MAP ADC Input Smoothing

08/19/99....
    This one is similar to the timing period jitter but I know what is causing it.  Also I'm not absolutely sure if it is really a problem or not, or if I am being too picky (probably)....  The MAP sensor has enough bandwidth that you can see the pressure changes in the manifold as different intake values open and close.  200 mv to 400 mv of ripple is not uncommon.  That is 4% to 8% of ripple (2.32 inches, 19 ADC counts, or a 15% change in injector pulse width at 2500 RPM !!  It really shouldn't ever be this bad because the ADC is really read close to the same point in time every timing period.).  What concerns me is that when the ECU reads the ADC it is usually on the steep portion of the slope.  A little jitter here and the ADC readings are going to be moving around quite a bit.  Looking at some of the data I've logged, the MAP ADC values are always noisy, even under a steady load.  The average value is steady, with a lot of noise riding on it.  I have observed a 5 to 10 ADC counts of jitter.  This means what is being fed into the ADC could be as much as twice as bad as that because the the ECU averages the last two reading together before reporting them.  It sure would be nice if you could feed good data into the ECU, you know 'Garbage In - Garbage Out'.  One thing that may be aggravating this MAP noise is the timing scatter problem.  Fixing that one will help this one because it will remove the jitter from when the MAP ADC is read.  I've been looking at a way to try to filter the MAP sensor output.  The filtering on the ECU itself is very high frequency so it doesn't help us at all. The problem with using an filter with a low enough cut off frequency to filter the intake value pulses will also slow down the real response of the sensor to load changes.  This is true if you use a electronic filter or a mechanical one (like longer hoses or some restrictions to the flow).  Back to one of these 'Catch 22' things.  I don't want to slow down the response of the MAP sensor because that will affect how the ECU responds to rapid changes in load but I also want to filter out a fairly low frequency ripple at the same time. The only way that I can see to accomplish this is to use some kind of preprocessor on the signal.  Take an ADC and digitize the wave form, average it, maybe apply some more smart filtering and noise reduction routines, before passing it on to the ECU.  Have been looking at quite a few micro controllers here lately.  Need something with a fast ADC, a couple 16 bit timers, and enough power to do all I want of it at cruising speeds.  It also has to be small enough to fit inside of the ECU itself (or at least that is what I am shooting for).  It also would be nice if it was something that I can program fairly easily with having to learn yet another processor.  So far I haven't located my dream chip yet.  I would like to do this at the same time as I do the In Car Programming modifications.

08/22/99
    Thinking more about this, some of this noise may be just an artifact of the timing period jitter problem.  May not actually do anything with this until the timing period is stable.  Fixing that may make this noise so small to be a non issue (it may be so small now that it is a non issue anyway.....).
 

Below are some plots of what I am seeing so far........


    The above plots are what I have observed idling around 850 RPM.  The lower one is just the MAP ADC data expanded over the same period of time.  You can see at time the MAP ADC values are around 10 ADC counts or higher. (The 'Y' scaling is 10 times so 1,480 is ADC=148).


    This plot was taken around 2500 RPM cruising steadily down the freeway.  About a 4 ADC count pp here.



    This one was taken while the motor was idling around 850 RPM.  The falling edge of the trigger event is the 10 deg BTDC point.  Shortly after this is when the MAP ADC is read.  Slight uncertainty in the reading time and the ADC reading will move up and down the slope.  Looking at this plot alone would not explain the noise I see when I log data from the ECU.



These above two were taken with the motor is free running around 2500 RPM....


Just some more MAP noise around 1000 RPM...


These two are not directly related to any problem but are kind of interesting seeing the relationship between the TPS and MAP.  The motor was idling and the throttle was given a quick jab.