.TRIMS1})/100"
*CLC-00-005
factor 0 2 .2 "(((({GM.AIRPERSEC}*{CALC.BEN1_B})/{GM.DYNAIR_LS2})*100)+{CALC.TRIMS1})/100"
# ================================================== ============================
*PRN - Parameter Reference Numbers
# --------------------------------
# See sae_generic.txt for more information on the *PRN section
#
#Code PRN SLOT Units System Description
#------------------------- ---- ------------ ---------------- ---------------- ------------------------------------------
CALC.TRIMS1 F001 CLC-00-001 % Tuning "ST1 & LTFT1 COMB"
CALC.TRIMS2 F002 CLC-00-002 % Tuning "ST2 & LTFT2 COMB"
CALC.ST_LT_AVE F003 CLC-00-003 % Tuning "ST1&2 & LTFT1&2 AVERAGE"
CALC.MAF_BEN1 F004 CLC-00-004 factor Tuning "MAF BEN1 FROM AIRPERSEC AND TRIMS1"
CALC.DYNAIR_BEN1 F005 CLC-00-005 factor Tuning "DYNAIR BEN1 FROM AIRPERSEC AND TRIMS1"

You can see from the attachments that I have applied this calc pid to, that the MAF BEN map why the PCM uses dynamic fueling at lower rpm’s, its way off..

[Gelf VXR]

additional screen shots

[swingtan]

The biggest problem you will have is knowing what is influencing the final airflow figure for fueling. When running both the MAF and the VE ( IE, below the High Speed Air Flow threshold - 4000 rpm ) I don't think anyone has a complete understanding of the roll that both the MAF and VE table play. The MAF is the primary measurement, but...

• does the VE act as a correction factor or a plain sanity check?
• How much difference does there need to be between the MAF and the VE before any "correction" is performed?

This will make using a single log difficult to tune both the MAF and the VE because it will be difficult to tell if the MAF or the VE needs the correction.

Here's the good news though...

What you would need to do is to alter the HSAF point to be below the normal running RPM, say 400rpm. This puts you in MAF only mode and bypasses the VE corrections. Then zero all LTFT and start logging all parameters.

If you log MAP, RPM, MAF ( FREQ and gm/Sec ), gm/Cyl then create a calc PID for gm/Sec ( taken from RPM and gm/Cyl ), you can create a BEN type MAP from the MAF gm/Sec and the Calculated gm/Sec. Then get the MAF dialed in accurately so you know it's "correct". Now, remember that the gm/Cyl PID is based from the VE table, so this will allow you to correct the VE while still running the MAF. Of course, to do this you will need to ensure the fueling with the MAF is pretty close to start with.

The thing is, that you will always need to reference any correction procedure, to get an accurate result. You can't correct 2 separate parameters with the same data if the variables between them change ( which in this case they do ).

Simon.

[Gelf VXR]

The biggest problem you will have is knowing what is influencing the final airflow figure for fueling. When running both the MAF and the VE ( IE, below the High Speed Air Flow threshold - 4000 rpm ) I don't think anyone has a complete understanding of the roll that both the MAF and VE table play. The MAF is the primary measurement, but...

• does the VE act as a correction factor or a plain sanity check?
• How much difference does there need to be between the MAF and the VE before any "correction" is performed?

This will make using a single log difficult to tune both the MAF and the VE because it will be difficult to tell if the MAF or the VE needs the correction.


Simon.

Hi Simon

I'm not sure that I need to know how the PCM calculated the final g/sec value as long as I know what the final value is, which I am pretty certain is the GM.AIRPERSEC pid.

My above calculations are incorrect, I need correct the AIRPERSEC pid by the combined ST&LT trims to arive at 14.7:1 RATIO when commanded AFR =14.7/1. If the fuel trim is -4%, the g/sec needs to be increased by 4% to correct.

(100+(-4))/100= 0.96

GM.AIRPERSEC/0.96 = Corrected AIRPERSEC

I want the DYNAIR and the MAF g/sec to agree with the PCM (AIRPERSEC) g/sec value, so

The corrected GM.AIRPERSEC divided by GM.DYNAIR to arrive at correction factor for VE table.

The corrected GM.AIRPERSEC divided by SAE.MAF to arrive at correction factor for the MAF table.

Ive just logged 2 and half MB of data and will apply the correction factors as described, re log and see if the GM.AIRPERSEC, GM.DYNAIR and SAE.MAF values come even closer together. I will post my findings.

All the above assumes that the injector flow rate is correct (GM LS7 values for GM 42lb/hr injectors), for which the g/sec pids are calibrated to.

[WeathermanShawn]

Hi Simon

I'm not sure that I need to know how the PCM calculated the final g/sec value as long as I know what the final value is, which I am pretty certain is the GM.AIRPERSEC pid.

Good luck if you can really compute each airflow contribution. You would get the tuning equivalent of the 'Nobel' Prize of scientific accomplishment.

The problem is in your assumptions. It could just be that in a MAF-enabled vehicle the airflow calculations for Trims my be all CYLAIR (my long-standing hypothesis). For example on a LS1, Spark appears to follow CYLAIR 100% (MAF-enabled). Airflow corrections from DYNCYLAIR do not seem to effect spark load. Perhaps closed-loop on MAF cars is also 100% CYLAIR.

I will be interested to see if your calculations can withstand scientific scrutiny. So, far my observations have been all anecdotal. My hypothesis has been that in a MAF-enabled vehicle, the MAF controls almost (if not 100%) of the closed-loop Trimming along with spark. The brief contribution of airflow correction from DYNCYLAIR last for about 1/2 second..then the airflow and fueling goes back to being controlled from the MAF. Thats not long enough to change the LTFT filter sequence..hence the MAF controls a whole lot more than given credit.

On my end I can't get it into a mathematical expression. If you do, it will be the first person to have ever done it. I would just question your assumptions. You might be using the end results thinking you are getting success, but how are you going to prove it. You might just be manipulating the end result to match your assumption. And how do you know if it is correct?

I respect your pursuit. I am just respectfully skeptical..

[Gelf VXR]

I know my assumptions my be wrong, but when performing autove, airpersec tracks dynair, the maf pid is still logged althought disabled. When performing maf cal, airpersec tracks maf pid, the dynair is still logged but ignored. When in cl, airpersec tracks the higher value of the two, so it stands to reason that airpersec is the g/sec value derived between the other two. Time and a few logs will tell

[Gelf VXR]

My pursuit is that all three pids are within a percent of each other

[WeathermanShawn]

When in cl, airpersec tracks the higher value of the two, so it stands to reason that airpersec is the g/cyl value derived between the other two. Time and a few logs will tell

Gelf..thats my main concern. Your making a very 'risky' assumption. It may not be derived at all between the two. How will you be able to make that legitimate interpretation?

Again, I appreciate you taking the effort. I am rooting that you can crack the equations, but don't you actually need to evaluate the computer code to know how the PCM then uses the airflow corrections? What if it ignores the first two airflow corrections, and then takes the third correction (as an average). How would logging reveal that?

Food for thought. This has been a long-standing pursuit. Again, good luck.

[Gelf VXR]

If I can get all three g/sec pids to agree with zero % trims, (positive on the lean bank) I will have achieved my objective. I'm not to worried about the internal calcs as long as the end result is as desired

I'll update on my progress..

[joecar]

You could try to correct the MAF just with a ben based on trims.

And you can calculate the per cylinder VE as follows and apply the same trim-based ben to it:
VE[g*K/kPa] = MAF[g/s] * DAT[K] * 120[s*rev/min] / RPM[rev/min] / MAP[kPa] / N[cyl]

where DAT = dynamic air temperature (blend of IAT and ECT)

[ i.e. in a similar manner to Shawn's LS1 calculated VE tutorial ]


I'm not too familiar with E40... on www.holdencrazy.com which year/models are E40 (are all the LS2's E40's)...?

I do not know what AIRPERSEC is derived from, when I think I might know then I see something that indicates otherwise.