[Diy_efi] Update on EFI without O2

[Marcello A. Belloli]

Hello Everyone,
     I finally got a chance to take my friends car to the track last
weekend to try and dial in the efi system.  By the time I'd arrived
back in the USA they had already tried to get drive it.  First think
I was told was that there was a very audible knock on acceleration,
even without a load.  10 degrees of timing was taken out to make it
go away.  They ( not me ) disconnected the coolant sensor in an
attempt to force the computer to provide more fuel.  I really don't
know what they were thinking with that idea.  The next thing they did
was install an innovate wideband O2 to try and get an idea of the the
a/f ratio.  They were never running lower than 14.9 to 1.  Once they
saw that they changed the fuel injectors from 240 lb. to 300 lb.  And
they were still lean.  At this point they gave up and finally decided
to wait for my return before going any further.
     Well, the first think we decided was with the 13.6 to 1 compression
ratio we needed to get fuel to go with it.  The fuel has to match the
compression if we want to be running any advance, and have any chance
of making power.  We tried 100 octane fuel.  On the dyno we still had
a knock, and now I was actively trying to remove the knock by
rewriting the spark table for the computer.  I ended up taking around
20 degrees of timing out around 5000 rpm.  Well, I didn't think it
was worth going any further unless we could bring back our timing. 
We went to 114 octane fuel, and put the timing back to stock.  We no
longer had a knock.
     I also had reconnected the Coolant Temperature Sensor before
contining onward.  The hard starting I had been told about while I
was away disappeared.  And the check engine lamp was now out.
     Now we could attempt to dial in the fuel table.  We had a timing
curve that made sense.  I figured we would try and get the fuel into
the ballpark before we did anymore playing with timing.  The engine
we had was a Honda/Acura with Vtec.  The way the computer is set up,
there are two distinctly different fuel, and spark table for when
vtec is engaged, and when it is not.  Vtec is basically honda/acuras
version of variable valve timing.  I found that I had a lot more
control over the ECU than I had thought I'd be able to have.  With
the help of Crome, and www.pgmfi.org I was able to raise the rpm for
vtec engagement, and lower the rev limiter so that vtec would not be
engage for our first runs.  I was also able with basically a flip of
a switch turn off the O2 sensor.  I was able to disable the O2 sensor
without removing it.  It was one of many options that I could change
using Crome.  We set the Rev limit to 5500 to keep the driver from
getting carried away.  After a couple mis-attempts at burning a new
chip we finally got going.  We set up the Innovate Wideband, and
simply turned on the record function for the entire time the car was
on the track.  I went with the driver for the first run, and quickly
realized I did not want to get back in the car.  I'm sticking to
programing for now.  I hadn't been in a car with sticky tires for a
long time.  We made no changes beside to the fuel table for our first
run.  We were simply trying to gauge where we were.  We quickly found
that we were never seeing an A/F ratio much below 14.5:1.
     I decided to pull out another project and use it for this testing.  I
had made a chip switcher eariler, that allowed the switching between
four different chips on the fly.  And here was another trick from the
Crome dept.  I was able to remove the checksum routine from all my
chips, and I never saw a check engine light during all our testing. 
The chips would easily change between each other during the testing. 
We were able to change between chips at any given RPM without
failure.
     We set it up with the non-vtec fuel map as the only thing to change
for these first tests.  We made 4 chips (one chip with four images on
it) each having a 5% increase from the previous in fuel.  We quickly
found that 20% more fuel put us in range.  We were in the 12-13% AFR
range.  So we stopped with the non-vtec map, and turned vtec back on,
raising the rev limiter to 7500 rpm, and dropping the vtec engagement
point to 5300 rpm.  We did the same thing with the vtec on, and came
away with 15% being what seemed the sweet spot.  Now all the time we
were doing across the board changes to the fuel tables.  We didn't
try and zero in on any one rpm.  We just wanted to get in the
ballpark, and make sure we weren't too lean anywhere.  We figured a
little rich was better than too lean.  Once we had a new chip with
20% increase in non-vtec fuel, and 15% in vtec fuel it became
apparent that we had added to much fuel to the non-vtec map.  We
dropped the non-vtec to 15%.  We now had an across the board 15%
increase in fuel.
     We ran out of daylight, and track time.  But we now have a drivable
car.  Very drivable.  Exhuast temps around 900F.  One cool thing
about disabling the NBO2 is that you can simply take it out, and put
the Wideband in its place for testing purposes.  We've even thought
of leaving it out for good.  The last engine that was in this car was
using oil, and went through O2 sensors often.  Our AFR is around
13.2:1.  That seemed to be where things worked best for us.  I must
say I am very glad we kept the 7500 rev limit on the chips towards
the end of the day.  This motor was still in the break in period, and
the driver was getting carried away.  I know that if we had bumped it
up to 9000rpm.  He wouldn have been riding the 9000rpm limit.  It was
running too good.  He had been saying earlier in the day, it doesn't
look like we are using any fuel.  Towards the end of the day he
changed his story, saying you can watch the fuel gauge go down as you
are driving.  The Rev limiter was keeping his foot out of it.  He
complained of not having as much power as he expected.  But it was
running very stable now.  We can go looking for power later.  We are
going to try dropping the octane down a bit now, and see what
happens.  With as much more fuel as we've added maybe we'll be lucky
and some of our knock will have been removed.
     Next we are going back to the dyno and checking all the cylinders.  I
feel kind of stupid for not having already done this.  We are going
to put Widebands on each of the four cylinders and see how evenly our
fuel is dispersed.  Make sure one cylinder isn't overly lean in
relation to the others before we go any further.  Maybe get a power
run, and see how much HP this thing is putting out.
     I think removing the O2 made tuning much easier.  If we had been
trying to get to 14.7:1 it could have helped.  I wish every computer
had the ability to have the O2 turned off.  And it was so much easier
than trying to come up with something to trick the computer into
thinking it has a perfect mixture, and make no adjustment to fuel
trim.

Later,

Marcello