Air Flow Measurement

[Mazda Ebrahimi]

tom cloud wrote:
> 
>         [ snip'a'lot ]
> 
> >You are very correct about the definitive MAP reading part.  I have access
> to the logic
> >of a natural gas fuel injection system, and they use an averaging scheme to
> dampen the
> >MAP sensor input!  I wouldn't be surprised if other do the same to
> stabilize the MAP
> >signal.
> 
> So, would one route a small line from each runner to a central 'orange juice
> can' (a la Ford) or some other 'integrating' device?  Clearly, the time
> constant would be a consideration (too big a canister, too slow response).
> Maybe just the time constant of the lines going to a central collector might
> be enough averaging to give reasonable MAP.
> 
> Anyway, thanks for discussing this with me.
> 
> Tom Cloud <cloud at peaches.ph.utexas.edu>The best way is to find a spot in the main plenum volume that is not 
near any runners.  However, with some manifolds, like Ford's split 
plenum 5.0L or 5.8L, the signal is still unstable.  As you said, a 
vacuum canister would dampen the signal fluctuation, but it also slows 
the signal down.  But I don't think that is very critical.  Generally 
speaking, you should use the MAP primarily under steady state 
conditions, and TPS under transients, therefore the responsiveness of 
the MAP sensor is not super critical.  Any how, without having access to 
detailed code, it is hard to determine how your system responds to an 
erratic MAP signal.  Maybe they already have an averaging scheme built 
into the software!
An alternative to the canister approach is to use an orfice tube before 
the sensor (this is the approach we use), but essentially they both do 
the same thing.

This was a very interesting discussion. Best Regards, M.Ebrahimi