Air Flow Measurement

[oecar1 at oec4.orbeng.com.au]

>Some of us dont have access to a dyno (hence the car window idea) ;)
>although I might be able to schedule some time in the wind-tunnel...

I guess I'm trying to demonstrate that even with access to all the
toys, it's still difficult (for me anyway!) to get working acceptably.

>Ok, why the heck does the airflow change directions?  Some sort of shock
>wave like in a 2cycle tuned pipe exhaust?  Wouldn't this just indicate lousy
>intake manifold/runner (whatever you all call them) design?  Seems it would
>give peaks and valleys to the torque curve (like resonances in the exhaust do).

The inlet system is usually tuned to optimise peak torque.  I can't easily
describe it this in detail, but imagine the momentum of the flow pulses used
to increase VE at peak torque are also being reflected off closed inlet valves
at other speeds.  I know from practice, its very dependent on cam timing.

>What do you mean by the freq response being asym?

The response to a step increase is faster than a step decrease in airflow.
This means that if you have pulsating airflow, the signal is skewed upwards,
including the average.

>I would expect the freq response of a MAF to be really poor compared to
>computational ability.

t63 for a 3 to 85 g/s step is about 5-10ms depending on the design.

>Ok, if I sample async and the sensor looks like a LPF w/ a corner at 1kHz
>it will roll off at -20dB/decade.  If it is resonant somewhere then put an
>electrical filter between it and...

  [snip...]

>...twice that, 2 MHz, and the aliasd signal gives less...

  [snippety snip..]

>...high order analog filters would cause).  Anyway, sampling wouldn't be a 
problem (and I could always use the signal in the analog domain).

A lot of this went over my head, but 2MHz is out of the question for an
automotive spec ECU.  I'd be lucky to sample at 1kHz, but I'd prefer to sample
at a lot less.  Remember you've got to do more than just measure airflow :)
If you just 'average' it with analogue circuitry (or even fast ADC and DSP),
the point I raise is that the average signal is not representative of the 
average airflow because of the linearity, response and rectification issues 
that I described before. 

>Ok, if the flow reverses direction, can't you just move the sensor
>'upstream' (even past the air filter)?  

I use an 1000 Litre volume between the airbox and the reference meter.  Even 
with this you will see some pulsations in the reference signal (another
MAF in this case).  Flow pulsations appear to be seriously hard to get rid
of.  I think you have to live with them and work your way around them.

>Sure seems a lot better than using MAP and some magical 'volmetric
>efficiency'.

I'd prefer to use MAP and learn (and adapt) the VE function using closed 
loop A/F control if I didn't need EGR (and lots of it - I'm working on a 
DISC engine project).

>How affected are MAFs by the relative humidity.

At 20degC, according to my calcs, the specific heat capacity of air (+H2O)
varies about 2% from 0 to 100%RH.

Andrew Rabbitt