-
Senior Tuner
VE Table Filter Factors
As I peel away idle "actors" I still have this fuel-loop surging.
That is, I see MAP, O2 and injector pulse width swing wide
while RPM sort of follows along; spark, commanded AFR and
IAC counts are steady. I look at injector PW vs MAP*RPM
and it's a "cloud" where you'd think it should be a straight
line fit, given that RPM isn't moving much. IPW / MAP is not
real tight either.
So this says to me, another actor and possibly just the
instability (loop lag) in the airflow measurement itself.
I see there is a time-filter for VE. But I also see this VE
Correction Factor, Initial vs ECT Zone which appears to be
putting huge (30%) "correction" onto the VE at lower MAP
and lower ECT. Since this "corner" is where I have the worst
problem I'm wondering if anyone can 'splain me some things:
- This correction vs ECT / MAP - is it straight multiplying
the current indexed table value? Why do I want so big and
so variable a correction?
- VE Corr Factor Filt - is this change-per-tick? What have
you seen for "sensible maximum"? Does this only "smooth"
the applied "correction factor" or does it also "smooth" the
table-traversal?
- Other smoothing filters - what else might "lag out" the VE
value as the motor slides across the VE table surface? Dynamic
Air Filt? I'd like to know how these all interact and which might
be the most effective in tightening up open loop fueled, open
loop IAC idle airflow/fueling.
-
Senior Tuner
what about MAF filtering?
I've been playing with logging MAF raw vs MAF (SAE) and calculating cylinder airmass out of it. the resulting cylairmass charts are way more jumpy than the ones we get by logging the PID. the smoothness doesn't show up from nowhere, there's gotta be some filters for it, but i went thorough tunes for 99 and 01 fbody, 04 z06, 05 gto and i dont see anything about MAF smoothing functions. Is it not available to us, or is it not there at all?
-
HP Tuners Owner
if you are talking about normal MAF operation there is a complex predictive filter algorithm you are missing that goes about the job of combining the MAF, VE, MAP and TPS transient information. Contains about 10 coeff's and uses the engine operating zone concept. I have discussed this eslewhere but it's something we have never added for fear of trying to explain it in detail.
If you are talking about SD then there is no filtering, the value that gets looked up in the VE table is pumped right into the final answer of the dynamic airflow calculation. The only real lag is the MAP sensor respsonse time and processing delay.
-
Senior Tuner
thanks for clarification Chris,
can we see the predictive algorithm? i wanna see what kinds of nuttyness we're up against.
-
HP Tuners Owner
It goes something like:
Cylair(t) = K1* Cylair(t-1) + K2.MAF(t) + K3.MAF(t-1) + K4.MAP(t) + K5.MAP(t-1) + K6.MAP(t-2) + K7.TPS(t) + K8.TPS(t-1) + K9.TPS(t-2) + K10.TransientCorr(t)
There is a set of K1, K2... for each operating zone 18 total.
TransientCorr is either MAF (for steady state), or VE*MAFVECorrFact and MAFVECorrFact is the ratio of the MAF cylair to VE cylair.
Judging from most cals, the transient effect is lightly weighted compared to the MAF and other terms.
while looking for the info i found this blast from the past, which i guess is common knowledge thesedays
http://www.hptuners.com/forum/showthread.php?t=2359
-
Advanced Tuner
Is it something that is worth making adjustable, I.E somein has moved the Maf much farther away from the engine as in a Draw thru turbo/blower app?