MAF Reading normal??

[Bad Habit Z]

Here are some screen shots of of my logfile from a dyno run awhile back. I am wondering if the MAF (lb/min) reading is supposed to jump around like this at WOT?

[WS6FirebirdTA00]

post your tune so we can see what the MAF table looks like.

[12secSS]

Are you copying the Dyn Airflow to your MAF table? If so, are you using the average, max, lowest, or last values? I tend to use the average.

[Bad Habit Z]

Since my scans seem to be so random at high RPM's I am using the average and pasting them into a excel sheet so I can look at them and compare the curves. Right now the MAF is disabled but I put in the numbers that match the curve of Dyn Airflow from my logs (until it starts getting rocky)

[WS6FirebirdTA00]

hmm, maf curve is smooth, but it looks like a very steep slope. Did you not use a WB to tune it? Is this the stock descreened MAF?

[jimmyblue]

If you suspect the MAF, you should log the frequency
as well as the g/sec (or lb/min but English units are so
20th century). The MAF frequency will have a little jitter
to it; this is one of the things that the Dynamic Airflow
does, filters that.

MAF output (as seen by the PCM) can be kicked around
by electromagnetic interference. If you run into obviously
kooky frequency "sports" at repeatable points, suspect
this and try moving the MAF part of the harness away
from ignition and injector driver signals, any other high
current high frequency (relative to engine speed) wires.
A hundred Hz or so of frequency jitter is kind of normal
but seeing it jump (or drop) more than that, wants more
digging into.

[Bad Habit Z]

hmm, maf curve is smooth, but it looks like a very steep slope. Did you not use a WB to tune it? Is this the stock descreened MAF?

I have not messed with the MAF tune yet. I disabled the MAF until I can figure out what is going on with it.

If you suspect the MAF, you should log the frequency
as well as the g/sec (or lb/min but English units are so
20th century). The MAF frequency will have a little jitter
to it; this is one of the things that the Dynamic Airflow
does, filters that.

MAF output (as seen by the PCM) can be kicked around
by electromagnetic interference. If you run into obviously
kooky frequency "sports" at repeatable points, suspect
this and try moving the MAF part of the harness away
from ignition and injector driver signals, any other high
current high frequency (relative to engine speed) wires.
A hundred Hz or so of frequency jitter is kind of normal
but seeing it jump (or drop) more than that, wants more
digging into.

So a little jumping around in Hz is "normal"?

Could someone post a log of a WOT run with MAF Hz logged so I can compare it to mine?

[EC_Tune]

Here's a 2007 Vette (CAI only) on a dyno run in fourth gear.
The C6's don't filter the MAF value as much as the earlier cars/trucks so you can really see the jitter especially in the spark table.
Config enclosed as well...

EC

[Bad Habit Z]

When I try to open that file I just get an error saying "unable to parse log". Screenshot maybe?

[jimmyblue]

Here is a snippet off a supercharged truck. I plotted MAF
freq vs MAP*RPM (plain speed density airflow, less the VE
fudge factor and constants). Here you have a truck in
part throttle, midrange, showing about 600Hz scatter on
the frequency (approx 10% error band). This is a single,
part-throttle acceleration sequence so fairly continuous
(normal), no surprises.

Now if you can't take 10% worth of noise you might say
the tune is not robust. On the other hand more than 10%
noise indicates you could have something bothering the
piece, its wiring or the airflow presented to it (stumbles
or reversion will drop or exaggerate the airflow-as-freq).
Whether the MAF is freaking from airflow unevenness or
vice versa, is a debug exercise.

Interesting to note the larger dispersion at lower airflows.
This I think is inherent to the MAF transfer characteristic,
there is more frequency-per-flow at low flows than high.
Though you would think that the precentage ought to be
relatively constant this appears not to be the case. You
can see some larger "sports" in the lower airflow while up
top it comes together, tighter.

Which is probably one reason the MAF is weighted less,
down low, in the airflow calcs.

[redhardsupra]

MAFflow vs MAFfreq is a 3rd order poly function. that means at low freq values, the differences in flow are minute. at higher freqs/flows the function provides much better precision and resolution. just like jimmy said, this is probably the reason why at lower map/rpm, MAF is not used much, mostly for comparison with SD calculations. i find tuning idle much easier in pure SD much easier than in the stock MAF/SD hybrid mode.

jimmy: the MAP seems to be stuck at 102kpa, and there's MAF readings on a supercharged truck, this means you're tuning it without really accounting for MAP (therefore airflow). how are you calibrating your MAF then? injector flow is way too unstable without major modifications to the fuel system to be used for backcalculating MAF directly, even beyond their max limit. am i missing something here, or do you know a trick how to calibrate MAFs without proper fuel flow info and ignoring PCM limits?

[jimmyblue]

I think the guy is pretty well talked into going 2-bar SD
(being as the MAP is pegged and his dirty MAF is still up
at 11700+ Hz, and he wants to up the boost). Less certain
is whether he'll spring for a manifold referenced fuel pressure
regulator and associated plumbing, return-style fuel lines etc.
(which eliminates the IFR table shortcomings). Seems like
the multi-bar OSes ought to extend that into negative VAC
but don't... so make it have zero MAP dependence and
fuggedaboudit, I figure.

I think you need a 3rd order poly to fit the MAF reasonably
but doubt that the true nature of it is poly3; the curve to
my eye looks more like a logarithmic or maybe square root.
Whatever the math is for airflow cooling power of a hot
surface, I dunno.

[Bad Habit Z]

I see what you are saying about the maf showing errors down low, but on the runs I have logged it seems to be more accurate of a curve down low and when it gets into higher rpms it starts to bounce around?

[redhardsupra]

bouncing around is usually a function of descreened maf and/or dirty airfilter.
if you do a scatter plot of mafflow vs maffreq with a dirty airfilter, there will be a strong signal, and a lot of very obvious outliers. These outliers will screw with your averages obtained by using AFR%Error, unless you get A LOT of data, which is not easy to do >7000Hz.

This is why i prefer base my MAF tunes from a 3rd order poly trendline of dynamic airflow plotted against MAFFreq. you can make it even better if you filter out the outliers using statistical or robust estimation methods. I have a prototype spreadsheet that does it, if you wanna experiment a little.

email/aim me or something if you're interested in cleaning up your MAF calibration.

[Bad Habit Z]

Once you have the maf tune correct even though it shows to be bouncing around in the log would it still be ok to use a dry nitous kit? I would like to use the spreadsheet you are talking about just to give it a try.

Thanks,