Opinions wanted: Whats more important, smooth graph transitions or 0 % AFR error

[angrygoat]

If you have to make a choice between having a smooth graph transition in VE for a given block of cells and having a 0% error between a block of cells which do you choose and why?

[Ironmancan]

Subscribed.

[BBA]

I choose smooth with 0% error. It's not possible because every engine/PCM has minor differences in certain conditions. Just remember it will never be perfect and even when it's really close, it will only be that close under one set of conditions. Let the weather change and do it again and you will see different values.

[RWTD]

Exactly as BBA said!

Da MAF is always a good choice, FYI.

[angrygoat]

I def understand what you are saying BBA, you want to get the bets of both worlds. But, lets say you just finished logging and you look at you VE error and you have a row of three cells at say 2000rpm or whatever and go go and look at your VE table in 3-d and the number for the middle cell is higher than the third cell causing a peak but your afr error for all 3 cells is 0%. HOw do you handle that, do you leave it because all 3 are zero or do you smooth it and induce a neg or pos error?

[SilverG]

Great question! I usually try to smooth them out some but my truck "wants" some small peaks and valleys according to the logs.

Maybe a filter issue with logging? Do you filter out Fuel Trim Cells > 19?

[angrygoat]

is that like the DFCO cell or something? I disbale the CFCO and DFCO

[ktoonsez]

is that like the DFCO cell or something? I disbale the CFCO and DFCO

I believe he is refering to closed loop with the fuel trims enabled and using fuel trims to tune.

As far as deciding on smooth curve or zero % error from AFR, I would always prefer the smooth option in most cases. Here is some examples:

Acceptable Example:
for RPM 1600 and MAPS 43, 48, 54
Values of:
800 850 840

Bad Example:
for RPM 1600 and MAPS 43, 48, 54
Values of:
800 950 840

See what I mean? That spike in the second example is too big and something would have scewed your log. Just remember that these numbers basically represent a proportion of Air to fuel mixture and common sense dictates that the more air you give the car, the more fuel you will need to maintain the desired AFR.

There are certain cases with big to really big cams with 20+ overlap that do have a dip in the middle of the MAP curve for a given RPM but it is for at least 3 to 6 MAP columns not just one column. Here is an example:

for RPM 1600 and MAPS 31, 37, 43, 48, 54, 60, 65, 71, 76
Values of:
700 750 800 770 750 730 750 770 810

Hope this all made sense and helps you out.

Ken

[angrygoat]

it does, thats kind of what was confusing. As either with MAP or RPM increase, air increases and as well should fuel. Thats why I was confused as to why I would have a dip or spike in the middle of two cells with all 3 being 0% error. Could this spike or dip be due to sudden throttle transitions or lack of GOOD steady state recording?

Since we are on smooth VE tables. SHould the VE table have a bell curve shape should it continue to ramp up till redline and max kpa. I would think with a N/A car that efficiency would decline toward the upper RPMs and hence the VE numbers would decline but in a boosted setup, would the efficiency continue to climb?

[SilverG]

is that like the DFCO cell or something? I disbale the CFCO and DFCO

Yes sorry, I was refering to using Trims. I disable those also but if you watch the FTC it still is > 20 sometimes. The help file for Volumetric Efficiency talks abut FTC.

Thanks for the input.

[ktoonsez]

it does, thats kind of what was confusing. As either with MAP or RPM increase, air increases and as well should fuel. Thats why I was confused as to why I would have a dip or spike in the middle of two cells with all 3 being 0% error. Could this spike or dip be due to sudden throttle transitions or lack of GOOD steady state recording?

Since we are on smooth VE tables. SHould the VE table have a bell curve shape should it continue to ramp up till redline and max kpa. I would think with a N/A car that efficiency would decline toward the upper RPMs and hence the VE numbers would decline but in a boosted setup, would the efficiency continue to climb?

As long as its just a little hump it is ok, it could just be due to inconsistency in throttle movement like you mentioned or just simply be the MAP sensor not being super, super accurate.

If you look at my VE table from the first page of my EQ <> VE 3.0 thread, that is typically how the cuve should look. You'll see a larger incline on the edges at 15 to 45 MAP than in the middle and see if fall off and decline at the far edge in the 105 MAP area. Also, you will see all MAP columns start a decline after max torque RPM has been passed, again look at the thread I mentioned and you'll see what I am talking about. Your's might not look exactly like mine as far as the whole curve in general but you should see similar things at the edges and after max torque is reached.