Port fuel injection was available on the 350 Olds motor used in some Cadillacs.
1975 ½ to 1979 Eldorados, and California emissions equipped 1980
Sevilles and Eldorados. Unfortunately, there is no stock big block FI
intake. You might be able to make one fit with
adapters to compensate for the added deck height.
It had a port injected manifold. The system used a manifold that was
basically the same as the carbureted version. The manifold was a very low
profile, dual plane type; sort of like a Weiland X-cellerator, but much
flatter. Looks like it should flow pretty good for
a mild street motor. The major differences being the injectors mounted
in the ends of
the ports and a hole in one of the ports at the front of the manifold for
an air temperature sensor. The throttle body looks like
somebody cut the top off a two-barrel carb.
It has no heat riser and
it looks to be a special casting to accomodate the fuel injectors.
It has 8 fuel injectors, 2 are mounted at
the end of each runner, each injector feeds an individual intake port.
There's a copper-looking fuel rail that feeds all the injectors. The
injectors look pretty standard issue, nothing special. I understand
that there were some aluminum EFI manifolds produced.
It has EGR. The valve is mounted in the same place as any of the
Olds engines. Four bolts hold down the throttle body. The diameter of
the barrels on the
throttle body are 2.25 inch, the same size as the secondaries on a mid
70s vintage Q-jet. The mount on the manifold for
the throttle body is about 4 inches square, and there's two round holes
in the middle that match the barrels. The hole that the throttle body
mounts over is bigger than most 2bbls and smaller than a 4bbl. The ports
that mate to the head are about 1.25 inch by 1.9 inch. The throttle body was
manufactured by Bendix, as was the fast idle valve. The TPI sensor is
Bosch and has a 5-pin connector (actually 5 spades, not pins).
The electronically controlled, 2-phase Bendix, fuel injection system
provided 180bhp @ 4400rpm,
and 275lb-ft @ 2000rpm, slightly more than the carbureted version, mainly
due to cooler manifolding. It was reported that this engine was able to
meet emissions standards without a catalytic converter, but was so equipped
nonetheless!
The computer was an analog unit that was located in the center of the
dashboard. It was a regular Olds 350, with a different PFI intake and distributor.
They were very often converted to carbs because the system did not run well at times,
and was also known for catching on fire, due to some
questionable rubber hose used for the pressurized fuel line. For anyone with one of
these cars, convert to today's FI hose!
The ECU's for these port injection systems didn't place any
programming in removable EPROMs. For instance, Cadillac used a
different ECU for each body style to help tailor throttle response
characteristics.
Fuel delivery started with two fuel pumps; one in the tank; one further up
the line. From there, the fuel was pumped through the injector banks and
to the fuel pressure regulator. The pressure regulator kept the pressure
at 39 psi. It was attached to manifold vacuum, so, apparently, that had
something to do with regulating pressure. From the pressure regulator the
fuel ran back to the tank. The fuel injectors were solenoid operated
pintle valves. The ECU signals them to open for precisely the right
amount of time to meter the fuel needed by the engine. Air delivery was
provided by a throttle body and a fast idle valve.
The system was most likely bank fired (4 at a time). There was no crank
sensor, so it couldn't be sequential. There was a sensor in the distributor
housing that indicated 1 pulse per 1 rev of the distributor. It was a closed
loop system; about as good as an analog FI system gets. The system is covered
in several books that relate to automotive electronics/history.
The system had pretty much all the standard sensors, except for a
mass-airflow sensor or oxygen sensor. It had a manifold absolute pressure
(MAP) sensor, so mabye that was used in place of a mass-airflow sensor.
This along with the bad temp. sensor location may be why the system had a
reputation for running lousy. The system also used a throttle position
switch, and air and coolant sensors. The air and coolant temperature
sensors are identical and completely interchangeable. There was an engine
speed sensor in the distributor. This sensor consists of a plastic housing
containing two reed switches and a rotor with two magnets. The ECU uses
the opening and closing of the switches to synchronize the fuel delivery
and to monitor engine rpm. The distributor was unique to the EFI, and a
regular distributor won't work with the system. The ECU is serviceable
only by a Cadillac dealer according to the manual.
he throttle body is not just an adaptor for the aircleaner, it uses
throttle blades to regulate air flow into the engine just like a carb
would. It also houses the fast idle valve and a Bosch throttle positon
sensor. It also has a buncha vacuum connections for the usual
accesories; vacuum advance, power brakes, PCV, etc. It takes a standard
4-bbl air cleaner but it has a spacer so that the air cleaner will clear
the HEI and the fuel rail.
One of the biggest design flaws was mounting the intake air temp sensor
in the manifold where it was affected by heat from the EGR valve and cooling
system - not good. These cars didn't tend to run that good.
If reusing the original 70's ECM, you would need the PFI dist - the
distributor is unique to this analog ECM. It's been verified from several
people that the output from the 7-wire module is the same as the module
from a TPI distributor, so a GM TPI ECU could be used. Fortunately, the
ECUs didn't control spark,
so if the 350 was rebuilt with a higher compression ratio, it would
be a relatively simple task to recurve the mechanical advance and install
an adjustable vacuum advance unit.
One would probably be able to use any GM V-8 (firing order 18436572)
TPI (not SFI) ECM (to start with, then roll your own, or use some other
aftermarket Chevy TPI setup. To my knowledge, the TPI (85-89 at least)
system just does 1 bank on one spark pulse, and the other bank on the next.
The beauty in the TPI ECM - since it is NOT sequential (injectors
fire one at a time in relation to each cylinder), there doesn't need to
be any device to determine TDC of cylinder 1.
Also, it's been said that sequential will only really be a noticeable
advantage over bank at low speeds. Much above 3000 rpm the difference is
negotiable.
Now for some speculation on performance potential. The ECU apparently
cannot be serviced and is pre-set for the (smogger) engine setup, a 180 net
HP 350. I think it might be possible to "dope" the inputs to the ECU so
that it alters the mixture accordingly. For example, playing with the
temperature input reading so that the car runs richer. The only problem is
the ECU probably has preset ceilings for how much fuel can be injected, so
the engine would likely lean out at higher rpm.
Another idea I have, which
I think is more viable, is to use some sort of adjustable fuel pressure
regulator to increase pressure so that (hopefully) more fuel will be pushed
through the injectors across the rpm range, and you could "jet" the engine
by adjusting the fuel pressure. This would be limited of course by how much
pressure the lines and injectors could withstand.
Another problem with the
system is the fact that it used a two-barrel manifold. A throttle body will
flow more than a carb of the same size but I have my doubts whether the
throttle/manifold setup could flow much more than 500cfm. It may be
possible to mount injectors and sensors in a high-performance manifold and
use a Cadillac ECU, but I doubt it's worth the time and expense.
If you are looking for one of the above cars with this setup, they usually
had an "Electonic Fuel Injection" badge on the fenders and grille.
Ask around at some repair shops.
[ Thanks to Kevin Wong, Greg Pruett for this information ]