Fuel injector operation

[curious125]

How many different methods are there of controlling the fuel injectors?

I am tinkering with building a fuel injection system for a small 2 cyl. engine. I was thinking of using pulse width modulation (pwm) to control the injectors.
Is the the norm, or am I way off in right field here?

 

[GregLocock]

so far as I know pwm is the usual method. I suppose you could use voltage instead, and a fixed pulse width.

Cheers

Greg Locock

 

[curious125]

I thought that pwm would make the most sense. But would there only be 1 pwm channel for all injectors, or would there be a pwm channel for each injector.

Seems to me that both could be true, as the difference between throttle body vs. multiport, or as neon vs. viper (low end vs. performance cars)

 

[GregLocock]

Either works, even with multiple injectors. Individual injector pulses (sequential injection) for each injector don't give you as much of a performance benefit as you might think (in fact throttle body injection is fine up to about maximum torque, on our engine).



Cheers

Greg Locock

 

[curious125]

I have seen on some other postings alot of info. about ECU memory maps. Are all fuel injection systems operating on a memory map system, or are there dynamic systems as well?

 

[GregLocock]

All modern ones use maps (or tables). By a dynamic system I think you may be talking about a feedback controller, as used on a Ktronic (is that what it was called, mechanical fuel injection anyway) or a diesel engine.


There are many feedback loops in an EEC, except when it is running open loop, when it essentially picks up the table values and uses them

Cheers

Greg Locock

 

[optimum]

A very very very long time ago in High School I sat down and designed a very crude system for a single cylinder B&S. I figured I could use the flywheel to trigger a fuel pulse and measure rpm. With such a simple engine fuel control is not that critical so it was more of a rpm based fuel map. I think I linked a potentiometer(SP) to the throttle with a simple and /or gate circuit. Example of 3600RPM pulse width would be high for say 3.6ms. at full throttle(5v) If rpm dropped at full throttle pulse with would be longer (555 timer) by a certain amount. Like I said very crude, and dont know how it worked because I never built it. Lawnmowers could not compete with my attention on the opposite sex.
Tim

 

[curious125]

I'm trying to do a similar thing, but with a two cylinder.
The main difference (i guess) is that i'm trying to make the system dynamic and do all the calculations on the fly instead of referencing a map. Is this a ridiculous prospect? or should I just accept that I need to use a map.

 

[Azmio]

Curious125,
I think it is a good idea. Anyway, here is what you need:-
1) Mass Air Flow sensor or Temperature+Manifold Absolute Pressure sensor (you have to choose either one)
2) Throttle body (cable or drive by wire) with potentiometer inside
3) Crank sensor with lots of teeth while at the same time having few missing teeth
4) Fuel injectors
5) Distributor or distributorless ignition coil
6) Lambda sensor

With these sensors and actuators combination you can have a pretty good closed loop engine control system. Of course, you can add in knock sensor, outside air pressure, coolant sensor, etc. to get your system more interesting.

Next you should get a programmable ECU and a Modac. If the engine is only for dyno and you really love programming, you can program your pc and buy the module to get all the sensors and actuators connected.

I personally love dynamic system because it gives more freedom in various operating conditions for example desert, rain, alpine, snow, etc. With the map based system, it needs to refer to the mapping and you will have to do a lot of parameter adjustment to adjust it to the operating conditions.

AO

 

[curious125]

My final goal is to have an 8 cylinder version of the system, but starting at 2 cylinders struck me as a good place to start.

I follow you with all the required sensors. I am curious about the crank sensor. 'lots of teeth, but missing teeth'? Would I not want 2 rings? one with lots of teeth, and another one with only a 1 or two teeth to reference TDC for cyl. 1, or even 1 ring modified to do both jobs.

I already have the processor that i'm going to use, and am slowly drawing control circuits each of the sensors.

 

[Azmio]

Curious,
Sorry for that, what I meant was the trigger wheel should have a lot of teeth to enable it to detect even the slightest movement. This is very important especially for more accurate operation. It will also enable the start up time to be reduced.
As for the missing teeth, new generation trigger wheel has more than 1 missing teeth to enable the ECU to detect whether the piston is at its top dead center or bottom dead center. This is very important during engine cranking.
If you do not have any missing teeth, you may need camshaft sensor reading to enable the ECU to superimpose the two signal and look for the BDC or TDC.
Anyway, all the best, dont forget to post more question along the way.