Rodders tend to regard fuel injection as black magic. This article will try to demystify fuel injection and show how it is practical for street rod projects.

Installing FI on an engine not originally so equipped is actually not that difficult, though it is not trivial. But there are definitely advantages in overall drivability and in part throttle performance. All other things being equal, the maximum power from an engine equipped with FI will not be significantly different from a carbureted engine in good tune.

Performance potential of any engine is determined by its ability to pump air. Performance cannot be improved just by supplying more fuel; there will not be enough oxygen to burn the extra fuel. Traditional hot rod modifications (porting, bigger valves, turbocharging, etc.) are intended to increase the mass of air in the combustion chamber during the power stroke. Then we can increase the amount of fuel and get increased power.

For best performance under all conditions, we want to match the amount of fuel to the amount of air. The ideal ratio is about 14.7 to one on a mass basis; i.e. 14.7 pounds of air to one pound of fuel. For our purposes, somewhat richer than this is acceptable.

Carburetors do not accomplish this objective very well across the engine speed and load (throttle position) range. Electronic fuel injection, in contrast, can do so very well. In addition, FI can more easily manage the much richer mixture for starting and for warming up a cold engine.

The bad news is that FI is a lot more expensive than carburetion. But then you can't take it with you!

The job of the FI controller is to determine the amount of air flow, and then to pulse the injectors so that the amount of fuel injected provides the desired air/fuel ratio. There are two basic approaches to accomplishing this:

• mass air flow systems in which the air flow rate is measured directly. These systems have an air flow meter in the inlet air ducting.
• speed-density systems in which the amount air is calculated from measured engine speed, air temperature and known air flow characteristics of the engine.

Mass air flow systems are ultimately more accurate, since deterioration in engine condition is automatically accommodated. Speed density systems cannot adjust for loss of compression (i.e. leaks past rings or valves) or other departures from as-new. However, for street rod purposes, the speed density systems are at least as satisfactory in practice and have a couple of advantages. The flow meter in a mass air flow system imposes a pressure drop which degrades performance somewhat. Also the air inlet ducting must conform to certain requirements for accurate metering. Speed-density systems have neither of these limitations.

One common misconception about FI should be cleared up before we look at an example. It is commonly assumed that we want to inject fuel directly into the combustion chamber through an open intake valve. This is not the case. On the contrary, fuel is injected with timing so that it the fuel droplets hit the back side of the closed intake valve. Because the valve is hot, the fuel will be vaporized and will mix with the air before entering the combustion chamber.

Another basic issue: recent factory fuel injection systems are called sequential, because each injector is pulsed individually at the optimum time. Many earlier systems from the 1980s and early 1990s fired the injectors in groups - so called batch fired. This is a compromise, since some injectors will be timed optimally and some will not. However, the batch fire systems have proven very satisfactory and have the advantage of not requiring a system of triggering based on crankshaft position sensing. We can simply use a pulse from the ignition system such as a tachometer output pulse from an ignition system such as an MSD.

The project described here is a speed density system with batch firing.
 

The Project

Back in the late 1970s I installed a Slant Six in a 1941 Dodge sedan. (I don't remember why, it seemed like a good idea at the time.) For a number of years it sat almost finished while I was distracted by other pursuits. A couple of years ago when I decided to "get a life", after considering repowering the car, I finally decided to leave the mighty six in place and have some fun with it. I discussed the idea of adapting fuel injection with an engineer who was involved in FI system design with Bendix. He convinced me to give it a try.

Today you can simply go to after-market suppliers, pay them a lot of money, and they will supply everything you need. (Though if your engine is not popular type like the small block Chevrolet, you still have some machine work to do to install the injectors.) But I thought it would be more interesting and also less expensive to try a junk yard approach.

Some research turned up the 1986-88 Fiero GT 2.8l V-6 as a suitable donor for controller, sensors and throttle body. (Yes, you still need a throttle body, but it's just an air valve.) It is a speed density system, batch fired and, very importantly, has the right number of cylinders.

Sensors required are shown in Table 1. The factory system has other sensors and control functions other than injector control, but we can omit these for a street rod.

Selecting and mounting injectors

Of course, we need a set of injectors. The Fiero V-6 is 2.8 liters in displacement and its injectors are rated at 15#/hr. The slant six is 3.8 liters; we need injectors that flow about 22 #/hr. Injectors from a batch-fired 3.8L Buick are rated at about this throughput, and are electrically compatible with those from the Fiero. This means that the Fiero controller should be able to drive them. I just bought a junk yard set, about $100 for all of them versus about the same amount each for new ones. I do recommend installing new o-rings on those junkyard injectors.
 

Mounting the injectors is perhaps the most difficult aspect of the project. The injectors should be installed so that the spray pattern points at the back side of the intake valves. We have to get into some machine shop work. Simply using a stock intake manifold would work. However, I had come into possession of a dual carb Offenhauser manifold for the Slant Six. (Imagine the impressive results from that modification!) This is a lot sexier than stock, because we can now run dual throttle bodies, but will probably not result in a noticeable performance increase over one, since the bore on a Fiero throttle body is quite large.

To mount the injectors you could have the machine shop make some mounting bosses, but I bought some after market components that were cheaper than the machine shop route. Then my friend Lorne Hanson machined the manifold to accept the bosses as shown in the figure above. We installed the bosses with epoxy rather than welding, because we were concerned that welding might distort the bosses, resulting in fuel leaks.
 

But before installing the injectors, we need a fuel rail. The inlet side of the injectors fits into the fuel rail with o-rings. Again I bought fuel rail material which we then machined to our requirements, determined from a junkyard fuel rail. We then assembled the injectors into the fuel rail and into the bosses and then carefully placed this assembly into the openings machined into the intake manifold. Now it's time for the epoxy.

After the epoxy set, we deburred and polished the inside of the intake manifold so as not to impede the air flow. The intake side of this setup, up to the ports, will really flow a lot of air with minimal pressure drop. So with porting, polishing and bigger valves (coming later), the air flow potential (hence power) of this engine will be greatly improved. The final part of the intake assembly was making a pair of adapters to mate the Fiero throttle bodies to the Offenhauser manifold. As part of this we opened up the manifold openings to match the 2 1/16 in. diameter of the throttle bodies.

One important lesson here: I assumed that I would need to utilize the idle air stepper motors in the throttle bodies so we made provision for idle air in these adapters. In the end, I am not using the stepper motors at all, so this job could have been a lot easier.

Fuel supply
 

A high pressure fuel pump is essential; pumps of this type can develop about 90 psi supply pressure. Again, you can buy after-market units for several hundred dollars, but I bought a junk yard pump from a Grand Am. The stock GM pumps are in-tank units, so I had the tank modified to mount the GM unit, as shown in the picture. It is recommended that you place a baffle in the tank to reduce the risk of starving the pump when the tank level is low.

The pressure regulator maintains a constant differential pressure between the fuel supply and the manifold pressure. Whatever the manifold pressure (i.e. vacuum), the fuel supply pressure is a set amount higher. Factory GM units maintain about 40 - 45 psi differential. These regulators are preset and not adjustable. But because the Slant Six and the donor engines have different air flow characteristics, we will need to be able to adjust the differential. In my case, I have found that a differential of 50-55 psi is about right. So I had to purchase an adjustable after-market pressure regulator. Note that these regulators must be referenced to manifold pressure. This is the 1/8" hose shown in the photo.

To finish the fuel supply side, we need a fuel line from the tank to the pressure regulator and a return line from the regulator back to the tank. I used 5/16 aluminum line and some industrial fittings for this. And one other thing: a fuel filter suitable for the high pressure in the supply line. You can also see the braided line that supplies fuel at regulated pressure to the fuel rail.

After running a 12 volt supply to the fuel pump, and hooking the pressure regulator to the fuel rail, you are ready to pressure the system.

Since I was concerned that I might have some leaks, I did an air pressure test first. Now you would not normally want to pressure test with air, as its dangerous. But the pressure here is not that high by industrial standards, so its ok, and I found and fixed several leaks with a soap test. This is better than having gasoline leaking all over your shop. Make sure you don't include the fuel tank in your pressure test!

Then I ran the pump to pressure the system, found and fixed one more leak and adjusted the pressure to 40 psi as a starting point.

The electrical side

A major part of the project is understanding what all those wires coming out of the controller are meant to do, which ones we need to use and how we hook them up. I bought a manual with a wiring diagram for the Fiero, and used this as a guide to do a test on the work bench. It takes time and effort to temporarily connect the sensors, 12 volt power supplies and injectors. You end up with a mess of wiring that you will throw away in the end. To simulate firing pulses I just borrowed a square wave function generator.

Be sure to make a diagram and check your hookup carefully. It would be easy to blow the controller if we do something wrong here. Eventually I was able to fire the injectors. At this point you don't know whether the performance will be acceptable, but at least you know you can pulse the injectors. The only choice now is to hang all this stuff on the engine.

I simply did a temporary hookup with the controller sitting on the fender and wiring running all over the place like a mess of spaghetti. At this point I was trying to use a primary pulse from a conventional ignition coil as a reference for the controller. When we tried to fire the engine, we quickly found this was unsatisfactory, even after we tried to shape the pulse electronically. We could not get rid of false, multiple firings due to ringing of the primary circuit.

A Chrysler junkyard electronic distributor and a Holley Annihilator box solved this problem, since the tachometer output from the Holley is just what the controller needs. After fixing a couple of vacuum leaks, the engine fired and ran quite well at idle in the garage.

Keep checking for fuel leaks. I found one minor one.
 

Now that it runs-- Factory FI systems use a stepper motor to allow a controlled amount of air to bypass the throttle body butterfly. This then influences idle speed and the controller signals the stepper until the desired idle speed is achieved. It seems that for street rod purposes we don't need this complication. The Fiero throttle bodies have a manual stop adjustment. I just used this to set the idle speed. If you wanted to drive your rod in our cold weather, this would probably not be satisfactory; you have no fast idle. But it seems just fine for my purposes.

I put Lorne Hansen to work fabricating some throttle linkage which actuates both butterflies and also has a lever to actuate a throttle pressure cable from Lokar. Another Lokar cable connects to the throttle pedal. At this point I played with fuel pressure and found that the engine seemed to like 50-55 psi (this just by ear). Before we could venture out on the road at all, I had to do something about that mess of wires and the FI and ignition controllers sitting unteathered on the fenders. I mounted the Holley box permanently. But still not convinced that the Fiero controller would be satisfactory, I moved it into the passenger compartment (held down with duct tape!) and ran the sensor and actuator wiring semi-permanently.

Now we are ready to idle this thing around the block. Ted Allan had previously fabricated a header setup for me, using large diameter tubing throughout. The mighty six sounds great and is running nicely. It wants to stumble a bit coming off idle and actually stalled a couple of times. Increasing idle speed basically stopped this.

Road Test

There is no vacuum advance hooked up - mechanical only. I set total advance to about 35 degrees at idle, about 50 at full advance. After a few more checks and fixes, its time for the real test. I drove out to Lorne's place - seemed fine on that short run. We hit a paved side road and went through a normal range of speeds, acceleration, deceleration etc.

We were both very pleased. The engine is extremely responsive, runs smoothly at all the conditions we checked. (No real high speed stuff, as if that would be possible anyway!) I think we have a runner here.

Remaining

There is still a slight tendency to stumble when just coming off idle. I believe this is because of the dual throttle bodies. The controller is being deceived, since it is programmed to react to the air flow increase from only one throttle body in proportion to what it sees from the throttle position sensor. Balancing or intentionally unbalancing the throttle bodies at idle may cure this.