Tuning techniques

[Dono]

Since you two are on this thread, and its kind of an sct tuning thread of sorts, Im going to do a quick hijack...

When you guys tuned load and boost, did you turn off WOT and follow this method of tuning:
http://info.efidynotuning.com/scaling.htm 'Scaling tables for BOOST (greater spark control)' by Decipha?

This seems quite interesting. I'll have to re-read how the a/f ratio is controlled under this method as we want to go rich under wot conditions and the narrow band o2's don't work under these conditions. I'm trying to learn what I can, and it seems to be one of those 'throw enough crap at a wall, and some will start to stick' things. I'm slowly gaining some small understanding.

 

[decipha]

henry stopped using WOT spark functions in the early 90s, so no worries, just simply spark down to borderline and dial it in

http://info.efidynotuning.com/saftot.htm

 

[2000StreetRod]

tuning methods

I probably read all of decipha's posted information on tuning and anything else I could find when I started tuning my Sport. I found a pretty good correlation of the functions decipha mentions and the CRAI8 source code I have. However, not so much with my CDE4 catch code calibration. I tuned using four different methods. Since I had the bench flow test results for my LMAF90 and thought it would be a good flow performance indicator I avoided changing the MTF to achieve desired AFRs. I also used the correct engine displacement to avoid incorrect load calculations since my borderline knock table is indexed by load and RPM. Unfortunately, because of the high flow of the Deka 60s and the low demand of my engine I was not able to achieve desired AFRs with injector slope tuning. I think next I tried altering engine displacement in the tune causing the PCM to miscalculate load. That didn't help much because my base fuel table is indexed by TP and RPM instead of Load and RPM. Another negative is that changing the displacement and therefore calculated load means having to adjust all of the spark timing. So I gave in (just to get the engine drivable) to using the method most commercial tuners use - inserting the manufacturer supplied injector characteristics and altering the MTF and then changing the spark timing.

I'm not proud of my current tune but it seems to work. I entered the actual engine displacement and my calculated intake manifold volume (includes the supercharger manifold and plenum. I restored the MTF to the bench flow results. I lowered the fuel pressure at the rail to the minimum I think is reliable with the Stealth 340 pump and my electronic PWM pump controller. I altered the injector slopes to get as close as possible to desired AFRs without having unpredictable things happening. And then I did the unforgivable offense of creating bogus commanded AFRs in the base fuel table to achieve the desired results when in open loop. I just estimated what might be needed for low TP values in the table because the PCM is normally in closed loop for them except at engine start and for a short duration (20 seconds) after.

At some point - maybe when I have to tear down the engine for some reason - I may purchase lower flow injectors that will allow me to achieve desired AFRs with injector slope tuning and using my accurate MTF.

 

[Dono]

Dale, I get it completely. I know I will run in to the same obstacles and probably use the same solutions. I wonder if this is a ford pcm thing, or a sct thing?

I love the idea of setting the base load without boost and then use the load values to determine boost and spark retard knowing when your in boost. I think that's genius. Using a map sensor attached to one of the analog ports as 4pointslow talked about earlier would eliminate the need to set a base line load of no boost by disabling the power adder. Then we would be able to log boost directly against load in logs. I have a boat load of learning to do. I hope my math skills are up to the task.

4pointslow seems to be having great success with the new injectors. I'm betting I will find that my 60's are larger than I'll ever need without tearing my block down the center also. If some of the other issues I'm expecting to encounter go away I will go down the road of smaller injectors also. In my case, with 2 extra cylinders and not making the same power as 4pointslow it stands to reason my injector duty cycle is at a lower percentage.

 

[4pointslow]

Tuning

Since you two are on this thread, and its kind of an sct tuning thread of sorts, Im going to do a quick hijack...

When you guys tuned load and boost, did you turn off WOT and follow this method of tuning:
http://info.efidynotuning.com/scaling.htm 'Scaling tables for BOOST (greater spark control)' by Decipha?

This seems quite interesting. I'll have to re-read how the a/f ratio is controlled under this method as we want to go rich under wot conditions and the narrow band o2's don't work under these conditions. I'm trying to learn what I can, and it seems to be one of those 'throw enough crap at a wall, and some will start to stick' things. I'm slowly gaining some small understanding.

You don't really tune load or boost, you usually tune the MAF transfer function and ignition timing. The air flow model gets altered when you add boost or change from a factory air intake system. Just leaving an air filter box loose on a 2003 ranger with 3.0L at work I found the fuel trims changed to about 14%, up from 0% +or- 5. Timing usually has to be retarded do to added temperatures in the cylinders from boost or higher compression.
You would/can also tune Fuel injector slopes if your fuel flow model is somehow incorrect or you just want to do it a different way.

As far as that post goes, that type of scaling is for when the amount of power made will exceed what the PCM can calculate and deliver in fuel, it has to do with the MAF transfer function. Older PCM's like ours have an approximate 63lb of air limit. To get around that you would scale things to make it work. I do not use that method on my truck, I don't need to. You probably wont need to either.

 

[Dono]

You don't really tune load or boost

Thanks, All this information leads me down all kinds of different paths.

 

[2000StreetRod]

adaptive learning

. . . you should never modify the lambses to get the wideband to spit out a different AFR, as soon as it goes into closed loop it will be correcting for that error and when adaptives have corrected your kam's will be all over the place and completely inaccurate
. . .

It is my understanding that adaptive learning is disabled when the PCM is in open loop. While in closed loop adaptive learning is based on the PCM comparing the O2 sensor outputs to the calibrated Stoichiometric Fuel Air Ratio. However, when using bogus commanded lambdas in open loop the Base Fuel Table Cold and the Commanded Lambda when Cyl Off table must be accounted for in addition to the Base Fuel Table.

If there is a problem when transitioning between open and closed loop I have not detected it.

 

[decipha]

no adaptive learning is not disabled in open loop

adaptives just can't update in open loop

in closed loop the ecu swings lambses around for the hego to switch, the hego will always switch at stoichmetric for the fuel your burning regardless if its e85 or straight gasoline

the only way to have accurate fuel control is to let the ecu reach its target each and every time by dialing the fuel on in

if you have a known or flowed maf curve go ahead and use it, wot a maf sweep to verify if the maf is correct, if your maf curve is correct everything from 2.5 mafv on up will have a linear fuel error, if it doesn't your maf curve is wrong, no if, ands, or buts about it

 

[2000StreetRod]

adaptive fuel tables?

I understand that the adaptive fuel function will assist when in open loop such as after engine start. I have been running with adaptive fuel turned off but will turn it on in the future. I have 3 different adaptive fuel tables in my strategy: Adaptive Fuel Table indexed by load and RPM; Adaptive Fuel Table AM indexed by air flow & RPM; and Adaptive Fuel Table Load indexed by load and RPM. The Adaptive Select Switch is supposed to select either the load or air flow index. I assumed that the Adaptive Fuel Table would reflect the position of the Adaptive Select Switch but it doesn't. If the switch only selects between the Load and AM tables it seems redundant to have the Adaptive Fuel Table.

 

[decipha]

yes it is, those tables are all the exact same, if you change a value in one cell the others will show it as well since its the exact same table

SCT only shows it 3 times for the different scaling axis dependent upon the adapt select switch

I always use airmass as i find its the most accurate

 

[2000StreetRod]

identical tables

I changed a value in the 1st table and it was reflected in the 2nd and 3rd. Then I changed a value in the 2nd table and it was reflected in the 1st and 3rd. Last I changed a value in the 3rd table and it was reflected in the 1st and 2nd. Thanks for letting me know there is really only one table with two different y axis normalizers.

 

[2000StreetRod]

LMAF MTFs

I input the PMAS and measured LMAF MTF data into Scilab and plotted them. I omitted the 16 volt point to avoid decreased resolution.
PMAS:

Measured:

I used Scilab because the data points are different and Scilab is supposed to be able to plot two matrices on the same page. But so far I've only been able to plot one matrix on a page. I'll try Excel next to compare the two MTFs. I wasn't able to make two plots on a page using Excel so I just entered all of the data and did a scatter plot.

Except for the two lowest values in both MTFs it appears the difference between the two MTFs is less than the size of the diamonds.