My opinions from a very comprehensive search of Australian SAE automotive papers and as many new car tech articles as I can digest. Well, there were new when I got them as a kid from Wheels, Street Machine, Motor, Autocar, Fast Lane, Car, Road and Track, Motor Trend, CarCraft, Hot-Rod, Popular Mechanics. Some of these date back to 1979, others are recent.
EFI will handle any cam. There are varoius types. Technically, asside form Varicam BMW's, there is no kind of gasoline injection system bar the throttle body type. The throttle body must exist somewhere to allow air control. It doesn't mater if its three feet before the intake, or 3" before the intake, there must be a metering device to make it work. The placing of the injector is important. Indirect Port injection isn't the bees knees for power. Some cars with injectors 6 inches away from the intake valves, piddling on to the back of the thottle blades, can rout a good port injection system at various points in the rev range. Direct injection systems require very specific design, and is not open to any changes bar specific machined alterations. It's essentially you-toucha-my-car-I braka-your -face-deal
Throttle Body EFI. Called, in the trade, pi$$'n'dribble. Single, dual or (almost) port injected. The basic type is the CFI units found in 1981 till 1985 302's. The TBI systems in Chevs, singles like in the 2.5 Iron Dukes and dual TBI, such as the 82/84 Cross-Fire 'vettes. The final incarnation is the indepedant runner systems like TWM's Weber DCOE version, which is used on SB and BB V8's with independant runners to the ports. These systems are TBI, but work with very short injector to valve distances. They are like the 427 and 289 GT40 Weber carb set-ups.
If the right type of multiple Throttle body injector, such as three from a CFI Taurus, is used, it can do a brilliant job. The Megasquirt guys have started doing a few on these on there site using Aussie Falcon EA 3.9 CFI injectors. One 460 runs just two on a Weber DCOE intake. As soon as you move away from an indepandant runner system, things get worse and worse. Last on the list is a single TBI unit. It has rotten fuel atomisation characterisitics, and works well only because an EFI system is easier to certify than a feedback carb because it can compensate for air temerature and fuel air ratio trim much better.
The method of metering the fuel is genrally by look-up table from a defined list of Manifold Absolute Vaccum values. The engine is dyno tuned or run on the road with a data logger or, on more basic systems, just with a rich or lean lambda display. You tune it at a given temperature, and then define the envelope you operate to. It may be a fuel only sysyem, or a fuel and ignition system. There is no manifold air flow meter or hot film potentiometer, or Karmen Vortex sensor to make things akward.
The port injection is a much better option for most of us, as you can get.
Here, There are many metering devices critical to emissions . Devices such as the
Manifold Absolute Pressure or
Manifold Air Flow meter or
Hot Film Potentiometer or
Karmen Vortex Sensor or
The Bosch L, LE, or LEII Air Vane Meter
These five systems are what you find on any Bosch K, J, L, D or EEC or Calpak/ Memcal or OBDI or OBDII system.
They are wounderfull set-ups, which have been engineered to a high degree of aptitdue by world class experts. Anything you can do your self will add kaos to order, unless you get a special assistance. Some kits, like the GM Calpak/ Memcal , are so friendly, you can just about do anything. Others, like the Ford EEC system, are very easy to damage unless you have a 25 year IBM veteran to help you out. Things are much easier with Intercetor or Extreme Edge interface kits, but the EEC's are much more nasty because when eggs like Chuck Missler and co designed the electronics in the late 70's, they added a raft of computer structures to create sequential injection options. Very complicated binary codes, not as simple as the GM stuff.
The status of bankfire or sequential firing is irrelevant, except in relation to engine start time. Sequential injection systems often start quicker, but the sequential systems are harder to calibrate than the bank fire set-ups. Both operates by pulse length, but the sequential has a timed event which varies a great deal. The GM Bankfire set-ups are really nice to tune, and they are not as fussy over cam timing. On a V8, one bank has one pulse length, like a big 4-choke carb on one side, then it fires the other bank with a similar pulse even timed at a different interval.
Some sequential systesm, like the pre OHC X-flow 4.1, have nasty vane meters and are not easy to work on unless you ditch the air vane and go for a more modern MAP or Hot Film Potentiometer.
My pick? The MAP set-up is the ideal one to have, as it has no close restrictions upstream. It can be a beggar to calibrate, as manifold temperature and flow can mean there is no limp-home if it fails. There may be other issues.
The thing I like is that you can do away with any downstream throttle blades on a six cylinder in-line engine, and use the 5.0 V8 throttle body to meter flow. This allows total freedom in intake design. In theory, a Hillborne injection system could be gutted of its throttle slides, and a big perspex or glass cover placed the intake trumbets. These could then be linked to one or two big throttle bodies, and the MAP sensor linked up to the manifold. You then have freedom to place the injectors in the right place, it'll look great, and the system can then be calibrated to bank fire or sequential.
Hope this helps. No opplogies for the loose terminology. Better scholars are around to sort out the tree-branch diagram or EFI systems. Those that are able to explain it better are not interested in the big picture. Those that are into the big picture, are not experts...