Craig Taylor said:
Hi all.
This is a quick pass at the design using an average of the injector voltage as an analog of the fuel volume.  It compares the air mass flow to the fuel volume flow and adjusts the fuel flow (via Control on the 555) to force the fuel to match the air flow. The signals are summed in the integrator made up of the opamp IOP1 and C5, R8.
 I ran a simulation to make sure it was stable.  The injector pulse smoothing filter values ( R7 and C7) are approximate, but they shouldn't be critical.  Since the constant pulse width, one pulse per rev approach gives a first approximation of the fuel requirements of the engine, allowing 1 second or more  for the feedback loop to adjust to the new MAF sensor voltage shouldn't be an issue.
For an actual system, some component values need to be optimized.  I used 15k and 200nF for the one shot RC time constant.  For a 555, the RC product of the components is very close to the actual pulse width; here it's 1.5E4 Ohms x 2E-7 Farads= 3E-3 seconds= 3ms.

That means that when the Control voltage is 2/3 of the supply voltage (6.66V in my case) the output pulse is 3ms.  This base value can be easily adjusted by varying R1.

You should be aware that increasing Control voltage above about 0.95xVsupply  will cause strange things to happen.  So don't make that voltage part of the normal operating range.  I would pick RC to be the middle of the expected injector pulse width range.  That allows the Control input pin to adjust the pulse width over a reasonable range.

Also note that I haven't make any attempt to "calibrate" the system.  I have no idea, for instance, what full scale ( 5V? ) on the MAF sensor output is equal to in liters/minute.  I don't even know if the MAF sensor output is LINEAR with air mass flow.  I sure hope so!

Craig Taylor

Paul responds:

Thanks for doing this Craig.

Q:Let me get this straight in my feeble mind. VG1 is a fixed frequency square wave generator correct?

     Right.  It simulates the pulse stream from the shaft angle sensor for a constant engine RPM.

Q:VS23 is the mass airflow sensor correct?

      Yes.  Here represented by a fixed voltage. Therefore  a constant air mass flow.

Q: Why not just connect the MAF sensor directly to the control pin?

      That would effectively multiply the pulse width by the mass flow.  But that's not what you want.  You want the amount of fuel squirted into the manifold to be a constant ratio to the air mass going in.  What this design does is to compare the air mass flow to the fuel volume flow and generate an error when the two don't match.  The error accumulates in an integrator,  which slows down the reaction to changes and so reduces the tendency for things to jerk around rapidly and cause instability.  But any error is eventually driven to zero by adjusting the pulse width to the injector(s).   That is, the fuel flow is varied until it reaches the preset A/F ratio.
   By averaging the injector pulses I derive a voltage that is proportional to the duty cycle of the injector drive pulses and therefore close to the volume/second of fuel flow out the injector.  There are errors, because of the injector dead time and other small variations, but it's probably good enough for a simple system.

It still needs the A/F ratio adjustment and some way to limit the injector pulse so fuel output isn't zerowhen the 555 output pulse goes below 1.x ms.  I was only trying to show the concept, not do a prototype design.  But it really isn't that far from being usable on a test stand. Calibration would require a squirrel cage blower (vacuum cleaner with a waste gate?) to drive the MAF sensor and a flow gauge to measure the fuel flow.  Also a surge suppressor in the fuel line just upstream of the injector to keep the fuel pulses from driving the flow meter nuts :-)

If you really want to try something like this, let me know.  The spice simulation I did said the basic design is stable and responds as expected, but quite a bit more work would need to be done to be sure it would be stable with rapid changes in MAF sensor outputs.

I'll run a simulation with the air mass flow changing and graph the resulting pulse width change to show some dynamics of the system.

Craig Taylor

I see. The MAF sensor output voltage is about the square root of the mass air flow. Correct?

What you are doing in effect is linearizing the output of the mass
air flow sensor. Correct?

Paul Lamar