More COS5 questions

[FreshDA9]

Ok so this only happens when flashing the COS5 i was reading it as it would happen with the 2002 OS flash as well.

[joecar]

The 2002 F-car OS should flash straight in with no problems... (you will still have to copy over your 12202088 tables... and make sure it runs on 12212156 at least as good as it did on 12202088... if it doesn't then you have to find out what went wrong).


With COS5/02020005, some of the tables will contain the wrong values after doing the cal-only flash from 12212156...

for example B3605 is replaced by B3647 which has the values that B3605 previously contained (which are wrong for B3647 since it has RPM on the vertical axis instead of ECT);

also, some of the new tables (A00xx), will be out of range (OOR), the Out-Of-Range tab will identify these to you (as will their blue dog-ear corners)... but also note that A0012 and A0013 will also be OOR (these two are not identified to you, but will have blue dog-ear corners)... you have to set all of these to in-range values...

note that the tunetool displays the closest max or min value for an OOR cell (and colors the corners blue)... when you set the OOR cell's value to max or min, the tunetool sees no change, so it does not save your edit... you have to set an in-range value other than max or min, save file, exit tunetool, re-start tunetool and if desired set max or min value on those cells.

Among other things, you have to do this:
- set A0002 to 100%.
- set B3647 such that in the last 2 or 3 columns (high MAP) the commanded fuel is sufficiently rich/safe.

[joecar]

More info: COS5.txt

[joecar]

Contents of COS5.txt:

===================================== COS5 notes =====================================

These notes are extracted from the COS5 thread on the forum.

Start from either of these OS's (with the same transmission as yours):
- 12225074 Australian Holden -> upgrades to COS5 01290005
- 12212156 Camaro/Firebird 2002 -> upgrades to COS5 02020005

NOTE: The latest COS5 calibrations are included when you install EFILive V7 software.


The COS allows adaptable spark (i.e. sliding between HO and LO tables depending on
detection of knock) even when the MAF is failed; to do this these parameters must
be set as follows:
- A0000 set to Disabled,
- A0002 set to 100%.


===================================== SETUP NOTES =====================================

All Custom Tables/Parameters:
OOR (out-of-range) cells have blue dogears in the 4 corners.
Fix all OOR cells (including in A0012, A0013), this is tricky, read the OOR forum threads.
A0012 and A0013 max display value is flaged as OOR, so fix these also.
The Custom tables/parameters use memory areas that were previously undefined and may contain
garbage values, so edit them to sane values in case you ever accidentally enable features.

A0003 TP VE Enable:
Set to disable (can be enabled later if you intend to run the TP VE Table).

A0008 OL ECT Commanded Fuel Multiplier:
Use warmup ECT vs BEN logs to adjust this table (adjust in OL).

A0014 IAT VE Multiplier:
Use seaonsonal IAT vs BEN logs to adjust this table (adjust in OLSD).

N20 monitor is done via the same input as the Valet mode;
once it is grounded then the two compensation tables become active, although,
there are two parameters to stop those two tables even if the N20 is on (A0017, A0018);
timing retard from A0016 will only function when table B5956 has a multiplier value of 1.00.

COS4 was going to be the N20 options only, but in the end it was possible to squeeze
everything in to one OS, though, on the later 2004+ OS's this might not be possible;
there is a few cars in Aus running COS4, it's been out for a long time, but these were
only used by workshops for N2O cars only.

B3647 OL Commanded Fuel:
This was previously B3605 which was ECT vs MAP, and is now RPM vs MAP, so it requires editing.
In OL, any B3647 cells that are stoich will enable STFT trimming (semi-open loop mode),
this is undesirable during any BEN correction of VE or MAF.


===================================== TP VE Table (A0007) =================================

Nearly all 2003+ OS's use DBW, it is not known if TP VE table would work too well on them.

The VE lookup pid only works when using it in [g*K/kPa] (if you do it in % it may report 0).

Given plenty of logging the BEN will be so spot on you'd think the WBO2 was stuck!

The one problem this does have is in an A4 transitioning in and out of Drive,
there is no detected load change for the VE, however that also has an advantage
too because the fueling remains constant and the spark/IAC handles the RPM drop.

Use the TP VE only in the area's you cannot get consistent MAP values in,
and then switch back to the MAP VE once things become steady.

Make sure DFCO is disabled.

TP VE Table setup:
0. Set A0003 TP VE Enable to disable.
1. You need to have a good VE table with BENS very close to 1.00.
2. You need to change your preferences to view VE Table in [g*K/kPa] units.
3. Log a pid called GM.VETABLE_DMA to build your initial TP VE table.
4. Create a map that has the same dimensions as the TP VE Table and use the PID above as the Data.
5. Go for a drive and log, try and hit as many as possible points in the TP VE Table map.
6. Now copy/paste the values in the TP VE Table map to the TP VE Table in your COS5 tune file
(experiment with the copy and fill function and the the Linear Fill functions).

After the initial build of the TP VE Table, enable the TP VE Table,
change the TP VE Table map created above so that the data is now using the BEN factor;
this allows you to fine tune that table going forward.

Add a little to your TP table that you have plotted to ensure it isn't lean.

You will need to play with cut in/out points for switching between TP and MAP VE Tables;
for initial tuning, run exclusively off the TP VE Table, this allows you to apply BEN corrections;
to use the COS5 on an ongoing basis, apply BEN corrections both tables in their active regions.

If you get bucking at low RPM low throttle, check your base spark and high octane spark tables;
if they are way different where the bucking occurs try closing the gap between them.

================================================== ==============================================