Why not use MAF for entire RPM range?

[LS1_Dragster]

I've tuned for both MAP(with MAF disabled) and MAF(400 RPM up) and the truck runs good either way. In fact I really cant tell any difference between the two.

So, why not run completely using MAF or MAP? Why mix the 2 or run the MAF from 4000RPM up?

Could you zero out VE table, set MAF at 0 RPM and run entirely using the MAF?

Just curious

[swingtan]

Short answer, Yes.

Long answer....

There are certain times where the MAF will simply not work correctly, some of these would be...


Big cam with high overlap: This causes air flow reverberation at certain RPM's that cause the MAF to read incorrectly.
Intake setups that move the MAF: Sometimes, relocating the MAF can cause measurement issues, especially if the MAF is directly in front of the TB.
The MAF tables in some OS's have maximum limits of 512gm/S air flow. In FI installations, the actual airflow can exceed this value, meaning you loose fine control of the fueling over this point.
The MAF "will" cause an amount of intake restriction. If you are going all out for power, you may decide to remove the MAF because of this. The amount is likely to be very, very small.
The MAF can introduce a "computing delay" for the fueling. The further the MAF is away from the TB, the longer it takes for it to read airflow changes. Because the MAP sensor is inside the manifold, it sees changes in pressure much sooner.


Some of the above are pretty marginal on the effects they would have. For me, I see two general rules for the MAF vs MAFless debate.....


For older ( slower CPU ) controllers like the LS1, MAFless seems to give excellent results when running COS3 or COS5.
For newer ( faster CPU ) controller like the E38, MAF based tunes work extremely well unless you run into the problems above. If the same sort of COS deail was available for the E38, this might change, but for a daily driver ( especially a manual ) a few of us have found that retaining the MAF makes for a more enjoyable driving experience.


It comes down to what the car is going to be used for. If you were talking about your drag car, I'd probably go MAFless no matter what the controller was. But I'd be trying both and seeing what worked the best at the track.

Simon.

[WeathermanShawn]

Dragster:

I have tried zeroing out the VE Table. Car (LS1) would not start. IMO it still looks for VE Values to start..both cranking and 'regular' VE.

You can run total MAF easily (non-idle). Where you might miss the VE Table's influence is on sudden throttle transients. It seems to take the MAF about 1/2 second to react..MAP appears instantaneous. I.E. Dynamic Fueling may be adversely affected. It is subtle, but important.

It is an interesting question that is not easily answered. And if for no other reason than a legitimate MAF failure, to have no values in the VE Table would be disastrous. So even though the MAF may be primarily controlling Trims, Spark, and steady-state fueling..you have to have a back-up if the MAF fails.

Good question..:)

[LS1_Dragster]

Great information guys.

Now, I think I read somewhere that there are certain transmission functions that will only work with the MAF operational like shift times D1108-D1110, is this true and is there more functions that get negated with the omission of the MAF?

There's a FI setup that incorporates the MAF in the throttle body. Wouldn't this give instant response or at least cut down on the reaction delay?

Could it be possible to run any size engine, any amount of cylinders and any amount of HP purely on the amount of air flow? If you new the amount of air being delivered you should be able to add fuel accordingly. To me if the MAF sees x amount of air flow it could represent a 4 banger at 50% throttle or a 632CI at 5% throttle and in either case it would inject the same amount of fuel to maintain 14.68 AFR or what ever you choose.

I'm just thinking out loud as I wonder why FI as it sits right now is fairly complicated with so many tables effecting each other. I've used the Holley commander 950 on several cars and it's so damn simple and easy yet works so well. And now I see so many self tuning setups and I'm wondering how they are doing it....

Lee

[joecar]

GM have made it like this (seemingly complex) so that it fail-safes itself when something goes wrong...

and it detects unsafe conditons and protects the powertrain...

and it helps you diagnose problems (DTC's and the various scanner modes).

[LS1_Dragster]

Anyone have info on this:

"Now, I think I read somewhere that there are certain transmission functions that will only work with the MAF operational like shift times D1108-D1110, is this true and is there more functions that get negated with the omission of the MAF?"

[WeathermanShawn]

Dragster:

Not sure on the tranny, but unless you run COS, most OS's will default to a Low-Octane Spark when the MAF is disabled.

Seems like EVAP is affected also, and I have seen some threads on Dynamic Fueling/Misfire Detection, etc..

A lot can be overcome with a COS, but it shows how MAF-enabled vehicles seem to be designed..well around the MAF..:grin:

[joecar]

I have observed effects that seem to indicate the shift times not following the tables while the MAF is failed...

[ I worded it vaguely because I don't have the data at my finger tips... but yes, I seem to vaguely remember this... my $0.02. ]

But, yes, various functions seem to disable while MAF is failed.

To find a way of running SD without failing the MAF would be advantageous.

[LS1_Dragster]

Yes, I'm running CO3 in all my cars!

It does appear that the factory relies on the MAF, maybe too much.

Lee

[swingtan]

If I was working at any auto manufacturer, the MAF would be "the sensor of choice". It's a great bit of gear and does so much in a single package. It doesn't measure "air flow" as such, but more along the lines of "intake charge characteristics". The "hot wire" measurement system instantly corrects for both air speed and air density at the same time. The "wire" its self has almost no "thermal inertia" so it reacts very quickly ( it doesn't help that the MAF is normally some distance from the TB though... ). When coupled with a fast ECM, the MAF is a very nice bit of gear and works very well.

RE: the use of a MAF on any engine. I think you'll find that different engines have differing combustion characteristics and will require differing amounts of fuel. While the same amount of air may be going in, burn rates will vary, optimal spark timing will change and even the RPM the engine is doing will alter the efficiency of the motor. In an ideal world, where all thing are equal, I think you are right. But once you change too many factors, like RPM,cylinder size, combustion chamber shape, valve timing/ lift, etc. then too much has happened.

Simon.

[redhardsupra]

MAF and SD are equivalent but not equal. They do the same job, but they don't do it equally well at the same set of conditions. So pick whatever works best for you. Personally, I like the hybrid model because of the double safety, but once I've came across a car that despite my best effort I could not make idle with MAF, so it's obviously not perfect. However, and this is very important point, once you pick a method, tune it right.

Here's my last take on MAF vs SD, I like it because it quantifies where MAF and SD shine and where they fall apart.
http://redhardsupra.blogspot.com/2010/02/maf-vs-sd-comparison.html

[LS1_Dragster]

redhardsupra - great info, thanks.

Now on to another question. Why couldn't you incorporate the MAF wires an circuitry right into the throttle body? Mentioned above, it was said there's reversion and turbulence at low RPM. If that's the case then so what, the stock settings use MAP under 4000RPM and MAF above. Now you have the best of both in a smaller and more responsive package. It seems that you could also lower the RPM when MAF takes over to sweeten the pot even more and to help tune large cams.

[redhardsupra]

_I_ couldn't incorporate wires into anything, because I'm a mechanical moron. I just mangle the equations and fondle the data ;)

On a more serious note, I think I can figure out a method to determine the optimal 'switchover' point over between MAF/SD mix. I feel a blog comin' on... thank you for the idea.

[swingtan]

As said, one of the biggest issues with retaining the MAF and modifying the engine is reverberation. In this case, the MAF works "too well" and can cause fueling issues. Big cams with lots of overlap can cause pressure waves in the intake that travel back up the intake to the MAF. These "waves" can be read by the MAF as the intake air slowing, stalling and even reversing direction ( not that the MAF actually indicates the reversal ). One method of reducing reverberation in the stock intakes was the use of "resonance chambers" in the MAF pipe. These would absorb pressure pulses and even them out to give the MAF a more steady signal. They also help overall airflow by reducing turbulence caused by the pulses. In a back to back test by one of the workshops over at www.ls1.com.au, it was found that the 2006 VZ SS intake gained 5kw at the rear wheels over the Holden GTS MAFless pipe. Overall intake size was very similar, but the stock VZ intake had the resonance chamber.

Hymey and I played with MAF activation a while ago and it did work to some degree. The idea at the time was to disable the MAF when airflow dropped into the idle area of the MAF. So if the MAF reported 3000HZ when idling, you would set the min fail frequency to around 3200HZ. The MAF would then fail and you would drop into SD mode for idle. I think there may be other ways to do this now though.

Simon.

[5.7ute]

_I_ couldn't incorporate wires into anything, because I'm a mechanical moron. I just mangle the equations and fondle the data ;)

On a more serious note, I think I can figure out a method to determine the optimal 'switchover' point over between MAF/SD mix. I feel a blog comin' on... thank you for the idea.

You have been a bit quiet lately.(JK)
Simon, once the maf fails using the frequency method, did it reenable once the maf frequency increased? Or did it stay disabled until the next key on?

[swingtan]

It re-enabled as soon as the test parameters allowed. Joel ( Hymey ) did most of the testing on his car and said that he may have felt the cut over between MAF and SD, but I can't remember now.

[DrX]

I have observed effects that seem to indicate the shift times not following the tables while the MAF is failed...

[ I worded it vaguely because I don't have the data at my finger tips... but yes, I seem to vaguely remember this... my $0.02. ]

But, yes, various functions seem to disable while MAF is failed.

To find a way of running SD without failing the MAF would be advantageous.

If that is the case, then running a SD tune could kill an auto transmission!

[Mr. P.]

If that is the case, then running a SD tune could kill an auto transmission!

It all depends on how the trans is setup; the way that we build our 4L60s we want these adaptive timing tables zero'd out anyways. With the factory 4L60 setup in trucks, I agree with you it would probably not last long!

[joecar]

If that is the case, then running a SD tune could kill an auto transmission!Yes, this is why it is very important to get the VE table correct (and avoid any portion of it being lower than actual)... i.e. if the VE table correctly allows the correct torque to be calculated, then the transmission will survive...

but it would be nice to have a non-failed SD mode.

[swingtan]

What would be "really nice" is to have a section of the COS that "reverse populates" the MAF data from the VE table. In this way, the PCM/ECM is fed a value that it thinks is coming from the MAF and continues to run normally. It might be that the existing gm/Cyl value, which is calculated from the VE, is converted to gm/Sec on the fly using the MAF calibration table as a correction. Or maybe the existing "speed density air flow per second" values can be used directly and the MAF signal just ignored.

Simon.

[DrX]

Yes, this is why it is very important to get the VE table correct (and avoid any portion of it being lower than actual)... i.e. if the VE table correctly allows the correct torque to be calculated, then the transmission will survive...

but it would be nice to have a non-failed SD mode.

The VE table will not be "correct" with respect to torque calculations if there has been injector scaling or if the base fuel pressure is different than stock. I have adjusted my trans pressure tables accordingly though. I have a trans pressure gauge and I can confirm that pressure does still vary with torque as commanded.

[joecar]

If fuel pressure is different than stock then the IFR needs to be recalculated to match;

if the IFR is correct (matches FP and injectors) then the corrected VE will correctly model the engine's air fill;

if the VE correctly models the engine's air fill then several things all work correctly (according to values populating these tables):
- spark timing advance,
- engine torque calculation,
- engine torque/traction limiting/protection,
- electronic throttle control,
- transmission line pressure calculation/control,
- transient fueling,
- and a few others.

Now, some large engines exceed the ability of the PCM to read MAF or to control injectors, so people will peform scaling to work around these limitations, but when this is done, those things I listed have to be addressed.

[redhardsupra]

What would be "really nice" is to have a section of the COS that "reverse populates" the MAF data from the VE table. In this way, the PCM/ECM is fed a value that it thinks is coming from the MAF and continues to run normally. It might be that the existing gm/Cyl value, which is calculated from the VE, is converted to gm/Sec on the fly using the MAF calibration table as a correction. Or maybe the existing "speed density air flow per second" values can be used directly and the MAF signal just ignored.

Simon.
I've done that, you can calculate the entire VE table from MAF, or vice versa. It's very simple, you can just do it with histograms. You set up a MAF calibration-like table, except for the airflow data you use not MAF data, but the results of the SD calculations.
getting VE out of MAF is a bit more complicated, but you start with setting up a VE-like table, and then for data you must use a custom pid, that takes MAF airflow (which is a lookup, and if it's calibrated correctly, then it's a correct lookup), and using the SD equations you can calculate VE (or GMVE) out of it. It's basic algebra, and it saves you a boatload of time. There is no need to calibrate the two tables separately. Although, doing it independently is a nice verification that both calibrations are correct, because if the resulting airflow from the two methods varies in a non-trivial manner, then you screwed up somewhere.

The one thing I'd like to work out in the future how to use both sets of data to figure out the cleanest data to use for the calibration. If we got two airflow estimators, why not use them both, and limit ourselves to just one of them? Any ideas?

[swingtan]

I've done that, you can calculate the entire VE table from MAF, or vice versa. It's very simple, you can just do it with histograms. You set up a MAF calibration-like table, except for the airflow data you use not MAF data, but the results of the SD calculations.
getting VE out of MAF is a bit more complicated, but you start with setting up a VE-like table, and then for data you must use a custom pid, that takes MAF airflow (which is a lookup, and if it's calibrated correctly, then it's a correct lookup), and using the SD equations you can calculate VE (or GMVE) out of it. It's basic algebra, and it saves you a boatload of time. There is no need to calibrate the two tables separately. Although, doing it independently is a nice verification that both calibrations are correct, because if the resulting airflow from the two methods varies in a non-trivial manner, then you screwed up somewhere.

The one thing I'd like to work out in the future how to use both sets of data to figure out the cleanest data to use for the calibration. If we got two airflow estimators, why not use them both, and limit ourselves to just one of them? Any ideas?


Yes, I've done that as well. Using the "VVE" in my case and setting up custom PIDs to get gm/S I was able to show the differences in airflow between the VVE and MAF, especially on transients. The problem is that that's only going to work in the Scan tool and has no impact on the ECM its self. It would be nice if the the OS could be set to feed the "MAFSD" data back into the OS to be used as the MAF data. Then things like Auto shifts would be easier to work out ( though still not perfect if you've had to bump the VE for changed injectors ).

On the auto shifts ( which I'm not an expert on ) I've been logging the estimated engine torque and then using that to "correct" the shift pressures. I can read the torque figures that the ECM is calculating and then I use that value for shift pressure adjustment. Fine tuning is then done in the force motor table. It seems to work OK, but may not be the ideal method.

Simon

[DrX]

The VE table will not be "correct" with respect to torque calculations if there has been injector scaling or if the base fuel pressure is different than stock. I have adjusted my trans pressure tables accordingly though. I have a trans pressure gauge and I can confirm that pressure does still vary with torque as commanded.

Oops! Should have said "if the base fuel pressure changes." Not that this is common, but I was thinking of my dual base pressure fuel system. As there is only 1 IFR table, there was compensation in my Boost VE table for the higher base fuel pressure when in boost.