Weighting of VE vs. MAF

[johnsZ06]

Does anyone know the percentage of weighting the PCM uses to calculate AFR between the MAF and VE table during normal closed loop operation?

[caver]

My understanding is that it uses the MAF exclusively.

[Dirk Diggler]

Below 4000 rpm and unsteady map values the pcm blends both above 4000 rpm the pcm looks at the maf. THis was posted by a member on tech. Search for a thread called "ve Cracked"

[superls1]

Below 4000 rpm and unsteady map values the pcm blends both above 4000 rpm the pcm looks at the maf. THis was posted by a member on tech. Search for a thread called "ve Cracked"

This was posted by the guys from HPT almost 2 years ago. It has been internet gospel ever since. I am not saying it is incorrect, but I have not ever seen anyone validate their description of 'how' the code works. I wish Paul or Ross would lend some input into this discussion. The basics of what the HPT guys wrote is thus:



RPM > 4000 *
----------
*trust MAF completely and ignore SD calcs (apart from MAF sanity checking purposes)

RPM < 4000
----------
if RPM <; 2400 and MAP<; 84 kPa then
Steady MAP threshold = 0.0 kPa
else
Steady MAP threshold = 0.8 kPa

If (SteadyMAP) then
Calculate MAFAirmass/SDAirmass ratio (used for Unsteady MAP operation)
Correction Airmass = MAF Airmass (filtered)
else
Correction Airmass = SD Airmass x MAF/SD Airmass Ratio (calculated during Steady MAP conditions)

Transient Corrected Airmass = previous Final Airmass + proportion of Correction Airmass

Final Airmass = fn(MAF Airflow, previous MAF Airflow, prev 3 MAP readings, prev 3 TPS readings,
Transient Corrected Airmass)



The full description is here (http://www.hptuners.com/forum/showthread.php?t=2359&highlight=4000).

It would be nice to put together a couple of real world examples, indicate which tables are being used, how, etc.

[Dirk Diggler]

or better yet some type of pid which indicated the amount of smoothing/bias at different maps/rpm

[johnsZ06]

This was posted by the guys from HPT almost 2 years ago. It has been internet gospel ever since. I am not saying it is incorrect, but I have not ever seen anyone validate their description of 'how' the code works. I wish Paul or Ross would lend some input into this discussion. The basics of what the HPT guys wrote is thus:



RPM > 4000 *
----------
*trust MAF completely and ignore SD calcs (apart from MAF sanity checking purposes)

RPM < 4000
----------
if RPM <; 2400 and MAP<; 84 kPa then
Steady MAP threshold = 0.0 kPa
else
Steady MAP threshold = 0.8 kPa

If (SteadyMAP) then
Calculate MAFAirmass/SDAirmass ratio (used for Unsteady MAP operation)
Correction Airmass = MAF Airmass (filtered)
else
Correction Airmass = SD Airmass x MAF/SD Airmass Ratio (calculated during Steady MAP conditions)

Transient Corrected Airmass = previous Final Airmass + proportion of Correction Airmass

Final Airmass = fn(MAF Airflow, previous MAF Airflow, prev 3 MAP readings, prev 3 TPS readings,
Transient Corrected Airmass)


The full description is here (http://www.hptuners.com/forum/showthread.php?t=2359&highlight=4000).

It would be nice to put together a couple of real world examples, indicate which tables are being used, how, etc.

Very interesting, thanks.

Would be nice if somebody could validate that. :hihi:

[bink]

or better yet some type of pid which indicated the amount of smoothing/bias at different maps/rpm


Great idea- that would be really cool to see!! :D

[superls1]

or better yet some type of pid which indicated the amount of smoothing/bias at different maps/rpm

I would be happy with just an explanation to start, but a PID would be pimp!

[bink]

or better yet some type of pid which indicated the amount of smoothing/bias at different maps/rpm

Shouldn't there be a "DMA" type pid for this?? GMPX or Blacky?????
:cheers:
joel

[superls1]

ttt, Ross or Paul. Can you shed any light on this? Would it be possible for you to post some pseudo code that we could interpret for ourselves?

It seems as though MAF and VE are like High/Low spark where the computer is always interpreting between the two. It would just be nice to know if there are particular situations where one is favored over the other (just like the whole Charge Temp Blending factor). Inernet gospel says when using MAF, VE is for sanity checks and transition. Some go as far as to say above 4000RPM, use MAF exclusively.

FYI, saw this tech paper on MAF's posted on another site: http://delphi.com/pdf/techpapers/2000-01-0546.pdf (http://delphi.com/pdf/techpapers/2000-01-0546.pdf)

If anyone wants to discuss, I have questions after reading it. :)

[Blacky]

Very interesting, thanks.

Would be nice if somebody could validate that. :hihi:
Yes, that is correct.
Although the values described are calibrated values, so they may not always be 4000rpm, 85kPa etc..

Regards
Paul

[Dirk Diggler]

Yes, that is correct.
Although the values described are calibrated values, so they may not always be 4000rpm, 85kPa etc..

Regards
Paul


So that means we have access to change em?

[TAQuickness]

my gears are turning already....

[Blacky]

So that means we have access to change em?

Soon... (as soon as we get the web site back up again).
New design web site should be available tomorrow.

Regards
Paul

[johnsZ06]

Soon... (as soon as we get the web site back up again).
New design web site should be available tomorrow.

Regards
Paul

I'm not sure I understand. Is the ability to edit B0120 going to be added to flashscan or is it only available with a custom tune?

[Dirk Diggler]

Sounds like it will be added and editable jsut like the rest of the parameters

[TAQuickness]

So, effectively, we could tell the pcm to not report MAF errors (P0101, P0102, & P0103), zero out the MAF table, and raise B0120 to an insane number to run in SD?

If that's the case, that'd be slicker'n goose poop!

[Dirk Diggler]

LOL you want to make positively sure that thing isnt being used. LOL

Take a look at the descriptinos in the screenie Paul posted. You will see that if the maf is diabled via any of the MAF DTC's the PCM ignores that value and applies no correction...

[Blacky]

Also, re-read the description posted by Chris from HPT.

The "dynamically calculated" airflow (at <4000 rpm) is not a purely MAP based airflow. It is predominantly a MAF based airflow with a correction factor applied by comparing the difference between the sensed airflow (based on MAF) and the calculated airflow (based on MAP, RPM, TPS, IAT, VE table etc).

The correction factor is adjusted/computed only during steady state airflow. During high airflow changes (stomping on or getting off the throttle) the correction factor is not updated - just used to override the skewed MAF readings that occur in high delta airflow conditions.

The *only* time the PCM uses the MAP calculated airflow exclusively is when the MAF has "failed".

Regards
Paul

[johnsZ06]

I don't know. It almost seems pointless at times to even use a MAF except for the occasional sanity check. :bash:

[TAQuickness]

Also, re-read the description posted by Chris from HPT.

The "dynamically calculated" airflow (at <4000 rpm) is not a purely MAP based airflow. It is predominantly a MAF based airflow with a correction factor applied by comparing the difference between the sensed airflow (based on MAF) and the calculated airflow (based on MAP, RPM, TPS, IAT, VE table etc).

The correction factor is adjusted/computed only during steady state airflow. During high airflow changes (stomping on or getting off the throttle) the correction factor is not updated - just used to override the skewed MAF readings that occur in high delta airflow conditions.

The *only* time the PCM uses the MAP calculated airflow exclusively is when the MAF has "failed".

Regards
Paul


I thought about that on my drive home and all I could say was "DOH! I need to edit my post!". Would be nice to completely eliminate the MAF without having to throw a code or 3.

[GMPX]

I don't know. It almost seems pointless at times to even use a MAF except for the occasional sanity check. :bash:

My 4year old daughter thinks my MAF is a great toy, looks pretty high tech amongst the Barbie dolls. :muahaha:

Cheers,
Ross

[VetPet]

While the argument of whether or not the MAF is of any value could probably be discussed until the cows come home, I think the question of greater value would be, what does the MAF give us that using SD does not and are there situations where a MAF is better to have than not? We all have different levels of modifications done to our cars and live in varying climates and have different uses for our cars. The other part that's not accounted for and seldom brought up has to do with emissions. For me this is a big consideration since testing is a requirement for tag renewal. Here's a quote from a Corvette Fuel Injection & Electronic Management Book about the MAF.

"MAF does a better job of maintaining accuracy during dynamics, including acceleration and deceleration. When you consider aftermarket modifications, remember that the MAF sensor compensates for intake or exhaust modifications that increase air throughput. But the MAP sensor doesn't."

You could easily test the validity of the MAF. Take a stock engine and using the scan tool,measure the gms/sec at a steady state rpm, either at idle or a higher rpm. Based on the gms/sec reading you can determine the frequency using the MAF table. Now, install a cold air intake system. Measure the gms/sec at the same rpm and interpret the frequency from the MAF table. The gms/sec should be higher because of a more dense air charge from cooler intake air. If the MAF is responding as stated above, the frequency you get should be higher than before, which the PCM will interpret as more gms/sec and therefore make a fueling adjustment by pulsing the injectors for a longer period of time.

We already know that the VE table needs to be adjusted for changes in the air intake system or exhaust. Why? Because there is no provision for the ECM to make changes to the VE table even though all of the information for calculating the gms/sec using speed density is there. The VE table is back calculated from the gms/sec to arrive at the values we see based on the stock engine. The VE table is the backup for the MAF, not the other way around. If MAF calculated gms/sec gets out of whack in comparison to speed density calculated gms/sec the PCM will set a trouble code. This is your sanity check and I'm sure that there are guys out there that have installed a different cam, heads, exhaust or intake system and set the code after starting their cars. That's when the VE table needs to be corrected to reflect the increased VE or air intake(gms/sec).

I know we are correcting the VE table using a BEN factor to account for the increased VE and that afterwards we can do the same thing for the MAF sensor. I guess what I'm not convinced of is that we're doing it because the MAF needs to be adjusted or because there are still discrepancies in the VE table and we're patching the problem by adjusting what is essentially a fixed measurement device. We can probably all agree that we disagree.

Thanks for listening. :notacrook:

[johnsZ06]

Thanks for listening. :notacrook:

Not at all! A very good explanation of "maf madness" as I like to call it! :muahaha:

[TAQuickness]

While the argument of whether or not the MAF is of any value could probably be discussed until the cows come home, I think the question of greater value would be, what does the MAF give us that using SD does not and are there situations where a MAF is better to have than not? We all have different levels of modifications done to our cars and live in varying climates and have different uses for our cars. The other part that's not accounted for and seldom brought up has to do with emissions. For me this is a big consideration since testing is a requirement for tag renewal. Here's a quote from a Corvette Fuel Injection & Electronic Management Book about the MAF.

Running a dry shot. But even then, your MAF cal has to be accurate.


"MAF does a better job of maintaining accuracy during dynamics, including acceleration and deceleration. When you consider aftermarket modifications, remember that the MAF sensor compensates for intake or exhaust modifications that increase air throughput. But the MAP sensor doesn't."

Lot's of thoughts here. MAF does a good job of reporting during uni-directional steady state flow conditions. It does not do well under bi/multi-directional turbulant flow (hense the reason for a MAF screen).

Simply put, when you change the breathing characteristics of an engine, both the MAF and MAP will reflect the changes. The cars performance is only as good as the respective calibrations. For an easy way to see this, look at the MAF cal for a truck vs Corvette - same MAF part numbers, very different calibrations.


You could easily test the validity of the MAF. Take a stock engine and using the scan tool,measure the gms/sec at a steady state rpm, either at idle or a higher rpm. Based on the gms/sec reading you can determine the frequency using the MAF table. Now, install a cold air intake system. Measure the gms/sec at the same rpm and interpret the frequency from the MAF table. The gms/sec should be higher because of a more dense air charge from cooler intake air. If the MAF is responding as stated above, the frequency you get should be higher than before, which the PCM will interpret as more gms/sec and therefore make a fueling adjustment by pulsing the injectors for a longer period of time.

You are right. But, again, although you measure a higher output frequency, that doesn't mean the calibration table has a good value in it.


We already know that the VE table needs to be adjusted for changes in the air intake system or exhaust. Why? Because there is no provision for the ECM to make changes to the VE table even though all of the information for calculating the gms/sec using speed density is there. The VE table is back calculated from the gms/sec to arrive at the values we see based on the stock engine. The VE table is the backup for the MAF, not the other way around. If MAF calculated gms/sec gets out of whack in comparison to speed density calculated gms/sec the PCM will set a trouble code. This is your sanity check and I'm sure that there are guys out there that have installed a different cam, heads, exhaust or intake system and set the code after starting their cars. That's when the VE table needs to be corrected to reflect the increased VE or air intake(gms/sec).

Again, this is not 100% accurrate. In the case of the VE table, the PCM reads MAP & RPM (keeping it simple) then looks up a value from the VE table. At the same time, the PCM reads a frequency from the MAF sensor and looks up a value from the MAF table. At no time does the PCM actually measure a mass flow! In both cases, the PCM looks up and/or calculates mass flow based on various indirect-mass-air-flow inputs.


I know we are correcting the VE table using a BEN factor to account for the increased VE and that afterwards we can do the same thing for the MAF sensor. I guess what I'm not convinced of is that we're doing it because the MAF needs to be adjusted or because there are still discrepancies in the VE table and we're patching the problem by adjusting what is essentially a fixed measurement device. We can probably all agree that we disagree.

I'd bet 25 cents that if the PCM says the MAF flows 247.078125 g/s at 9500 Hz, that same 9500 Hz on a flow bench <> 247.078125 g/s and that if you plugged that flow bench value into the PCM, your trims/ben would still be off.

If you decide to retain your MAF, you will need to adjust both the VE and the MAF tables after modifying the engines breathing characteristics for proper operation. This is because the stock values are only valid for the stock engine hardware +- the 1,000% GM engineering tolorances.

The reason we correct the VE table first is: 3) it's easier with the tools available to us. 1) you can fully eliminate the MAF from the PCM calculations, but you cannot fully eliminate the VE table from the calc. 2) Once the VE table is correct, you now have a known good air flow table to reference your MAF to.


Thanks for listening. :notacrook:

you're quite :welcome: And thanks for entertaining my random thoughts.

[VetPet]

You're quite :welcome: TA. I'm kind of playing the devils advocate regarding the MAF and enjoying the feedback. I like to keep the wheels turning..... prevents rust.

:cheers:

[Highlander]

Can this be edited for e38 e67?