ECM Tolerances? EEPROM PCM in LT1

[Kurek, Larry]

Guys:

I have a 1994 LT1 Camaro, that has a bunch of the basic bolt on
modifications. However, I have always felt the car was a bit weak for
what I had, but I couldn't figure out what was up. I ran a couple of
Diacom runs and have a knock retard problem (10 degrees of retard) at
pretty much any RPM over 4000 to about 5500 or so. The effect of this
retard was obvious on the dyno runs I made....the power/torque curves
went up smoothly until about 4000 rpm, where they got very choppy, and
the rate of increase fell off dramatically. So, I thought I had a
problem with low octane fuel. Since I was running 93 anyhow, I popped a
can of 104 boost in, to no avail. So, I felt I would have to just live
with it.

Well, I recently was able to get a hold of another PCM from a friends
car (Dave), so I decided to do a bit of experimenting. First, I swapped
knock modules between the computers. While I still had knock, it was
shorter in duration and only peaked around 8 degrees or so. So, I
figured I'd try the whole PCM next. I swapped it in, and, until it
learned my car, I got some retard, but now usually around 6 degrees,
with an 8 here and there. Well, once the computer learned my car ( a
couple of miles of driving), I subsequently ran my FIRST Diacom run from
a stop through third gear (@ WOT), with NO knock at all. I was
shocked...and pretty happy (and still need to talk to Dave about this).
In normal driving since then, I still get a few spikes every once in a
while, but they are two-three frames tops, and peak at about 6 degrees.
The car feels MUCH stronger.

Since the car is identical, uses all of the same sensors, and has the
same operating characteristics...what in the heck is going on? Well, I
thought I would do some analysis of all of my Diacom runs. After
spending a few hours last night with a database, and Excel, I have some
interesting observations...and need some further opinions.

First, all analysis was done as follows: I screened all of the Diacom
data to only provide data points that were made with throttle positions
of 90% or greater...so that I was in WOT mode. I then did an average of
each run for the range of RPM. So, what I ended up with was a data
table, by rpm, with all of the Diacom values averaged at that speed. I
then graphed them all (each measurement seperately, but all runs on the
same graph). 

Results:

First, both the left and right O2 readings were remarkably different
between the two ECM's (I'll refer to them as the LK...original one, and
DW...Dave's). The LK had both O2's in the 920-940mv range.
Interestingly, the curves started out around 900 at  lower RPM's, but
curved up as RPM climbed. The DW curves were in the 880-900mv range and
were flat from low to high rpm.

Second, with the SAME engine and associated hardware, the airflow (MAF)
rates were slightly higher with the DW than the LK. 245 gps peak with
the LK and about 260 with the DW. There were some unusual peaks with the
DW that were not in the LK...peaks in the 310-315 GPS range.

Injector pulse widths were on average lower with the DW than the
LK...confirmed by the lower O2 readings. However, the curves followed
the same paths from low to high RPM...which doesn't explain the rise in
the O2 curves of the LK.

Finally, the battery voltage measure in the DW was consistently .2 - .3
volts lower than the LK. This surprised me, since I think the low
battery compensation in the PCM actually raises the pulse width for
lower voltages. Granted, the actual measurements were 13.1 in the DW vs
13.4 in the LK, so I don't know if this is a factor or not. I suspect
they are both in the normal operating window of the PCM, so no
adjustment was needed.

So...what is going on here? Since the car was the same, and the ONLY
thing that changed was the PCM, I assume something is up here. What?
Both the LK and DW PCM's are 94/95 with the current GM programming. I
can see driven devices (injectors, advance, etc) being different between
computers if they have different programs, but why are the sensors
behaving differently, as far as the PCM is concerned? They obviously
aren't, since they are the same sensors. My first guess at this was
maybe the production tolerances in the PCM, especially the internal
voltage regulator. Since the PCM provides 5v and 12v regulated voltages
to sensors (and a 450mv reference bias to the O2's), maybe that could be
it. However, the TPS voltage at WOT between the LK and the DW PCM's was
identical...4.9 volts. Maybe there is a circuit in the PCM that measures
voltage (must be, since it determined +bat), and there is a bias in it?
Obviously it is reading my battery voltage as lower with the DW than
LK...but would this carry over to all of the sensor readings as well?
Maybe it is interpreting a "real" 13.5 volts at the battery low, so it
is interpreting a real 5.2 volts as 5 volts reference? I'm lost...

Any ideas? Need more info?

Thanks!

Larry Kurek