Re: VE stability

[Scot Sealander <Sealand@clarityconnect.com>]

"John S. Gwynne" wrote:
 
> I've been collecting tuning data through the ALDL now for 3 1/2
> months and correcting my VE table with BLM and O2 data ('730 5L sbc,
> sc, cam, headers, axxc bin).

You have the basis to do an even better job of collecting data.  Use the
extra RAM space you added to collect data as it runs, using much finer 
resolution than the BLMs... Try forcing the BLM to 128, and then
recording the
min/max integrator for the operating range you are trying to fix.
I found that this method works great!  Got the VE table much closer, and
it
took less time.  I am using a 16 x 16 table of operating conditions as a 
base.  (MAP and RPM)


> I looks to me as if something is varying with operating conditions to
> prevent me from narrowing this last 4% of error (it's not that much,
> but it has my curiosity). The two conditions I see this best are 1-2 hours of
> around-town driving versus 4 hour of highway mixed with 1-2 hours
> of around-town. It looks to me like something in the system is changing
> characteristics after a couple of hours of drive time.
> 
> Has anyone else seen this?

Sure.  ;-)


> A few of my thought of what may be going on.... appreciate comment and
> suggestions....

Ok, give it a try.

 
> -- baro measurement screwed up because of sc
> I'm collecting data on this now. The condition does seem more
> pronounce at partial throttle conditions.

Definitely shut off the "engine running" baro update.  This will get
messed up with the SC.


> -- EGR
> EGR partial pressure correction needs updated for exhaust system

Probably, but there is more to the EGR thing.  Do you consider the added
temp that the EGR adds to intake?  I worked on getting non-EGR VE right
BEFORE tackling the VE with EGR.


> -- inlet air temperature reading inaccuracy
> MAT sensor may be influenced by block/manifold temp

Absolutely.  Get the IAT out of the intake manifold. In my experience,
it
does not work well there.  I put the air temp sensor after the
intercooler, and
then add an airflow adjusted amount of temp to try and get an accurate
cylinder 
air temp.  This then references a table that uses a theoretical drop off
in 
air density for temp. It looks real good now, but it has all been done
with no 
EGR.  I am building a new model for cyl air temp during EGR operation.

Also, idle can heat soak the engine.  You find that the BLMs change as
the 
engine idles long term.  I added an "idle temp adder", that slowly ramps
up as 
it idles. This seems to fix the BLM creep at idle.

I also added much needed resolution to the VE table.  I think some of
the
VE variation problems at low MAP (idle to off-idle) was due to the large
VE 
changes in that area of the VE map.


> -- engine/cylinder conditions
> carbon in the cylinders changing the burn after highway driving

Possible, but in my opinion not likely....


> -- sensor temperature sensitivity
> MAP sensor may be somewhat temperature sensitive (1-2%)

I think Motorola calls out a range, got a data sheet?


> -- limitation of the tuning procedure
> maybe, but it does look like it in the data.

A possibility.  Did you shut off cannister purge?  This will drive the
BLM lower when it first turns on.  If you forget it is on, it will
show up as a hole that grows into a peak if you try and fill it.
Ask me why I know..... ;-)


> -- O2 sensor
> temperature/accuracy sensitivity of the O2 sensor.

Could be, but GM tries to take some of this into account in the code.

One thing you have to remember, are you considering ALL the variables
that
change as the engine runs?  I was thinking of adding something for the
expansion that happens at the throttle plate at low throttle angles.
But not too much error there....  The airflow based temp adder may
consider
this already, albeit indirectly.

Scot Sealander

----------------------------------------------------------------------------
To unsubscribe from gmecm, send "unsubscribe gmecm" (without the quotes)
in the body of a message (not the subject) to majordomo@lists.diy-efi.org