NoGo

iTrader: (2)

You have to go the other direction. Multiply by 103.

2349 / 30 = VE% 78.3
2349 / 28.96 = VE % 81.11

81.11 / 78.3 = 103.58%

Anyhoo, this is all pending that the formula is correct. Might want to wait until gameover and I get done duking it out to find the right formula.

HumpinSS

iTrader: (1)

Since the variation from the old formula to the "theoretical" new formula is 4 percent or 3.58 this means i need to multiply my VE values by 104% rather than 96%. Hope this is clear

NoGo

iTrader: (2)

Okay, cranked through more of the data again today comparing Gameover's equation to real data and my equation to real data.

As usual, its all grey area again.

It appears that gameover's equation is more accurate for lower MAP values and RPMs, while the Nogo formula is more accurate for the higher MAP / RPMs.

For lower RPM / MAP:
The NoGo formula yields a standard error of ~11% when tabulated against real world data on the same car.
The Gameover formula yields a standard error ~2% (damn good) when tabulated against real world data on the same car.

For higher RPM / MAP (and WOT):
The Nogo formula yields a standard error of ~4% (damn good) when tabulated against real world data on the same car.
The Gamover formula yields a standard error ~20% when tabulated against real world data on the same car.

Just thought I would share......

ChrisB, you've been awful quiet.

66ImpalaLT1

ChrisB has been MIA for almost a month. I'm trying to get a hold of him now about another PCM problem (unrelated to this.)

Brains

iTrader: (7)

damn...

gameover

please show an example calculation using your formula in detail so i can follow.

NoGo

iTrader: (2)

No problem.

I will do three points and indicate the error against recorded data.

Massflow = (VE * MAP * RPM * Displacement) / IAT

Point 1 - Typical WOT:
VE = 2349 m3
MAP = 100 kPa or 1 Bar
RPM = 5600
Disp = .00566 m3
IAT = 300 K

Massflow = (2349 * 1 * 5600 * .00566) / 300
Massflow = 248 g/sec ----- Actual recorded ~250 g/sec (less than 2% error)

Point 2 - Typical Idle:
VE = 1370 m3
MAP = 45 kPa or .45 bar
RPM = 900
Disp = .00566 m3
IAT = 304 K

Massflow = (1370 * .45 * 900 * .00566) / 304
Massflow = 10.33 g/sec ----- Actual recorded 9.22 g/sec (11% error)

Point 3 - Typical Cruise

VE = 1635 m3
MAP = 40 kPa or .4 bar
RPM = 2836
Disp = .00566 m3
IAT = 304 K

Massflow = (1635 * .4 * 2836 * .00566) / 300
Massflow = 34.99 g/sec ----- Actual recorded 37.19 g/sec (6% error)


That is pretty dern close if you ask me.
To obtain the same values using your formula simply multiply my results by 1.175

According your equation, the results and the resulting errors.
Point 1:
Massflow = 295 g/sec (16.5% error)

Point 2:
Massflow = 12.13 g/sec (17.5% error)

Point 3:
Massflow = 41.11 g/sec (9.6% error)


Repeatably, your equation will yield a significantly higher error for the data that I have. The only situation in which I get your equation to be more accurate is in the very high vacuum situation. However, I believe the accuracy of the MAF breaks down at these circumstances due to the very low flow velocity.

It doesn't matter to me which equation turns out to be correct. Either way it is bound to save time when converting to MAFless.

I am curious how you got your formula. Did it come directly from the PCM or did you have to do some interpolating. The only thing that makes me question it is the use of constants that I don't know where they originate from.

gameover

My formula comes from the PCM.

Some observations:
1. 2349 m3 is a very big volume (like really really big)
2. this 2349 VE number is not the actual number stored in the PCM like i said before
3. what makes you think the stock PCM's VE table is accurate? If your MAF fails you would probably prefer richer mixtures at high RPM than lean - hence tell the PCM you have more air entering than you actually have and let the fuel trims go -ve to compensate if they have to. Anyone who has tuned a MAFless car can tell you the stock VE table is out by quite a lot.

Sorry, but I cannot follow your equation. Specifically, i can't make the units of the result to be g/sec.

FYI, we display the VE table in % in the editor.

HumpinSS

iTrader: (1)

How correct are these forumlas


http://www.installuniversity.com/ins...n_9.012000.htm

NoGo

iTrader: (2)

I'm not saying what you have is incorrect, I'm merely trying to figure out why some things don't match up (for both your and my equation). Again, it doesn't matter to me who is correct, I would just like to get this figured out.

On with the show.....

It doesn't really make sense to question who's method is correct, because it is the same solution
The only difference between your solution and mine is a scaler of 1.175 (no matter how much algebra you go through)

One of us is off by one or several constants. My questionable term I believe is the displacement. Your questionable terms I believe are your two constants. The problem is, I don't know what either of your constants stand for. Where do these come from? What do they stand for? If they weren't just blind numbers I would be more inclined to go your direction.

Additionally, the formula that you present doesn't match any of the data that I have. It is off 10-20% for 80% of the VE table. The formula that I presented is 2-10% for 80% of the table.

1. Yes 2349 m3 is a big volume, but you can use liters as well, you just have to use liters every where else in the equation. I used meters cubed because the VE table doesn't have any decimal places. 2.349 liters works just as well when used with a 5.66 dispacement. It is just a matter of opinion (or scaling).

2. If the VE number is scaled by a constant like you claim, then it doesn't really matter what values are in the PCM. The formula will just have 1 extra constant.

3. Actually, the stock VE tables are very accurate for a stock vehicle. The VE table for the 98-00 are a little goofy down low, but the VE tables for the 01 and up vehicles are very accurate. Anybody doing MAFless tuning knows that you really only need to do it on modded cars, and of coarse modded cars are going to have a different VE table. Large cam cars shift the VE table as much as 30% in the lower RPM ranges. I know, we run a hybrid program of MAF, non-MAF to get the cars to run with the big cams.


Where do your constants of 8192 and 51.2 come from? What do they stand for?

gameover

i am not questioning who is right or wrong just trying to understand your formula (and it's surprising accuracy) because mine is based on the PCM code.

run thru the units of all the terms you use, the result is not g/sec. you have:

m3 * kg/ms2 * m3 / K which is m5.kg/Ks2

8192 is the multiplier the PCM uses to store the g/cyl number. likewise the 51.2 is the multipler it uses to store the MAP value. To work out things in PCM units you need to include the internal value scaling factors.

For example in the PCM the Cylinder volume of 0.708L is stored as 0x5AA6 or volume x 32768. Also, RPM is stored internally as RPM x 5.12 to give more resolution.

The PCM g/cyl calc for the VE table is:

g/cyl x 8192 = VEvalue * MAP * 51.2 / (IAT * 32)

if you compare this to the gas laws

g/cyl = Mn = M * P * V / (R * T)

so you can see that the VE table value is the M*V/R part of this times the scaling factors.

VEvalue * 51.2 / (32 * 8192) = M*V/R

VEvalue = (28.96 * V * 32 * 8192 / (51.2 * 8.3145) = 17833*V

where V is in litres.

This aligns closely with the PCM's calculation of g/cyl at Baro. Where it actually enters the cylinder volume directly and the pressure it uses is Baro.

deezel

Very interesting discussion. Thanks for sharing.
I have been trying to work out the units on all this math, too. I'm not taking sides here, just trying to figure out the math and units in my own way. Here's what I've got so far...

1) Ideal gas law used to calculate theoretical volumetric flow

PV=nRT or (m/M)=PV/RT and m=PVM/RT

now the units on the last equation:
grams=(Pa)(L)(g/mol)/(J/mol K)(K) and since J=N*m and Pa=N/m^2
grams=(N/m2)(m3/1000)(g/mol)/(N*m/mol*K)(K)

Using NoGo's first data point (5600 rpms) this gives:
grams = (100000 Pa)(.00566 m3)(28.96 g/mol)/(8.314 J/mol K)(300K) = 6.572grams
this is really grams per 2 revolutions since the full displacement is used but intake stroke is every other revolution, so that makes 3.286 grams/rev.
Now take 3.286g/rev * (5600 rev/min) * (1min/60sec) = 306 grams/sec theoretical air flow (or 0.8196 g/cyl)

The MAF from NoGo's data gives 0.6696 g/cyl so %VE would be (.6696/.8196) or 81.7%.
The VE table value of 2349 used with 2896 (100*molar mass of air) as 100% value gives VE = 81.1%.
NoGo's calculated value of 248g/sec gives 0.6643 g/cyl and VE = 81.05%

So far, it looks like NoGo's calculation is pretty close. But what about the units???

2) So what do the units of "VE" have to be in order for NoGo's equation to work out?

NoGo has massflow= VE*MAP*DISPL*RPM/IAT
And the units... (g/sec) = ("VE")(bar)(m3)(RPM)/(K)

So, "VE" has to have units of (grams*K)/(bar*RPM*sec*m3) for the equation to work. If you plug in conversion factors (1 bar = 100kPa ; 60 sec = 1 min), known constants (R=8.314(J/mol K) and M=28.96g/mol) and figure the displacement volume is per 2 revolutions you find something very interesting...

"VE" in NoGo equation = (g*K)/(bar*RPM*sec*m3)
= {(28.96 g/mol) * (100kPa/bar) / ((8.314 J/K mol)*(60 sec/min)*(2 rev/displacement))}* CONSTANT
= (2902.7)*CONSTANT

And what do you think the constant is? Its %VE.

I'd be willing to bet that the true 100% VE value in Edit is 2902.7 and not 3000 or the molar mass of air times 100 (2896).

3) I still haven't been able to work out the units for gameover's equations. Something about the PCM scaling factors is messing me up. Anyway, no offense to gameover, I just don't have another 3 hours to sit and go through the math right now...

Keep the discussion alive, please, gentlemen. I am interested to see if we can figure out units on gameover's formulas and reach a consensus on the meaning of the "VE" value in the PCM.

cheers,
deezel

87_ta

Sorry to but in thread, I have asked this question here before but never got an answer..
To make closed loop fuel changes (not global) do I Re work VE or re scale lower MAF tables? Or both? I do understand that VE is referenced during Openloop but what about closed? Where do I make my changes?

Or could someone point me to a thread that answers this?
I have much experience with OBD1 tuning in my 87 7730 ECM, Which is only SD so all of my changes are made VIA VE tables.
I am helping brother in law with his LS1.

Bink

Very interesting - Thanks for the labor!!
joel

gameover

The values for the VE table you are using are NOT the actual values stored in the PCM. The equation the PCM uses to work out CylAir is exactly as i have written it. Nowhere in the equation does it reference displacement. so the VE value must contain cylinder volume information.

The PCM unit scaling factors are how you store decimal numbers in binary, otherwise everything is integers.

The VE table is 5120 * V * M / R or the Effective Cylinder Filling Volume x 5120 x 28.96 / 8.3145 and amazingly 5120 x 28.96 / 8.3145 = 17833. The 5120 factor here is to increase the resolution of the stored number.

now look at NoGo's equation and check out the value of 5.12*28.96/8.3145, where 5.12 is the factor between the VE values you are using and the actual values in the PCM... that number looks familar right

100% VE for a 5.7L would be 0.708*5120*28.96/8.3145 = 12626 (or 2466 in your numbers)

Hope that helps...

HumpinSS

iTrader: (1)

So can we get a mathematical formula for calculating VE for use in EDIT for the mathematically impaired

NoGo

iTrader: (2)

HumpinSS I don't think we've quite ironed out which formula you want to use yet.

Gameover, still my concern with the formula that you have extracted from the PCM is its relation to real world values.

For Example:

Below is an extract of a stock VE table for a 2001 F-body. The numbers are displayed per 400 RPM (read from left to right and carrage return).
Using your formula to tabulate the %VE for a typical WOT, shows VE values at 100% for 4600 RPM and 96% for peak power. Now I'm sure you know cars, and know that if these cars came from the factory with that kind of VE we would never have to mod them. 100% and 96% is touching on race car territory.


MAP 100 69 78 82 83 83 84 86 87 91 94 99 97 96 93 90 89 89 89 89 89


Using the formula that I present will yield the below %VE values for a typical WOT. These look alot closer to what we see. Even going by the link posted by HumpinSS ( http://www.installuniversity.com/ins...n_9.012000.htm ) they solve for %VE as 76.57%. According to the table that I present, VE is 77% at this point.

MAP 100 58 66 69 70 70 71 73 74 77 80 84 83 81 79 77 76 76 76 76 76


Even though your formula comes from the PCM, I can't help but think it is still missing something. No matter what data I crunch it against or what VE table I make with it, it seems to be off by about 15%.

The biggest beef that I have with my formula is that back-calculating to the ideal gas law it needs R in order for the units to cancel out properly. It does seem to work rather well with actual data though.......

Bink

NoGo - why does your formula have to synch with the VE calc in the PCM??The PCM calculation/equation is for the use of a predetermined VE table - predetermined by GM. Couldn't they have tweaked it as neccessary to work with their mass produced sensors and PCM?? In which case your values would also need to be tweaked to work with their system (as would gameovers)?? Just a thought.
joel

Bink

Too Damn Funny...I know for a FACT you aren't mathematically challenged!!!
joel

NoGo

iTrader: (2)

Hey Binks,

I am wondering along the same lines. My next question is if backup operation is the same as normal VE operation.

Gameover,

How does the PCM function in back-up or P0101 mode? Are there any scalers that are applied to the VE table or is the VE table value used directly via your calculation.

I can understand the PCM using your calculation under the condition that it applies its own scaler. From your earlier statements (prior VE thread) it appears that the PCM does do this, as you illustrated below:

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)

In this case the VE table would act more as a ballpark value representing merely the proper slope and "shape" of the VE curve rather than the exact values during normal operation. As we have seen between your and my calculation, the difference in accuracy is 1.175.

However, it doesn't appear that the computer could properly apply fueling if it directly used the VE table during backup or P0101 operation. Perhaps it applies a fixed scaler if it didn't have a predetermined scaler stored or didn't have time determine one (ie: MAF unplugged at start-up....)

Is the P0101 or backup VE operation the same as normal operation VE determination?

Thanks