72's and IM240

[Carl Ijames]

I needed bigger injectors last year for my 86 TType, and wanted to see how 
well I could make 72 lb/hr injectors work on the street even though they 
are much bigger than I really need.  They were pretty easy to get tuned up 
in 93 octane mode (cold start, WOT fueling, etc.) but the day I've been 
dreading came to pass last month  I received my Maryland emissions 
inspection notice.  MD uses the full IM240 dyno test, which I actually 
thought an advantage since the biggest problem with injectors this big is 
at idle where they just aren't happy at such short pulse widths.  You can 
usually smell a little gas in the exhaust, and the idle BLM's and IPW's 
hunt up and down even after everything is warmed up.  I had red-stripes 
before these, though, and these don't stink as bad at idle as those did 
.  On 93 octane my usual tune is 36 psi fuel pressure (vacuum hose off), 
and a maximum injector duty cycle in 3rd gear of about 70% (I said that 
they were big).  I lost about 1-1.5 mpg compared to the red stripes, but 
most of my driving is now short trips in town where the motor doesn't have 
time to warm up, so that may account for a lot of that.  On a recent road 
trip to NC (about 375 miles each way), burning gas purchased in southern VA 
to try to get away from the garbage (RFG) they pump here in MD, I averaged 
20 mpg which is about what I used to get with the red stripes.

Anyway, on to the testing and tuning.  I started by borrowing some time on 
a friend's exhaust gas analyzer to see just how bad it was.  With no cat, 
idling in park at 170-175 deg F and about 750 rpm, the CO was 0.67-0.72% 
and the HC was 550-620 ppm, with CO2 8.2%.  The CO2 is mostly to prove that 
there are no major exhaust leaks so I'm not going to list it anymore.  The 
HC readings especially were highly dependent on recent engine conditions, 
going well over 1000 ppm after idling for many minutes, and falling below 
300 ppm after holding a high idle for several seconds.  This machine also 
responded slowly, needing about a minute to get back to zero after removing 
the probe from the tailpipe.  Holding 2100-2200 rpms until the readings 
seemed stable, the CO remained 0.6% but the HC dropped to 120-150 ppm.  In 
both cases the injector pulse widths were 1.7-1.9 msec.  After this 
baseline I optimized the fuel pressure.  Higher pressures should give a 
better spray pattern but even shorted pulse widths and thus less 
reproducibility.  The optimum was at 33.5 psi (vacuum hose off), where the 
idle CO was 0.41-0.47% and HC 530-730 ppm and the 2200 rpm readings were 
0.56% CO and 130-135 ppm HC.  Even though the idle HC looked bad I think 
that this was mostly due to the slow response of the meter and the fact 
that the engine had idled for about an hour by this point.  After the 2200 
rpm test the idle HC was under 300 ppm but I didn't wait for it to 
stabilize.  I admit I was pretty pleased already since I think that this 
was almost good enough to pass the old tailpipe sniff test, and the cat was 
still in the trunk.

Next step was to make an emissions chip.  I didn't bother with raising the 
idle speed since this was going on the dyno.  I did raise the fan on 
temperature to 190 deg F and the fan off to 185 deg F, to let the motor 
warm up some even with the 160 thermostat.  I also raised all of the PE 
entry thresholds to their maximum values so that no matter how heavy-footed 
the test driver was the engine would stay in closed loop.  This included 
the PE MAF, LV8, LV8 hysteresis, TPS, and TPS vs. rpm thresholds.  For 
safety I set the wastegate duty cycle vs. rpm table to all zeros so the 
maximum boost would be about 12 psi.  At 33.5 psi fuel pressure and my 
usual injector constant ($1B) the idle BLM was 143, and I left this alone 
for now.  The big change I made was to make the engine run leaner than 
stoichiometric.  This is supposed to lower CO and HC at the possible 
expense of NOx, and is the whole rationale behind oxygenated fuel in the 
first place.  Unfortunately for the EPA, cars with O2 sensors don't really 
run leaner with oxygenated fuel, they just burn more of it (I call this a 
70's solution based on 60's technology, implemented in the 90's, and I'll 
get off my soapbox now ).  There are three sets of O2 thresholds in the 
chip:  closed loop upper and lower, rich-lean low and high, and rich-lean 
low and high slow trim.  The first set is used at location $3CEB in the ECM 
code to decide if the O2 sensor is warmed up yet.  The second set is used 
at $3D0A to see if a rich-lean transition has occurred, and if so to update 
the O2 cross-counts, the rich/lean flag, and the INT timer.  The final pair 
is used at $3E52 to actually compute the INT.  I have no idea why they used 
separate constants for the second and third pairs.  I have access to four 
GM chips for the TR's.  These are the chip that came in my early 86, the 
late 86/87 chip, the emissions recall chip, and a "hot GM chip" that I've 
been told was part of a series of Stage chips planned by GM.  Here's the 
important data from these chips (use a non-proportional font):

Addr   86     86/87  emissions  hot   Contents
     CMW0602 ACXA0942 BBKJ9464         Chip label sticker
      $018A    $0090    $2200   $0D42  Prom ID
3419  0.60     0.60     0.60    0.60  CL Upper Threshold
341A  0.30     0.30     0.30    0.30  CL Lower Threshold
341B  0.46     0.46     0.37    0.35  Rich-lean Low
341C  0.67     0.67     0.58    0.50  Rich-lean High
341D  0.50     0.50     0.41    0.30  Rich-lean Low Slow
341E  0.64     0.64     0.55    0.50  Rich-lean High Slow

These thresholds are in volts, and convert to raw chip values as N = 
E(volts) * 226.  For example, 0.60 volts * 226 = 135.6, which rounds to 
136, which is $88, the value found in all four GM chips at address 
$3419.  As you can see, GM leaned out the engine with the emissions chip, 
and further in their hot chip.  I don't have access to a wideband O2 sensor 
so I don't know if GM was starting at the stoichiometric air fuel ratio 
(about 14.7) and leaning the mixture from there, or had started with the 
original chip on the rich side, and was trying to get down to 
stoichiometic.  This may explain why some people say the emissions chip 
doesn't run as well as the original chip, since in general richer is 
smoother (until you go too far ).  I've been using the hot values in my 
street chips for the last few years, in the hope of slightly better gas 
mileage, but have no real evidence of improvement except that the idle is a 
little rougher.

For my emissions chip I set both Rich-lean thresholds to 0.15 and 0.5 
volts.  This lowered the idle BLM from 143 to 130, a change of 10% which in 
theory if the original BLM represented stoichiometric gives 143/130 * 14.7 
= 16.2 as the new air/fuel ratio.  I don't have an EGT gauge mounted, but 
the exhaust manifolds and downpipe did seem hotter than normal after a test 
drive.  I read somewhere that around 16 is the best air/fuel ratio for 
lowest emissions, so I was happy to see where this first guess landed and 
didn't try any other values.  The analyzer now said that the idle CO was 
down to 0.3% and HC 450-500 ppm, with engine temps between 185 and 188 deg 
F.  In drive, with the slight extra load of the torque converter, these 
dropped to CO 0.17-0.30% and HC 110-127 ppm.  In park at 2400 rpm they were 
0.37-0.4% CO and 70 ppm HC  yes, I was smiling .  The last step was to put 
the Terry Houston 3" cat on and go for a 1-2 mile test drive to warm it up 
and make sure that drivability was okay and that it didn't go into PE 
mode.  Everything was just fine.  Back on the analyzer, idling in park, the 
CO was 0.06% and the HC was 10 ppm, and idling in drive the CO was 0.03% 
and HC about 7 ppm.  Okay, I was grinning, not smiling, at this point 
.  Brake torquing up to 2000 rpm in drive gave the same numbers as idling 
in drive.  I left the test sheet in the car so I don't have the exact 
numbers, but I fast passed the IM240 test the next day, with readings for 
CO, HC, and Nox all between 5 and 10 percent of the limits.  In MD, if the 
readings are clean enough during the first 15-30 seconds of the test they 
stop there and call it a fast pass.

Anyway, that's the long winded version of how I fast passed the IM 240 test 
with 72 lb/hr injectors on a 3.8L V6.



Regards,
Carl Ijames     ijames at netaxs.com

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