PCM Tuning Process Flow=
<=
o:p>
<=
span
style=3D'mso-list:Ignore'>1.&n=
bsp;
DISABLE ALL TORQUE MANAGEMENT – Th=
is
will eliminate all torque management within the PCM. Ignore this step for a
standard transmission (M6) and continue to step 2.
A. Open the VCM Editor>Edit=
>Transmission>Torque
Management
B. Set Abuse Mode Enable =3D F=
alse
C. Set Abuse Mode RPM, Abuse M=
ode
TPS and Abuse Mode Speed =3D 0
D. Select>Abuse Mode Torque
Reduction vs. RPM. Set all values =3D 0 <=
/b>
<=
o:p>
<=
span
style=3D'mso-list:Ignore'>2.&n=
bsp;
LTFT TUNING –
A. In the VCM Editor>Edit&g=
t;Engine
Diagnostics>General>MAF Sensor Fail Frequency =3D 0. This will set a =
P0103
code and turn on the SES light. Ensure that the P0103 DTC is enabled and you
are seeing P0103 in the DTC list. Don’t worry about the DTC at this t=
ime.
B. In the VCM Editor>Edit>Engine&=
gt;Spark
Advance>Main Spark vs. Airflow vs. RPM Open Throttle/Moving. Copy the Hi=
gh Octane
table to the Low Octane table. The computer reverts to the low octane table
when a MAF failure is indicated, this will assure optimal timing.
C. Start the VCM scanner>Hi=
stogram
display. File>Connect. Then Tools>VCM Controls>Fuel & Spark>=
;Fuel
Trim Learn>Reset Fuel Trims.
D. Changes to the LTFT’s do not take effect immediately – t=
he PCM
requires about 50 minutes or roughly 100 miles to allow for the PCM to rele=
arn the
fuel curve. Try not to enter PE mode while driving and logging for this
procedure. Log about 30 minutes of driving at many different speeds and
conditions. Try to hit as many cells in the histogram as possible. Stop log=
ging
and save the log. Do NOT turn off the engine until the log is saved or it w=
ill be
lost. Go to VCM Scanner>Histogram display>LTFT's=
.
Open the VCM Editor>Edit>Engine>Airflow>Main VE and select Prim=
ary
VE vs. RPM vs. MAP.
E. The goal is to get ALL LTFT’s between -5 and +5. Positive LTFT's indicate fuel is being added because of a lean
condition. Richen this cell by increasing the VE table value by the amount =
of
the LTFT value. The operation is opposite for negative LTFT's.
If LTFT =3D (4), VE cell value is 67, result would be (67)+(4)=3D71
- increasing the VE, which is adding fuel. If the LTFT was (-4), the result
would be (67)+(-4)=3D63, decreasing VE and thus =
reducing
fuel. To decrease LTFT values, a smaller number or number closer to zero, A=
DD the
difference between the positive LTFT value and zero to the corresponding ce=
ll
in the VCM Editor>Edit>Engine>Airflow>Main VE>Primary VE vs.=
RPM
vs. MAP table. To increase a LTFT value, a larger number or number farther =
away
from zero, SUBTRACT the difference between the LTFT value and zero and SUBT=
RACT
from the corresponding cell in the VCM Editor>Edit>Engine>Airflow&=
gt;Main
VE>Primary VE vs. RPM vs. MAP table. For example, In the VCM Scanner>=
Histogram
display, the (.8, 40) cell, 800 RPM's and 40 kPa, is 4. To bring the VCM Scanner>Histogram
display>LTFT cell (.8, 4.0) DOWN to 0 from 4 ADD 4 to the (.8, 4.0) cell=
in
the VCM Editor>Edit>Engine>Airflow>Main VE>Primary VE vs. RP=
M vs.
MAP table. If the VCM Scanner>Histogram display>LTFT cell (2.0, 30) is
-10, SUBTRACT 10 from the (2000, 30) cell in the VCM Editor>Edit>Engi=
ne>Airflow>Main
VE>Primary VE vs. RPM vs. MAP table to bring it UP to 0. This will not w=
ork
out exactly but will be VERY CLOSE.
F. Repeat steps D-F until ALL values in=
the
VCM Scanner>Histogram display>LTFT are between -5 and +5. Try to
complete this on the same day for best results as LTFT values can vary +-4%=
per
day.
G. Once all values are between=
-5
and +5, look at the VCM Editor>Edit>Engine>Airflow>Main
VE>Primary VE vs. RPM vs. MAP>3D Surface graph. If the 3D Surface gra=
ph
looks choppy, click on polynomial smoothing ONCE. This will smooth out the
table values and provide a crisper throttle response. The table can also be
hand smoothed using the 3D graph. Look for spikes in the table and
increase/decrease the cells around the spike, creating a smooth table. Now
rescan, and go back to step E.
=
3.&n=
bsp;
WOT PE TUNING<=
/span> – Do=
this
only AFTER all LTFT's are -5 to +5. This method=
uses
the stock narrow band oxygen sensors which are not accurate for this type of
tuning.
A. Open the VCM scanner, do not
worry about resetting the fuel trims they should be learned at this point. =
If
not, it takes roughly 100 miles or 50 minutes of driving to set the LTFT's.
B. Open the VCM Scanner>His=
togram
display and do a nice 0-70 or preferable 0-100mph run. Look at knock retard
FIRST. If knock retard is present, skip to section 4. If knock retard =
is not
present, continue to the step C.
C. Open the VCM Scanner>His=
togram
display>Air/Fuel tab and look at the 100(kPa=
) row.
Most cars seem to like narrow band oxygen sensor reading between 890mv - 90=
0mv.
D. For example, at 100(kPa), 3200(RPM) the narrow band oxygen sensors are at
950mv. We want to bring that down to 890mv. Go to the VCM Editor Engin=
e>Fuel>Power
Enrich, PE Enrichment>V8 Mult vs. RPM. =
Make
sure Plus and
Selected are bubbled in. In thi=
s case
the narrow band oxygen sensor is reading rich, so bring it down by SUBTRACT=
ING
.01. NOTICE THE DECIMAL!!!! VERY IMPORTANT!!! If lean, BELOW 890mv then ADD=
.01
at a time. This is a small increment but we do not want to hurt the motor. =
E. After making the changes, g=
o back
to step B and repeat until the oxygen sensors are in the 890mv to 900mv ran=
ge.
4.&n=
bsp;
ELIMINATING KNOCK RETARD -
A. In the VCM Scanner>Histo=
gram
display>Retard, look for ANY knock retard. For example, cell (4.0, .20) =
shows
4 degrees of knock retard. This should be 0, so SUBTRACT 4 from the VCM Edi=
tor>Edit>Engine>Spark
Advance>Main Spark vs. Airflow vs. RPM Open Throttle/Moving>High Octa=
ne (4000,
.20) cell. Values cannot be less than zero in this table. =
B. In the VCM Editor>Edit>Engine&=
gt;Spark
Advance>Main Spark vs. Airflow vs. RPM Open Throttle/Moving>High Octa=
ne, go
to the (4000, .20) cell AND/OR whatever other cells that have knock retard =
and
SUBTRACT the amount of knock retard that is present in the Histogram displa=
y from
the value that is in the corresponding cell in the VCM
Editor>Edit>Engine>Spark Advance>Main Spark vs. Airflow vs. RPM
Open Throttle/Moving>High Octane table. Subtract by simply clicking on t=
he Plus selection and in the box type=
-4 or
whatever number you have to subtract by and click commit.
C. Scan again and verify NO kn=
ock retard
is present. If still present, repeat from step A. <=
/p>
=
5.&n=
bsp;
A4 TRANSMISSION SETTINGS -
A. Ensure all Torque Managemen=
t is
disabled. If not, see Section 1.
B. Open the VCM
Editor>Edit>Transmission>A4 Shift Speed. Set WOT Shift Enable %TPS=
=3D
90.
C. Set WOT Shift Disable %TPS =
=3D WOT
Shift Enable %TPS-10 or 80 if you used the parameter in step B.
D. Look at VCM
Editor>Edit>Transmission>WOT Shift RPM vs. Shift. Set these =
table
parameters to the desired WOT shift RPM for each gear. Keep in mind there i=
s a
slight delay at the shift point that will cause the engine to exceed these =
RPM
settings. Ensure the VCM Editor>Edit>Engine>Fuel Control>Fuel
Cutoff, DFCO>RPM Limits>P/N Cutoff RPM is roughly 500 RPM higher than
these settings. We don’t want to hit the rev limiter during the WOT
shift. Set Normal<=
/st1:City>,
Performance, and Hot tables to the same parameters.
E. VCM
Editor>Edit>Transmission>WOT Shift Speed vs. Shift--PLEASE PROVIDE=
ME
WITH A GOOD LINK FOR THIS. I KNOW THERE ARE TABLES/CALCULATORS PER GE=
AR
OUT THERE.
F. VCM
Editor>Edit>Transmission>Shift Speed vs. %TPS vs. Shift =3D Leave =
stock
parameters.
G. Go to VCM
Editor>Edit>Transmission>A4 Shift Properties>Desired Shift Time=
vs.
Torque>Normal.
I basically guessed here, and could use some input. For the first half=
of
the torque band, I set shift time to .500 so you get nice soft, smooth
shifts. Starting about midway, I decreased to .250 and for last 1/4 I
changed to .100. I heard you do not want to go below .100 or else you =
will
run into some kind of gear crossing? Please feel free to fill in here.=
H. Go to VCM
Editor>Edit>Transmission>Base Shift Pressure vs. Torque vs. Gear. =
Okay
this is kind of weird and I don't understand it, but what I PERSONALLY did =
was
again take half of the chart and to the left. Take this and set to a =
LOW
# like 10. I have a shift kit in my car, and setting first half gives=
me
nice smooth shifts. You would never know I had a shift kit or torque conver=
ter
in my car. I then took the middle and started beefing up shifts in
increments of 10 then increments of 15. By far right of table I have
shift pressure up to 96. Now, when you drive my car at 0-1/4 throttle=
it
is SMOOTH. 1/2 throttle, you can feel a nice crisp shift. WOT it chir=
ps
tires from 1-2 and 2-3 shifts nice and hard.
I. Upshift/Downshift pressure modifiers - I don't unders=
tand.
If you do, teach me and I'll update.
J. If you have a shift kit, le=
ave
max line pressure at 90. If not you can probably set to 100.
=
The Basics <=
/o:p>
=
&nbs=
p; &=
nbsp; IDEAL
=
; <=
/span>(Rich) =
&nb=
sp; (Lean)
Base (ltrims) &=
nbsp; -&nbs=
p; 0 &nbs=
p; +
=
=
; <=
/span>(Lean) =
&nb=
sp; (Rich)
WOT (O2's) .800 .880-.890 .900=
=
=
; (Rich) =
; (Lean)<=
/o:p>
A/F (WOT) =
12.5 12.8 or .9 13.3
=
A/F (non-WOT) PCM wi=
ll try to maintain 14.7:1
Other Important Values
IAC (idle) =
:
30-50
IAC (load)&nbs=
p; : <120
Injector Duty Cycle : See the chart below<=
/o:p>
Knock Retard
: 0<=
/p>
LTFT’s=
=
: -5 to +5
MAF Flow =
:
Timing =
:
26-28 WOT, 15-18 idle
WOT 02 =
: .880-.890 for narrow band sensors=
Injector Duty Cyc=
le Chart
- The general rule is to not=
exceed
80% duty cycle.
|
|
Duty Cycle
|
|
|
|
10%
|
20%
|
30%
|
40%
|
50%
|
60%
|
70%
|
80%
|
90%
|
100%
|
|
RPMs
|
|
|
3000
|
4
ms
|
8
ms
|
12
ms
|
16
ms
|
20
ms
|
24
ms
|
28
ms
|
32
ms
|
36
ms
|
40
ms
|
|
3500
|
3
ms
|
7
ms
|
10
ms
|
14
ms
|
17
ms
|
21
ms
|
24
ms
|
27
ms
|
31
ms
|
34
ms
|
|
4000
|
3
ms
|
6
ms
|
9
ms
|
12
ms
|
15
ms
|
18
ms
|
21
ms
|
24
ms
|
27
ms
|
30
ms
|
|
4500
|
3
ms
|
5
ms
|
8
ms
|
11
ms
|
13
ms
|
16
ms
|
19
ms
|
21
ms
|
24
ms
|
27
ms
|
|
5000
|
2
ms
|
5
ms
|
7
ms
|
10
ms
|
12
ms
|
14
ms
|
17
ms
|
19
ms
|
22
ms
|
24
ms
|
|
5500
|
2
ms
|
4
ms
|
7
ms
|
9
ms
|
11
ms
|
13
ms
|
15
ms
|
17
ms
|
20
ms
|
22
ms
|
|
6000
|
2
ms
|
4
ms
|
6
ms
|
8
ms
|
10
ms
|
12
ms
|
14
ms
|
16
ms
|
18
ms
|
20
ms
|
|
6500
|
2
ms
|
4
ms
|
6
ms
|
7
ms
|
9
ms
|
11
ms
|
13
ms
|
15
ms
|
17
ms
|
18
ms
|
|
7000
|
2
ms
|
3
ms
|
5
ms
|
7
ms
|
9
ms
|
10
ms
|
12
ms
|
14
ms
|
15
ms
|
17
ms
|
|
7500
|
2
ms
|
3
ms
|
5
ms
|
6
ms
|
8
ms
|
10
ms
|
11
ms
|
13
ms
|
14
ms
|
16
ms
|
|
8000
|
2
ms
|
3
ms
|
5
ms
|
6
ms
|
8
ms
|
9
ms
|
11
ms
|
12
ms
|
14
ms
|
15
ms
|
Fuel Trim Cell Info
0 : Non load&nb=
sp; (? Coast or idle)
1 : Non load&nb=
sp;
2 : Non load&nb=
sp;
3 : Non load&nb=
sp;
4 : Non load&nb=
sp;
5 : Non load&nb=
sp;
6 : Load &=
nbsp; Part throttle
<=
span
class=3DGramE>7 : Load &=
nbsp; Part throttle
<=
span
class=3DGramE>8 : Load &=
nbsp; Part throttle
9 : Load &=
nbsp; Part throttle
10 : Load &=
nbsp; Part throttle
11 : Load &=
nbsp; Part throttle
12 : Load &=
nbsp; Part throttle
<=
span
class=3DGramE>13 : Load &=
nbsp; Part throttle
<=
span
class=3DGramE>14 : Load &=
nbsp; Part throttle
<=
span
class=3DGramE>15 : WOT &n=
bsp;
16 : Idle &=
nbsp; PARK, A4
17 : Idle &=
nbsp; NEUTRAL, A4
18 : Idle &=
nbsp; Engine warm?, AC on, M6
19 : Idle &=
nbsp; Engine warm?, AC off
20 : Idle &=
nbsp; Engine Cold?, AC off
=
21 : Non load&nb=
sp; (? deceleration)
22 : WOT &n=
bsp;
Non load cells 0-5,=
21
Idle 17-20
Part Throttle 6-14
WOT 15, 22
AFR Info
H=
ere is a
chart that I have that should give you some insight into what standards are=
for
lean cruise etc...
9.0:1 BLACK SMOKE (NO POWER)
11.5:1 RICH BEST TORQUE @ WOT
12.2:1 SAFE BEST POWER @ WOT
13.3:1 LEAN BEST TORQUE @ WOT
14.6:1 STOCHIMETRIC AFR ( CHEMICALLY CORRECT )
15.5:1 LEAN CRUISE
16.5:1 BEST FUEL ECONOMY
18.0:1 CARBURETED LEAN LIMIT
22.0:1 EFI LEAN LIMIT
=
PE Tuning Info=
Naturally
Aspirated
Air/fuel ratio for peak p=
ower
is 12.8. If tuning on a Dynojet, try for 12.2 &=
#8211;
12.5 air/fuel ratio. If tuning on a Mustang dyno or
on the street try for 12.8 – 12.9. Some tuners say that LS motors run=
the
best at 13.1. The AFR curve should be rich up to max torque then lean out
slightly up to maximum RPM and then go a little rich a few hundred RPM’s beyond maximum for safety.
PE Delay
If the RPM is below the delay RPM def=
ined
it will blend in PE at the PE enrichment rate. That’s why many people=
set
the enrichment rate to 1. With a lower RPM delay PE will apply immediately
above the RPM specified and full PE will activate at the RPM specified. &nb=
sp;
Normal practice for automatic transmissions is to set that RPM limit to the
stall speed of the converter. For manual transmissions set it a little less
than the take-off RPM. An enrichment rate of 1 effectively negates the
delay RPM. For cars that experience tip-in knock retard at WOT this is
often the solution.
<=
/p>
VE Tuning Info=
Higher
VE values add fuel (tellin=
g the
PCM you have more useable oxygen in the cylinder)
Lower VE values
subtract fuel (telling the PCM you have less useable oxygen in the cylinder=
)
If using a MAF sensor, only tune the =
idle
and part throttle areas of the VE table, 4000rpm and below. If not using a =
MAF (Speed
Density), tune the entire operating area of the table.
A rough VE table will be more suscept=
ible
to burst knock retard.
1. I personally smooth =
each
time, but I don't think it's necessary. I don't stick with the raw polynomi=
al
results, however. I have a spreadsheet which compares the poly value to the
range the value should be within to stay within my AFR range. I don't let t=
he value
fall outside of these bounds.
3. The adjoining cells
should be smooth not spikey=
.
If it's a dip, it should look like a U, not a V. You should tweak the spike=
and
the values around it to smooth it out.
4. I'm guessing that the
max VE cell value you'll see for a stock vehicle would be ~100-110. I think
I've seen VE tables from FI cars which are in the 150+ range.
its kinda like this. from the =
factory
the ve table is the backup controller, the maf is the primary measuring device for airflow, the =
maf reading is double checked against the ve table to make sure nobodies gone crazy. Now in a p=
erfect
world you VE table would be perfectly tuned to match your engine then you p=
lug
the MAF in and perfectly tune it to match your VE table.
Now for the imperfect world most of us live in.
Log your rpm, mass air flow g/sec (use imperial pounds if you must), mass a=
ir
flow frequecny, TP, dynamic cylinder air and fu=
el
trims, save the logs
Now what you will have to do is open the log in excel and figure out what f=
requency
areas of the maf need massaging to bring your f=
uel
trims in line, keep in mind that changing the VE alone will not make much if
any change to the fuel trim but changing the MAF flow VS frequency even 1%
makes big changes. Also keep in mind that if you change maf
without changing VE you can start setting codes and getting flat spots and =
bogs
in throttle response.
confused yet?
Its not that hard, histogram shows which VE cell=
, if
you look at raw data in excel you will easily find the MAF frequency that w=
as
in play when that particular cell was in use. My rule of thumb is if I add =
1%
to a MAF frequency range I will add double (2% in this example) in the VE c=
ells
in that range.
VE and Burst Knock<=
/i>
Once the VE table is correct, tune out
any detected burst knock by increasing the Edit>Engine>Spark
Retard>Burst KR Enable Delta Cyl Air Thresho=
ld vs.
RPM table. The ultimate measure is whatever it takes to eliminate the error
between commanded and measured AFR.
SD Tuning, LTFT’s and MAF Ta=
ble
Scaling
scale the VCM
Editor>Edit>Engine>Airflow>MAF Calibration>MAF Airflow vs.
Output Frequency table positive to get the LTFT’=
s
back to where they were when it was MAFless (SD=
). The
point is to get an accurate VE table and then adjust the MAF calibration ta=
ble
to agree with the VE table at the observed LTFT values.
MAF Sensor Info
MAF Tuning – In Work&nbs=
p; <=
/b>
1.)
Make sure you log Dynamic Airflow vs MAF Freque=
ncy
(Hz) In HP Tuners it is measured in lb/min so we will have to convert this
later for the MAF table (g/sec).
2.) Go do enough driving to log a variety of MAF frequencies. You probably
won't get a whole lot of data above 10,000 Hz or below 2000 Hz, but get as =
much
as you can. Cruising on the highway is a good place for this as you can cov=
er
all rpms and a wide range of mph.
3.) Save the log run and export the data into an Excel readable format and =
sort
the data by MAF frequency (smallest to largest).
4.) Section off MAF frequency ranges that register with the frequency point=
s on
the MAF table (i.e. For MAF table freq =3D 3000, you want to use the data y=
ou
logged in the range of 2940 to 3065) Take the average of all the Dynamic
Airflow data in this range. The reason you want to use this range is so that
the average is calculated using a sort of "swing error" that
straddles the calibration point itself.
5.) Once you have calculated averages for each range (this will be very ted=
ious
and take quite a bit of time, but using excel functions makes it much easie=
r)
you will have new MAF Airflow data to rebuild the table with. 1 lb/min is e=
qual
to 7.58 g/sec so do that calculation and you will derive a new MAF table.
6.) For all the calibration points that you were missing data for (above 10K
Hz, below 2K Hz) you can either shoot in the dark and scale up accordingly,=
or
if you choose to log raw MAF air readings in tandem with Dynamic Airflow and
frequency, you can calculate the variance b/w your dynamic airflow and MAF
airflow and scale up by the trends you see on either extreme. (i.e. If as y=
ou
get closer to 10K and you notice the dyn airflo=
w is
10% higher than the stock MAF airflow, then you can go ahead and
"assume" that above 10K Hz it will most likely behave the same,
otherwise you could try to log 155+ mph runs)
I know this seems like a very painstaking way to do this but it worked very
well for me and using a lot of excel functions I was able to reduce the tim=
e on
this project significantly. I have verified that my MAF reports almost exac=
tly
what dynamic airflow the motor sees and therefore does not cause any confus=
ion
for the trims (this other bit of business in this thread is not MAF related=
).
Some people have tried using scatter plot functions to derive equations that
will supersede any manual calculation, but having to "best fit" t=
he
curve for the logged data leaves an element for bias and human error. Manual
calculation appears to be the most error free method that I can think of.
Hopefully someone can come up with an easier way!
LS1 VCM Main Airmass Calculation
The VE table looks as though it is in
meters cubed; it is just not used like a conventional VE table. The VE valu=
es
are such that the PCM can directly back calculate to g/cyl,
the primary means to determine fueling and timing.
This is what makes it so confusing. You can't solve for air mass, you have =
to
solve for g/cyl.
The equation
VE =3D ((mass flow * IAT / (MAP * RPM * Displacement))<=
/span>
Mass flow: grams/sec
IAT: Degrees Kelvin
MAP: Bar
RPM: RPM
Displacement: Cubic Meters
To solve for the mass flow in g/sec simply re-arrange the equation.
Mass flow =3D (VE * MAP * RPM * Displacement) / IAT
A very important calculation the VCM must make to ensure correct fuel mixtu=
res
under all driving conditions is the dynamic air mass calculation. This is t=
he
main calculation the VCM uses to determine how much air it should use for t=
he
Base Fuel Calculation (Inj PW). The air mass
calculation is not simple and uses various combo's of MAF and SD inputs
depending on engine operating conditions such as current engine RPM. It may
also make decisions based on whether the engine is in a steady load state
(steady MAP) or unsteady load state (Unsteady MAP). Note, that these thresh=
olds
change with RPM and MAP. E.g. at higher RPM or high MAP readings you have m=
ore
leeway before the VCM decides that you have an Unsteady MAP condition. This
unsteady MAP definition is basically there to decide if a throttle transient
has occurred (or other) i.e. the MAF input is known not to be accurate under
these operating conditions.
Under normal conditions (i.e. all sensors working properly) in the code I&#=
8217;m
looking at it is like this:
(caveat: many of these thresholds may vary between code revs and vehicle ty=
pe)
RPM > 4000
----------
trust MAF completely and ignore SD calcs
(apart from MAF sanity checking purposes)
RPM < 4000
----------
if RPM < 2400 and MAP < 84 kPa then
Steady MAP threshold =3D 0.0 kPa
else
Steady MAP threshold =3D 0.8 kPa
If (Steady MAP) then
Calculate MAF Air mass/SD Air mass ratio (used for
Unsteady MAP operation)
Correction Air mass =3D MAF Air mass (filtered) <=
br>
else
Correction Air mass =3D SD Air mass x MAF/SD Air =
mass
Ratio (calculated during Steady MAP conditions)
Transient Corrected Air mass =3D previous Final Air mass + prop=
ortion
of Correction Air mass
Final Air mass =3D fn(MAF Airflow, previous MAF Airflow, prev 3 MAP readings, prev=
3 TPS
readings,
Transient Corrected Air mass)
There are 9 coefficients to this filter (and a total of up to 16
different sets of coefficients depending on operating conditions). It is wo=
rth
noting that the previous value is weighted heaviest followed by the 2 MAF
terms, so MAF dominates IMHO).
There are also a number of checks at the end to make sure thing=
s do
not exceed certain limits.
To summarize:
1. High RPM behavior is totally based on MAF
2. Mid RPM behavior has an allowance on Steady MAP behavior before it switc=
hes
to Unsteady MAP
3. Low RPM behavior (where the bulk of the fuel cells are) is dictated by
unsteady MAP behavior that is still mostly dominated by the MAF input with
small tweaking from SD)
If the VCM decides that a throttle transient has occurred (unsteady MAP), t=
he
airflow incorporates a "correction" from the SD calculations. This
value is the SD calculated air mass multiplied by the previous ratio of
measured MAF air mass to calculated SD air mass (this normalizes it, since =
you
are worried about the transient deviation from MAF air mass only). The way I
understand it is this, imagine you are cruising (MAP is steady), you have a
decent vacuum in the manifold and you change the throttle position quickly.=
Air
rushes into the manifold to service the change in air demands from the engi=
ne
itself but also to try and fill the vacuum. The result is that the MAF reads
higher than it should at that point in time (spikes). This is more pronounc=
ed
at low RPM where the engine airflow is smaller and the relative proportion =
of extra
airflow due to filling vacuum is higher, also the MAF is known to be more
inaccurate at lower RPM and more non uniform airflow. IMHO, the SD correcti=
on
is to account for filling and emptying of the manifold during throttle
transients and also to smooth the MAF's spikine=
ss at
lower RPM’s.
The bottom line is that if the engine is at a steady load state or operatin=
g at
high RPM then the airflow is 100% based on the MAF once you get thru all the
filters and calculations. And the SD calcs only=
get
used for transients and smoothing lower speed operation.
At no stage does the MAF get ignored completely in these calcs
(the dominating terms of the main filter calc are always MAF based).
An interesting point to note is that removing the MAF basically bypasses the
whole system and directly sets the Final Air mass value to be the result of=
the
SD lookup (it also disables things like knock learn and a few other nice
things). Most have taken to calling this "Backup SD Mode" which i=
s as
good a name as any I guess and I meant to allow the engine to run with a fa=
iled
MAF (although it is quite possible to tune using this mode (e.g. HSV GTS 30=
0kw
comes factory this way). There is another way to disable the MAF system
completely (i.e. without setting the DTC's) and=
by
tuning of the thresholds and other flags you can get a fully functional SD =
tune
happening, the so called "True Speed Density Mode".
LTFT and STFT Info<=
o:p>
STFT Info
The PCM uses the Short Term Fuel Trim
(STFT) for real time fueling corrections. They have a 10% window that will =
only
affect the Long Term Fuel Trim (LTFT) if the fueling correction is beyond t=
he
10% window for 10 seconds. If this condition exists, the LTFT is changed and
acted upon during the next PCM timed cycle, which is every 10 minutes.
LTFT Info
The
purpose of LTFT’s is to compensate for en=
gine
and sensor age and variation over time.
+
LTRIMS, PCM is sensing a lean condition and adding extra fuel
- LTRIMS,
PCM is sensing a rich condition and subtracting fuel<=
/p>
Positive long term fuel t=
rims
(LTFT's) indicate the PCM is commanding more fu=
el in
order to compensate for what is being perceived as a lean condition. By doi=
ng
so it maintains a stoichiometric air-fuel ratio of 14.7:1 during closed loop
operation. That is, less than wide open throttle.
Negative LTFT's indicate the PCM is commanding =
less
fuel due to what is perceived as a rich condition although the air-fuel rat=
io
remains the same at 14.7:1 and in reality it's neither truly lean nor rich.=
However, positive LTFT's can result in a rich a=
ir-fuel
ratio during WOT operation because the PCM will add fuel in addition to the=
PE
table. If they were negative no fuel is subtracted during WOT as that could
result in a lean condition resulting in detonation.
If the LTFT's are positive, often resulting in =
a rich
AFR during WOT, horsepower may be gained by getting the LTFT's
as close to zero or slightly negative during closed loop. In this way no
additional fuel is added during WOT. If the LTFT's
are negative no additional horsepower can be gained because the AFR at WOT =
is
then determined solely by the power enrichment table without any additional
fuel being added except during CAT over temp conditions.<=
/p>
To put it simply what's
happening is the Mass Airflow Sensor (MAF) is calibrated to expect outside =
air
temperatures and temperature change rates within a specified range. When you
add an aftermarket intake you often get colder and therefore denser, more
oxygen rich air, than what the MAF is calibrated for. It interprets this as
more air than expected when in reality there really isn't. In turn the PCM
tells the injectors to stay open for a longer period of time by lengthening=
the
Injector Pulse Width in order to inject more fuel into the combustion chamb=
ers.
By doing so the air-fuel ratio remains at 14.7 parts of air for every one p=
art
fuel. Hence it is neither lean nor rich but rather right where it's suppose=
d to
be.
But, when you go WOT the PCM remembers it had to add additional fuel during
closed loop and adds this extra fuel in addition to a predetermined amount
called for in the Power Enrichment vs. RPM table. This results in a too rich
condition at WOT and a loss of horsepower. Or to put it another way: not as
much H.P. as you can obtain should that extra fuel not have been added.
For this reason you want the LTFT's as close to=
zero
or slightly negative during closed loop so no extra fuel is added during WO=
T.
You do this by using scanning software and a program such as LS1 Edit, etc.=
to
get the LTFT's correct. Once they are you can t=
hen
tune WOT using a wideband O2 meter and typically adjust the PE vs. RPM table
for the AFR you want. Note: some applications such as nitrous or forced
induction cars usually require a richer AFR than a normally aspirated car. =
What is closed loop you ask? Closed loop operation means the front O2 senso=
rs
(forward of the catalytic converters) are used to help determine the AFR and
offer feedback to the PCM as to the current AFR. The PCM then adjusts the
injector pulse rate to maintain a 14.7:1 AFR. So it's just that, a closed
feedback loop.
What does open loop operation mean? Well, instead of using a closed feedback
loop (the O2 sensors are not used for input) the PCM uses a lookup table th=
at,
to put it simply, is just a table that says "at this RPM use X amount =
of
fuel." This is called the PE vs. RPM table or "Power Enrichment v=
s.
RPM" table.
<=
/p>
Deleting Rear Oxygen
Sensor’s
Driver Side Codes:
137 - HO2S Circuit Low Voltage Bank 1 Sensor 2
138 - HO2S Circuit High Voltage Bank 1 Sensor 2
140 - HO2S Circuit Insufficient Activity Bank 1 Sensor 2
141 - HO2S Heater Performance Bank 1 Sensor 2
Passenger Side Codes:
157 - HO2S Circuit Low Voltage Bank 2 Sensor 2
158 - HO2S Circuit High Voltage Bank 2 Sensor 2
160 - HO2S Circuit Insufficient Activity Bank 2 Sensor 2
161 - HO2S Heater Performance Bank 2 Sensor 2
Go to Edit>Engine
Diagnostic’s>DTC’s>Error Mode =3D=
3 for
each of the codes above.
Go to Edit>Engine
Diagnostic’s>DTC’s>SES Enable=
d =3D Off for each of the codes above.
Idle Info
General Operation<=
/b>
The PCM calculates the IAC position based on a number of Airflow calculatio=
ns
and estimations, the final idle airflow value consists of the following two
main components:
&=
nbsp; -
Base Idle Airflow (Base + LTIT)
&=
nbsp; -
Adaptive Airflow correction (STIT)
&=
nbsp; LTIT
=3D Long Term Idle Trim
&=
nbsp; STIT
=3D Short Term Idle Trim
Note: These names are made up=
to
more easily understand than proportional, integrator, <=
span
style=3D'mso-spacerun:yes'> slow filtered idle airflow, =
etc.
The first thing to realize is that the PCM only runs the adaptive idle cont=
rol
routines when at idle conditions (below certain TPS and MPH limits). Howeve=
r,
the base airflow routine is controlling the IAC position during all driving
conditions (things you are aware of already, such as the Throttle Cracker, =
and
if you set your Base Running Airflow values too high you get cruise control
etc.). The combination of all these components is what I call the Base Idle
Airflow in grams/sec.
Base Idle Airflow
The Base Idle Airflow is combination of looked up values from various tables
within the PCM and also incorporates a Long Term Idle Trim (LTIT) correctio=
n.
This airflow directly controls the IAC position when at non-idle and is the
"starting point" for adaptive idle control. The base airflow cons=
ists
of the sum of the following individual components:
Base Running Airflow
- this is the main Idle Airflow when in PN (A4 only) or Gear (A4 or M6)
- a table vs ECT
Startup Airflow
- additional airflow during engine startup and initial run period (decays to
zero in the first few seconds of engine operation)
- a table and a few delays and decay rates
Startup Spark Retard Airflow
- airflow correction to account for startup spark retard (if used)
Fans On Airflow
- Additional airflow to account for increased engine load during cooling fan
operation
- Two values depending if one fan active or both active.
DFCO Airflow
- used to set IAC position during DFCO
Throttle Cracker Airflow
- additional airflow to open the IAC based on MPH and RPM
- zero during idle conditions
- a table
Throttle Follower Airflow
- controls rate of closing the IAC valve during throttle closure
- zero during idle conditions
- a few tables of initial value and decay rates
Long Term Idle Trim Airflow (LTIT)
- a slow moving correction based on the adaptive idle routines (think LTFT's for fuel)
- the idea of this correction is to bring the Short Term Idle Trims (STIT) =
to
zero
- it has +ve and -ve limits
- a calculated value
AC Airflow
- airflow correction for when the AC is on, this is a torque based calculat=
ion
that estimates how much torque the AC is pulling and calculates an airflow
correction to compensate.
IAC Park Airflow
- airflow used to calculate IAC position when ignition is off and engine not
running
- used in place of all of the above
- a table
Adaptive Idle Control
The whole point of the idle control routines is to maintain the desired Idle
RPM. The PCM therefore needs to "close the loop" and use the Idle=
RPM
error as a feedback to provide this control. The monitoring of the Idle RPM
results in a Short Term Idle Trim (STIT) that provides the fast moving clos=
ed
loop control of the IAC valve. Again here it is very analogous to the STFT's and feedback from the O2 sensors. That’s=
why I
chose these names rather than Proportional, Integral, and Derivate.
Okay, so t=
he
PCM has a Desired Idle RPM it is trying to achieve and it is constantly
measuring the current RPM and calculating an Idle RPM error value. The PCM =
uses
various aggressive and not so aggressive algorithms to control the STIT, to
provide fast convergence (and also stall saver capability) but also reasona=
ble
idle stability.
&=
nbsp; During
all this, the PCM is maintaining a fairly complex state machine of, Are we =
at
idle?, Is the engine transitioning back to idle?=
, etc.
The PCM does remember a few different last known state of the STIT, for example, when you turn on the AC the PCM sto=
ps
updating the "ACoff STIT" and starts
updating the "ACon STIT" (again here =
think
Fuel Trim cells). The idea of this is that when you turn the AC off the PCM=
can
quickly return to the original IAC operating point. For A4 vehicles you also
have the PN/Gear dimension as well.
&=
nbsp; A
good example of the STIT in action is if you have an M6, you have your foot=
on
the brake and you partially let the clutch out and you feel the engine pull
harder to try and maintain the desired idle RPM. If you were logging the IAC
steps or the desired idle airflow you would see it increase. Monitoring the
LTIT and STIT is a very good tool to get your Base Running Airflow values
correct, ensuring your LTITs are not maxing out=
on
the limits and troubleshooting PN/Gear and Fan On/Off stumble etc., especia=
lly
after head/cam install.
Desired Idle Airflow
The net result is that the PCM takes the Base Idle Airflow (including LTIT)=
and
then adds the STIT to come up with a final Desired Idle Airflow (which
generally is available as a PID for logging). Then there is a final step th=
at
takes this airflow value and translates it to the actual IAC valve position=
or
the ETC TPS position. It's basically just a unit’s transformation for=
the
most part and the IAC and the ETC have their own control routines and state
machines that effectively take this idle airflow as an input.
&=
nbsp; The
"Desired Idle airflow part, which is the final "airflow" val=
ue
the idle control routines deliver to set either the IAC motor position or t=
he
ETC position.
&=
nbsp; After
the "Desired Idle Airflow" is calculated, it is then translated i=
nto
an "Effective area" value in square millimeters (mm2).
This is the cross sectional area required to deliver the airflow desired
(taking into account air density and pressure ratio across the throttle/IAC=
).
Now at this point the calculation branches to either IAC or ETC.
&=
nbsp; If
IAC is installed the "Effective Area" is translated into a number=
of
"steps" that delivers this area (a table of IAC Steps vs. Effecti=
ve
Area).
&=
nbsp; If
ETC is installed then there is a single value that translates "Effecti=
ve
Area" into "Desired throttle area percent" units of % area p=
er
mm2. This number is then handed over to the ETC routines that
control the ETC TPS %.
The ETC lo=
gic
is quite simple in that it has two main inputs the Accelerator Pedal Positi=
on (APP%) and the "Desired throttle area percent"
(IAC%). In the ETC code there is a maximum value that the IAC% is clipped a=
t as
a safety check (mostly on the throttle cracker), but the Throttle Cracker,
Follower, Adaptive and various compensation routines are all the same tables
(i.e. the ETC code has nothing extra).
&=
nbsp; Now,
the ETC looks at the APP% and checks if it is 0. If it is not 0 then the
commanded ETC position is a direct function of the APP%+IAC% (accounting for
throttle cracker). If it is zero then the ETC position is controlled via the
IAC% (as you would expect).
&=
nbsp; Once
this ETCDesiredThrottleArea% has been passed th=
ru the
numerous limiter functions (ETC RPMlimiters, MPH
limiter, TorqueMangement etc.) the final ETC
Rotation% is calculated via a simple transfer function of ETC Rotation% vs DesiredThrottleArea%.<=
br>
&=
nbsp; There
are maximum slew rates, minimum position checks and a few other parameters =
here,
but in terms of idle nothing else comes into it.
&=
nbsp; Unless
you have modified your ETC in anyway, there would be no reason to change the
ETC% vs EffectiveArea scaler or anything else there as far as i
can see.
What else?
In addition to the Idle Airflow routines the PCM also has an RPM based idle
spark correction "closed loop" operation that it uses to control =
the
idle RPM. Since the spark advance can move much faster than the IAC, it can
provide very fine control of idle speed. When logging you will see this as a
jagged spark advance chart, most noticeable with cams at lower idle RPM’s where the spark advance generally oscilla=
tes
between its min/max allowed values as the engine lopes at idle.<=
/span>
<=
/p>
Idle Tuning
To set idle speed go to V=
CM
Editor>Edit>Engine>Idle>Idle RPM>Target Idle RPM vs. ECT tab=
le
and change cell values to desired idle RPM. If idle RPM is modified up or d=
own,
the VCM Editor>Edit>Engine>Idle>Base Running Airflow>Idle
Airflow vs. ECT table must be adjusted up or down also.
If a lean idle condition =
is
present modifying VCM Editor>Edit>Engine>Fuel Control>Open &
Closed Loop>Idle Proportional Fuel Tables =3D Off
will most likely eliminate it and any low RPM surging.
A Basic Idle Tuning Strat=
egy
- Set idle speed to desired RPM. 900-950 is good fo=
r cars
with aftermarket cams.
- Go to all your spark tables and set the park and
drive idle values to about 22 degrees.
- Now switch to a scanner that lets you see IAC cou=
nts
and TPS voltage. This is where we will spend some time.
- We want IAC counts to be 40-60 for cars with
aftermarket cams. The stock cam runs about 60-80 counts. A car with an
aftermarket cam will want less IAC counts.
- To reduce IAC counts turn off the car. Turn the i=
dle
set screw clockwise to open the throttle blade a LITTLE. Unplug the TP=
S,
turn the key to the on position, DO NOT sta=
rt the
car, for 30 seconds. Turn the key off and plug TPS back in.=
- Start the car and begin scanning. Monitor the IAC
counts and repeat step 5 until the IAC counts come into line. It will =
take
about 4-6 times to get the IAC counts correct.
Note:
Keep in mind that TPS voltage must be in the .4 to .6 range. If above or be=
low
this voltage, the PCM will fall into the wrong cell at idle. Check the TPS
voltage each time the set screw is adjusted and adjust as necessary.
Another adjustmen=
t that
will help start-up and idle is to go to the cranking VE table and multiply =
the
whole table by about 80% to lower it. Now go to the primary or secondary, i=
f the
PCM has it, VE table and drop about 3-6 counts off the idle area.
<=
/p>
Open Loop Idle Fueling
The VCM
Editor>Edit>Engine>Idle>Base Running Airflow>Idle Airflow vs.
ECT table is like a software choke that is used to control the mixture at i=
dle
when the engine is in open loop. It controls AFR by opening or closing the
throttle blade slightly to meet the values in the cells at a given coolant =
temperature.
Closed Loop
Idle Fueling
For closed loop idle fuel=
ing
the 400, 800, and 1200 rpm cells at the lowest MAP value’s in the VCM
Editor>Edit>Engine>Airflow>Main VE>Primary VE vs. RPM vs. MAP
table has control over idle fueling.
Tuning the VE table in HPTuners - Overview
Wideband and Narrowband methods=
The LTFT value is your indicator of how much error is in your Volumetric
Efficiency table. If Block Learn is at 0% everything is just right. If your
STFT is more than 3 % away from 0 , the LTFT val=
ue is
still "learning". A motor is considered well tuned to have LTFT
values between +- 4%. Not every motor can achieve this though.
Step
by Step for SD tuning Wideband Method:
1: Unplug MAF sensor
2: Disable the SES lights for MAF codes P0101, P0102, P0103 (No check engine
light.) Do not completely disable the codes or the PCM will not fall into SD
mode. Only turn off the SES light, DO NOT DISABLE THE CODES THEMSELVES!
3: Change all points to 1.13 in the Open Loop F/A vs
ECT vs MAP table (commands AFR of 13.0)
4: Change all points in the Closed Loop Enable Coolant Temp vs
IAT table to 250* (Disables closed loop)
5: Copy High Octane table to the Low Octane Table (computer reverts to low
octane table when MAF is unplugged, this assure’=
s
optimal timing)
6: Change all points in the Power Enrich Fuel Multiplier vs
RPM table to 1.0 (disables PE mode)
7: Use your wideband and HPT histogram to verify AFR of 13.0
8: Make adjustments to the VE table accordingly to dial in a 13.0 AFR using=
the
desired formula - current afr/13.0 =3D VE multiplier
example: cell @ 1600 rpm Map 50 is showing the air/fuel to be at
11.7 in the histogram. 11.7/13.0=3D0.9 Lets say that your =
VE
table has a value of 48 listed at 1600 rpm, Map 50. Using the above formula=
you
would multiply that value times (.9). 48*.9=3D43.2. 43.2 would be your new =
VE
value. Continue using this formula until all data has been plotted. <=
br>
9: Hand smooth VE as described by Magnus (a smooth VE results in crisper
throttle response), upload new .bin and repeat steps 7 and 8 until all cell=
s in
the histogram are 12.8 - 13.2.
10: Change all points in Open Loop F/A vs ECT <=
span
class=3DSpellE>vs MAP table back to stock (re-enable stoich
commanded AFR)
11: Change all points in the Closed Loop Enable Coolant Temp vs IAT table back to stock (re-enables closed loop
operation and fuel trim leaning)
12: Change all points in the Power Enrich Fuel Multiplier vs. RPM table bac=
k to
stock (re-enables PE mode)
13: Use wideband and PE table to dial in desired WOT AFR. (Optimum HP at WO=
T)
Step by Step for SD tuning Narrowb=
and
(stock) o2' Method:
1st a couple of things to keep in mind:
a: This method will not be as accurate as the one above, but it should get =
you
real close for part throttle.
b: Please keep in mind that depending on where y=
ou
have your PE enable settings, you will probably only get "0"
LTFT values in your histogram after 4,000 RPM.
c: For WOT you will still need a WBo2.
d: There will always be a 2-4% change in learnin=
g from
day to day. Weather, fuel, and engine dynamics vary quite a bit. It’s=
the
VCM's job to learn these corrections. It =
is recommended
that all tuning sessions be done in like weather conditions.
e: There are two methods of using the LTFT's to
correct the VE table, below is a description of the two taken from the HPTuners help file. I will try to elaborate a l=
ittle
more on them.
1: Unplug MAF (Replace with strait bellows if you do not have intentions of
ever using a MAF again ie. always speed density=
.)
2: Disable the SES lights for MAF codes P0101, P0102, P0103 (No check engine
light.) Do not completely disable the codes or the PCM will not fall into SD
mode. Only turn off the SES light, DO NOT DISABLE THE CODES THEMSELVES!
3: Copy High Octane table to the Low Octane Table (computer reverts to low
octane table when MAF is unplugged, this assure optimal timing)
4:Write your file to the VCM and go for a drive approx 20-30 minutes, and t=
ry
and hit has many cells as possible. Use the "default" config file and scan your LTFT's=
and STFT's into the histogram.
5: After you have logged sufficient data, it is now time to look at what
changes you need to make. This will require looking at both the LTFT and ST=
FT
data on a cell by cell basis.
a: If the majority of your LTFT cells, are great=
er
than a 10% swing (+- 10), then refer to method 1.
b: If the majority of your LTFT cells, fall betw=
een
+-10% then refer to method 2.
Method
1 using the LTFT and STFT corrections from the scanner histogram data:<=
/b>
The VE table corrections can be derived by the following formula:
(+-LTFT%) + (+-STFT%) =3D Final % Learned
If your LTFT histogram value @1600 RPM, 30 kPa =
Map is
+7 and the associated STFT value is -2 your total % fuel correction is +5. =
Use
about ½ the correction (2.5%) on the VE table cell @1600 RPM 30=
kPa Map to avoid overshoot.
Repeat this process for any LTFT histogram value until you see a +- 4%
range. After you reflash your VCM with the new =
VE
table, use the VCM Controls function to reset the fuel trims before you
start recording new data. This will give you a “clean slate” to
work with.
It will take you the better part of an afternoon to log the LTFT, make the
corrections to the VE table, then verify the results on the histogram again.
Method
2 using the STFT corrections only from the scanner histogram data: =
This will require you to disable the LTFT process.
Use of this method requires your VE table to be no more than 10% off. If yo=
ur
VE table is beyond this, start with method 1 first.
Depending on your model you can either turn the LTFT enable option to ̶=
0;Off”
(Edit>Engine>Fuel Control>Open &Closed Loop>Long Term Fuel =
Trim
Enable =3D OFF) or set the LTFT minimum Engine Coolant Temp (ECT) to i=
ts
maximum value. Using the scanner reset your Fuel Trims. Go for a drive
and record the STFT's.
The above step of disabling the LTFT learning process will let the VCM run =
the
STFT correction, but no make any LTFT corrections.
The advantage to this is you only have to use the STFT table on the histogr=
am
to get your fuel % correction values. You no longer have to do the
(+-LTFT %) + (+-STFT %) math or keep track of the LTFT cell number in =
use.
The number derived in the histogram can now be used as a 1:1 ratio in the VE
table. So, if you are showing -7 in your histogram, you could subtrac=
t 7
from the corresponding VE cell, the net result should be a value close=
r to
0 during your next logging session.
This will take you about ½ the time to log the STFT, make the
corrections to the VE table, then verify the results on the histogram again=
.
Once you have achieved a histogram of near 0 values, you can go back into t=
he
editor and re-enable your trims.
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
<=
o:p>
LS1
Tuning Guide
This document will attempt to layout a
step by step process to guide you in the tuning of your ls1 powered vehicle=
Requirements
In order to properly tune your car th=
ere
are a few items that you need to have available
1. HP Tuners
2. Laptop computer<=
/p>
Outline
There are many different ways one can
tune a car, but below I have outlined a methodical process which should work
well.
1: Set absolute parameters
2: Start car, verify
3: Tune idle
4: Tune part throttle
5: Tune WOT
1: Absolute
parameters
At this point the laptop is connected=
to the
car and the base program is downloaded and saved (*always* start with YOUR =
base
program). The number one thing to remember when editing the PCM is to make =
only
one or a few small changes at a time.
First start=
with the
"absolute" parameters. These values can be known and set wi=
th
total objectivity. These generally include gear ratio, fan turn on temps, r=
ev
limiter, etc. Choose a desired idle rpm and enter it initially. The main id=
ea
here is to keep the program as stock as possible - we only want to change
values that have a concrete value - leave shift points, fuel, spark, and ot=
her
subjective values alone.
The only tricky parameter to set here
will be injector constant. This is because the LS1 does not have a manifold
vacuum regulated fuel pressure, so as manifold vacuum changes (and rail
pressure stays constant) the pressure drop across the injector change, so t=
he
injector flow/constant itself changes. If a vacuum referenced regulator is
used, (some supercharger setups, etc.) then simply fill in a constant value
across the range.
2: Start ca=
r,
verify
Now we will start the car and verify =
that
everything is working properly. Before we actually key on we should have our
scanning/logging software hooked up and ready to go. The suggested minimum
parameters to be logged are:
RPM, MAF flow (g/sec or lb/min), MAF
Frequency, MAP, Spark Advance, Knock Retard, Injector Pulse width (left and
right banks), O2 sensors - B1S1 and B2S1, IAC Position, L-trim (left and
right), S-trim (left and right).
In addition to any parameters of spec=
ial
interest, all data should be logged to disk.
Now we are ready to start the car. St=
art
it and begin logging. First verify everything seems mechanically sound (oil
pressure, etc.). Next, look at the MAF flow and/or frequency. As you blip t=
he
throttle this value should increase/change. Let the car run for a while to =
heat
up and go through its DTC tests. Watch out for any kind of SES light. If any
are observed determine if it is a mechanical/electrical problem or if it is
simply caused by a new engine combination (camshaft, etc.). Once the vehicl=
e is
up to operating temperature verify that the O2 sensors are responding to
changes in the throttle. If the vehicle seems to be running decently (no
pinging or potentially problematic situations) drive it around for awhile.
Monitor the STFT’s, =
LTFT’s,
and O2 values and ensure no great splits exist, exhaust leaks, etc. If the =
O2
sensors are old or just aren't switching fast enough (anti-freeze and RTV c=
an
both kill them easily) they should be replaced before continuing any furthe=
r.
3: Idle
Now that we have verified everything =
is
in proper working order we can begin tuning. Idle is the best place to star=
t. Previously
we set the desired idle rpm - subjectively decide if this rpm correct. If n=
ot,
change the value and re-evaluate. Once the desired idle rpm is achieved, we=
can
begin tweaking it for stability. Take note of the IAC counts. In a no load
situation (neutral, no ac) they should be no lower than 30, and no higher t=
han
50. A hole may need to be drilled in the throttle body or enlarge the one t=
hat
is already there to bring down the IAC values. Do this until they are
acceptable. Now put a load on the car (D if automatic, and put the AC on). =
The
values here should be no higher than 120 or so. If they are enlarge the hol=
e.
Repeat the above process until the IAC
values fall inline. If the idle is still unacceptable then try adjusting the
timing. Be careful of adding to much timing - it can give a great no load i=
dle,
but with any kind of load will become erratic. A "=
hunting"
idle is a sure sign of too much timing. To adjust the timing at idle the ba=
se
spark tables are the easiest place. The tables are scaled vs. rpm and g/cyl of airflow. The rpm part is evident. To calculate=
the
g/cyl use the follow=
ing
formula:
g/cyl =3D 15 * MAF(g/sec) / RPM
This formula takes mass flow per unit
time and converts it to mass flow per cylinder. The 15 is a constant which
corresponds to the characteristics of a V8 running a 4-cycle combustion cyc=
le.
The easiest thing to do is to load the log file into excel (export it as a
CSV), then create a formula in excel which applies the formula above.<=
/o:p>
4: Part Thr=
ottle
Not that the idle is correct and we h=
ave
verified that the car is in proper working order we can begin part throttle
tuning. The first step in this is to record a long log file of driving, a
minimum of 20 minutes, but the longer the better.
Once we have a log file we can begin =
the
data reduction. The first element we will tune will be the fuel delivery. At
part throttle the computer uses the MAF meter to find the amount of air
entering the engine. It then calculates the amount of fuel required to main=
tain
a 14.7:1 A/F ratio. It injects this fuel by controlling the injector pulse
width. The O2 sensors, which are very accurate at 14.7:1, provide feedback =
to
the computer and let it know how close it is to the goal. The computer uses
this feedback to tweak the fueling of the motor to achieve a proper 14.7:1 =
a/f
ratio. This "tweaking" is exhibited to through the STFT and LTFT =
parameters.
These values indicate how the computer is correcting. Since injector flow a=
nd
pulse width are known with great precision, and we have no control over the
internal algorithms we will assume that any inaccuracy (which is exhibited =
by
nonzero trim percentages) is a result of an incorrect MAF transfer function=
.
In tuning part throttle we will tweak=
the
MAF transfer function according to the LTFT values we logged. There are 2 w=
ays
of doing this, the simplest is to view the LTFT values, average them, and s=
cale
the entire MAF table by a percentage which will give the LTFT’s
a 0 to -4. LTFT’s are in units of percent=
so
this is easy. If the average LTFT’s are a=
round
+5 and we want to shoot for -4, then we would just multiply the entire tabl=
e by
109% (or an increase of (+5 - (-4))=3D9 percent). Likewise if we were at -1=
0 and
wanted to shoot for negative 4 we would decrease the entire table by 6 perc=
ent,
or multiply by 94%.
Once this is comp=
lete repeat
the logging process above and check the new LTFT value. We want to a=
void
positive LTFT values since they will be applied at WOT and will lead to inc=
onsistent
fueling. Negative values are okay, though we shouldn't go too far out of wh=
ack.
The second option is a little more
complicated. It uses the same premise above, but instead of taking the aver=
age
value it applies a localized LTFT correction to each point of the MAF trans=
fer
function and derives a new curve. This method is not for everyone, but in
certain instances is very useful.
After repeating the above method unti=
l LTFT’s fall in line, fueling should be complete=
. Now
we can address spark. Spark advance is a rather difficult item to tune dire=
ctly,
but here is a suggested method. This method relies on a properly functioning
knock sensor without any desensitization.
Assuming the car is naturally aspirat=
ed
and does not ping with the stock timing advance: Take the entire timing tab=
le
and increase it by 5 degrees. Now start driving the car while logging. Try =
and
emulate every possible driving condition. If pinging is detected at any poi=
nt
back out. If the car pings constantly reduce timing across the board two
degrees.
When done logging export the data to a
CSV file and open in it excel. Here we will make a pivot table. Create a co=
lumn
with g/cyl, spark retard, and rpm. Use these th=
ree
items to make a pivot table. Scale the table with g/cy=
l
on the x axis and rpm on the y axis. Put spark retard in the middle and set
it's mode to average. You should group the axis along the same lines as they
are grouped in the PCM.
We now have a table of the average sp=
ark
retard taken out at each timing point. Now go to the table in the PCM and
subtract 75% of this value from the actual spark advance at each point where
spark retard occurred. Re-log the car. Repeat the procedure until no spark
retard is detected. The timing curve should now be tuned.
If the car is an automatic we will no=
w start
tuning shift pressure, shift points, and TCC.
5: WOT tuni=
ng
The first thing to do is make a quick=
WOT
pass in a low gear (a low load) and check both O2's and knock retard. O2's =
are
NOT accurate or precise at this a/f ratio, but can still be used for a ball=
park
estimate. If they aren't 850-950 we will adjust the PE vs. RPM table accord=
ingly.
This table is where all fueling changes at WOT are made. If knock retard is
present we need to localize it to a point in the timing table, so using the
method above for part throttle tuning, we will do the same thing for WOT
tuning.
If either spark or fuel is changed th=
en
go back and check the other by logging. A wideband O2 sensor is required to
accurately set the fuel map. If wideband feedback is available the a/f rati=
o will
generally end the richest at your torque peak and leaning out from there to
peak horsepower and then a little rich before and after the shift point for
safety.
Once fuel and spark are set then begin
playing with the shift points and transmission parameters automatic cars. I=
f it’s
a manual transmission we are good to go!
Follow-Up
After a week or so you need to re-ver=
ify
all your logged values and ensure they haven't drifted. If they have, repeat
the processes necessary to bring them back in line.