Tuning with fuel trims

[68problemchild]

Just wondering, since short and long term fuel trims are pretty accurate, can you use the % change in fuel trims as a sensor to adjust overall fueling?

I have never gotten my 02 sensor to work but need to roughly dial something in as I am having lean codes on both banks.

Most of the time the fuel trims are not pinned so I have a real % reading to work with.

Sorry this is a newbie question.

Also my STFT are normally below 10% while long terms can get as high as 24%. Do I ignore LTFT''s and just tune with STFT's?

[mr.prick]

Use LTFTs if you're going to tune this way and
be aware that improper fuel injector settings will screw up fuel trims.

[WeathermanShawn]

If your Narrowbands (NB) are in good shape they can definitely be used in the closed-loop portion of a Tune. I personally use LTFT's as in essence they are STFT's integrated over time. But, you can use either.

[Taz]

Gentlemen, as always your advice is sound - for Gen III tuning. Patrick’s Camaro is using an E38 (6.2L L92 - aftermarket cam, LS3 intake and injectors) and T42 (4L70) setup. I have never undertaken the VVE tuning tutorial as outlined by Swingtan … but I believe that’s probably the only accurate tuning option for this particular setup ?

Gen IVs are not my area of expertise … merely thinking out loud …


Cheers,
Taz

[WeathermanShawn]

OP, check out the Swingtan's Tutorial thats applicable to your OP..

Thanks Taz..

[68problemchild]

I can't catch a break.

Anyone have any ideas on this unit? Searching for a V2 compatible O2 sensor.

http://forum.efilive.com/showthread.php?15957-02-meter-can-it-work-with-V2


Thanks for the input guys. Been trying to get the O2 meter to work for about 6 months now. Its been back to Innovate 3 times. No joy.

[swingtan]

Depending on the Gen IV controller being used....

The E38 and maybe others suffer from "LTFT Persistent Memory" the LTFT cells ( and others ) are not cleared on calibration updates and possibly even full flashes. This can lead to errors being introduced if they are either not eliminated all together, or are not taken into account if you are performing CL NB corrections. My advice would be....


Disable LTFT correction in the tune. Please note that this will affect the PE mode fueling as LTFT's are used in PE mode. So limit the use of WOT from this point on.
For the E38,
B1508 = 120'C
B1509 = -140'C
B1510 = 120kPa (for NA) or 256 (for FI)

Clear the LTFT data be either using the DVT controls in ScanTool, or performing an ECM reset.
Drive the car logging...
STFT's
Commanded AFR
MAP
RPM
MAF Frequency

When driving, use gentle throttle applications, and try and keep steady loads on the engine. Also vary the engine RPM and load but not in quick succession. YYou want long periods of stable loads. This is best done on a dyno, but can be done by going for a long drive. At least 30 minutes after the car is fully warmed up.
Take the log data and set up a map to suit what you are trying to correct ( VVE or MAF ). The STFT's are in a "%" reading, so you will need to create a custom PID to convert the +/- percentage into a BEN. Model your new map on the existing VVE or MAF BAN maps, but using the STFT-BEN as the data.
Perform the std Auto VVE / MAF process to apply the correction factors to the VVE or MAF.
Rinse and repeat till happy.


Be warned... This process will optimise your closed loop cells, but do nothing for the cells controlling fueling in the PE range. This means that even if you turn on the LTFT's again, their effect will be reduced in the PE range. While most factory tunes run quite rich in these cells, you may find that by reducing the LTFT effects, you end up leaning out in the PE cells.

This is why all "decent" tuners will recommend the use of a Wide Band controller to ensure the heavy load ranges do not run excessively lean.

Simon.

[nevinsb]

Great post! I used a similar approach when I tuned my truck with the LTFTs, and the part that gave me the most grief was creating the map with BEN values. I'll have to try the tutorial out when I get back this summer.

[68problemchild]

Thanks all! New AEM is scheduled to arrive tomorrow.

Simon, after I get the MAF dialed in, what table do I use to start tuning on for fuel? (Assuming we are back to Wo2 again and I get the new one working)
Also, on my table in scan, I have the options of showing max value, min values, or average. Do I use average?

Great stuff guys.

If this thing does not work, at least I have a plan!

[joecar]

Does anyone know the details on how to serially connect the AEM to V2...?

[68problemchild]

I found this but its lacking detail for the lap top and V2 wiring.
http://www.aemelectronics.com/Images/Products/Installation%20Instructions%2030-2310.pdf

Thanks!!!

[swingtan]

Simon, after I get the MAF dialed in, what table do I use to start tuning on for fuel?

Can I get that in English please?

RE: Table, are you highlighting data and then looking at the details? If so then yes, use the average, but it's not the best way. Using the BEN map will allow ScanTool to gather the data for all values and then average them automatically, across the entire log file.

RE: AEM controller, the AEM has a serial data steam and the V2 is already capable of understanding the AEM data. So all you would need to do is to use an adapter plg from the AEM to the V2 and to select "AEM UEGO" from the serial wideband option in settings.

TAQuickness did a tutorial .....

http://forum.efilive.com/attachment.php?attachmentid=3844&d=1220567224


Simon

[joecar]

I found this but its lacking detail for the lap top and V2 wiring.
http://www.aemelectronics.com/Images/Products/Installation%20Instructions%2030-2310.pdf

Thanks!!!In that doc, see Fig. 5.

Does the AEM come with a serial cable (that has a DB9 connector on one end)...?

If the AEM connects directly to a PC/laptop using that cable, then it will connect to V2 with a null-modem adapter used to join the AEM cable to the V2 cable (see the LC-1 example).

Does TAQuickness offer ready-made AEM cables...?

[68problemchild]

It came today.

No connectors just five wires.

OK so red can go to switched power. Black can go to ground.

That leaves:
WHITE - Connect to Lambda + Input.
BROWN - Connect to sensor ground. Connect to power ground if sensor ground is not
available.
*GREY – Connect to RS232 serial port on laptop/pc. See section on Serial Data
Viewing
*optional – only needed for laptop/pc viewing of data.

So grey to pin 5 on the RS 232 connector then to null modem then to top of V2 and that is it?
Do I need to tell the tuning software X volts = X afr? Or I just set V2 to "AEM"?
That would leave brown and white unconnected.

Thanks!

[nevinsb]

You are hooking this up to your laptop, or the V2? The V2, just take a piece of phone cord and attach the ground/serial to it. You don't need a DB9 or null modem. Think TAQuickness had instructions for getting it hooked up on his site.


http://www.taquickness.com/taqcables/TAQCables_sGEN.pdf

Just noticed too, mine had a blue wire for serial, the directions might need to be updated.

[68problemchild]

That is helpful.

Looks like

Phone cord/V2 to AEM

White to AEM Nothing
Red to AEM Grey (Serial - ??)
Orange to AEM Brown (Sensor Ground)??

CarPos switched to AEM Red
Car Ground to AEM Black

This scenario leaves both white wires unconnected. White on AEM is supposed to "Connect to Lambda input". Whatever that is.



PS, my phone cord has 4 wires. Red, Black, Yellow, Green.

[joecar]

Yes, it looks to me like you simply do this (in addition to power/ground):

AEM Grey -> V2 RJ12 pin 2 (Rx)
AEM Brown -> V2 RJ12 pin 5 (Gnd)

[joecar]

AEM White (Lambda +) is used for sending Lambda to the logging device (V1 or V2) as an analog voltage.

You don't need this, so just tape it off.

[joecar]

The pics/diagrams in this thread show pins 2 and 5 on the RJ12 plug:
showthread.php?8115-FlashScan-to-LC-1-Serial-Cable-Diagram (http://forum.efilive.com/showthread.php?8115-FlashScan-to-LC-1-Serial-Cable-Diagram)

take note, in the pics/diagrams the RJ12's lock tab is facing away from you.

[68problemchild]

Thanks Joe. When I put my "phone plug" into V2 have four phone wires (I am guessing this is an RJ12 plug with 6 potential wires) hanging to mate up: Black, Red, Green, Yellow.

Sounds Like AEM grey goes to position 1 wire on the "phone plug"?
and AEM brown goes to position 5 wire on the "phone plug"?

Thanks for all the help. Will wire in tomorrow and hopefully be rocking with Swingtans tutorial.

Unrelated, When I replace the wide band in one bank for one the NB, wont the motor run like crap with just one NB functioning or do they become irrellevant while tuning?

[68problemchild]

http://phone-jack.org/rj11

Mine only has four wires.

[joecar]

I thought it was pins 2 and 5 on the RJ12 plug...

watchout, an RJ11 may or may not fit properly.

[joecar]

Running with just one NBO2: if motor runs rough, this indicates that VE and/or MAF are not correct (whichever one the ECM is calculating airmass from).

[68problemchild]

RJ11 Fits but wires are in the wrong spot and only 4 of them. Need to order an RJ12 cable. None of the local stores have, not even Radio Shack. I cannot believe I have to special order this thing.....Well, another week gone by.....no o2's and no word on the meter I sent to Innovate two weeks ago, for the 3rd time!

[nevinsb]

Phone cables can be misleading, the one that connects to the handset is 4 pins, but it is a smaller connector than the land line. Might be easiest just to get one from TAQuickness if you can't find the RJ12 connector and a 3 pair cable at your hardware store.

[68problemchild]

Thanks. This info should help me locate what I need locally.

[68problemchild]

10249


This is the fuel STFT log per swingtan. This is as far as i got. Ben map??? What is a ben map and how do I create from this? Thanks guys. Judging by the pinned trims I need some fueling mods!!!!!


Thanks!!!!

[68problemchild]

Here is the tune....Any help appreciated.

[joecar]

I'll take a look when I get home later tonite.

[joecar]

You didn't get the serial wideband connection going...?

[joecar]

To create a BEN map, load up one of the existing ones, change the pid on the Data to the pid for your BEN (either analog wideband serial wideband, LTFTBEN, or STFTBEN), save the map with a different name.

[68problemchild]

Thanks. This should help alot. My STFT's are extremely high.

So with the Ben map, how do I cut and past the data on the the VE table? I assume this should be a multiple of a percentage or 30% = 1.3 X VE cell value. EG my trims indicate 30% lean so I multiply each cell by its corresponding % adjustment? (I assume the only VE table is Naturally aspirated(EFI?) B8101?)

Thanks.

No return yet from Innovate.
AEM unit has no RJ12 connector and same is on order from the internet to plug into V2. Nothing on either yet.

[joecar]

On the BEN map, in upper left corner, rightclick->Copy With Labels;

then in VE table, in upper left corner, rightclick->Paste->Paste with Multiply.

BUT, E38 has VVE tables, so follow Swingtans tutorial.

[68problemchild]

OK, tried paste with multiply and got some crazy numbers.
Paste with labels and "add" seemed to add the % copied from the STFT ben map. So a 1000 value cell in VVE and 30% SHTF gives the cell a new value of 1300 in VVE. Does this sound correct?

I think then I selected the "auto gen" box on Coeff's and then Generate VVE. Table did not change until I closed it out though.

Sorry if this is stupid. This is my first try at this and no O2 meter.

[68problemchild]

How does this new table look in B8101?

Also, if this process does nothing for PE, how do I mod PE?

[joecar]

Ok, do not multiply % values...

edit the BEN pid to produce a multiplier...

examples:
- if STFT is +5%, then BEN is 1.05,
- if STFT is -5%, then BEN is 0.95,

i.e. if you're using STFT from both banks, then BEN = (SHORTFT1+SHORTFT2)/200+1.

I think LTFT would give you a better result since LTFT is STFT integrated/averaged over some period of time.

Are you following Swingtans VVE tutorial...?

I don't have an E38 to play with, so VVE is not my forte.

[68problemchild]

Yes. Following swingtan. Pretty much the bible.

OK, you boys are pretty smart. I realize there is no ben map for fuel trims. How do I create a ben map? I can see ben maps for various 02 sensors but clueless how to pick one and then edit the formula to create a LTFT ben map.

Sorry about all these noobie issues.

Thanks so much.

Pat

[joecar]

Pat,

If your calc_pids.txt file is not stock, post it here, we can edit it for you...

[68problemchild]

Saw this link. It almost reads like in absense of data (eg in my case 02 data) it can adjust PE as well using LTFT's? Forgive me its all greek.
http://forum.efilive.com/showthread.php?15236-A-New-Twist-on-CALC.-VE-Table..Computing-the-Entire-VE-Table.&highlight=calc+ve

Doing a name search on my pc, I located the calc_pids.txt. Here it is:


# ================================================== ============================
# File details
# ------------
#
# This section defines various details about the file format.

*FILE

#Parameter Value Description
#---------- ---------------- ---------------------------------------------------
VERSION 7.1.1 File version
DECSEP . Decimal separator used in this file



# ================================================== ============================
# Units
# -------------------
# See sae_generic.txt for more information on the *UNITS section

*UNITS

#Code System Abbr Description
#-------- ---------- -------- -------------------------------------------------------------

None None "" "No units"


# ================================================== ============================
# Add slot definitions here
# --------------------------------
# See sae_generic.txt for more information on "SLOT" formats
#
#Units Low High Fmt Expression
#------------ ------------- ------------- ---- --------------------------------------------------------------



# ================================================== ============================
*PRN - Parameter Reference Numbers
# --------------------------------
# See sae_generic.txt for more information on the *PRN section
#
#Code PRN SLOT Units System Description
#------------------------- ---- ------------ ---------------- ---------------- ------------------------------------------

[WeathermanShawn]

Problem, the CALC.VE has been successfully tested out on the GEN III platform, but not on your OS.

As with >50% of modern-day Tuning it now comes down to how well one can use a computer..I.E. making BEN, MAPs etc..Swingtan's Tutorial will get you the results you need. The CALC.VE is pretty unique in its ability to target all phases of Tuning. Unfortunately I don't believe it will be applicable to your OS.

You are always welcome to use the instructions as far as cut, copy, paste, BENS, etc..but this method is a LS1/GEN III application.

Hope you find the answers you need. Good luck..

[68problemchild]

Roger that.

I guess what I am trying to do is change a polynomial with a multiplier which is fine with me if it works. The above process will be for another day. Happy to take a run at it at a later time.

In the interim, I will be thrilled if I can get a ben map created to produce cut and pastable data with the LTFT's. I have them turned off now. I probably should turn them on, turn off STFT's (location?) and get this new ben map installed. Hopefully you all can help with that.

Looking at my file above its empty.

THanks!!!!!!!!!!!!!!!!!!!!!!

[WeathermanShawn]

You should be able to turn on LTFT's quite easily. You don't need to 'do anything' to STFT's..in closed-loop they need to run. LTFT's are just a lot of STFT's averaged over time.

I have zero experience with your OS, so I hesitate to comment much. Having said that I thought almost as OS's came with a BEN Pid. Its the essence of Tuning. That file you 'attached' looked like an empty sae.generic file, not a calc_pids.txt file. What does your calc_pids.txt file look like?

Look in your computer here:My Documents\EFILive\V7.5\User Configuration\calc_pids.txt

Good luck..

[68problemchild]

That is my calc pids txt file in the location you described. I just cut and pasted the contents for simplicity.

[68problemchild]

So my file should not be blank?

[joecar]

Pat, where are you with this...?

[68problemchild]

Well I kind of dropped the ball since my O2 sensors came back. Now its back to trying to get them to work.

[rpmauto]

I dont have any experience with the E38, but I tuned two GTO's with the E40 last summer with the original version of CALC.VE worked awesome. Just obviously make the maps match the tables. I am curious about the E38, I might have to pull the tune on one and have a look.

[Taz]

The E38 is a totally different animal ... can't use traditional LS1 PCM style AutoVE or CALC.VE tuning techniques.


Regards,
Taz

[Dieselman]

Hello all,

How do I make a Calc Pid for STFT's and LTFT's to convert it into a Ben map for a E38???

This is my Calc Pid file - 11482

I went into the file but it may aswell be written in chinese :confused::confused:


Cheers

Ben

[Taz]

Hello Ben,

Have you read through Swingtan's E38 VVE tuning tutorial ? It details tuning the VVE and MAF.

Attached below ....


Regards,
Taz

[Dieselman]

Hello Ben,

Have you read through Swingtan's E38 VVE tuning tutorial ? It details tuning the VVE and MAF.

Attached below ....


Regards,
Taz

Yes I have done that and my VVE is ok apart from a few cells that I just can't get 100% spot on. I was just going to have a play with this and see if it is quicker to get the VVE close while being in the safety of Closed Loop (I have changed the cam)

My factory tune was actually actually putting in + trims through the whole rev range. I have a Tech Edge WB for doing my WOT work. I have talked to Swingtan and realize that the STFT will only work during closed loop and if LTFT's are not disabled it could make the PE cells either rich or lean.


Hope this makes sense!!!


Ben

[Taz]

OK ...understand what you are trying to achieve ...

Take a look at WeathermanShawn's CALC VET thread .... specifically the calc_pids.txt file posted ... obviously you can't use this technique with an E38, but some of the CALC PIDs may be of assistance to your current endeavours ... take a look at the CALC.LTFT, CALC.LTFTBEN, and CALC.SELBEN PIDs ... these examples may help you set up your own CALC PIDs for the E38 and your WBO.


Regards,
Taz

[joecar]

You will need to find out which E38 pids to replace some of those pids in that calc_pids.txt.

[Dieselman]

Thanks Guys :)


I have absolutely no idea how to make a CALC PID. If some one could help that would be great :help2::help2:


Cheers

Ben

[joecar]

Ben, no problem, I can help you with calc pids.

[swingtan]

FYI, you can't really follow Shawns Calc VET tutorial for the E38 as there is at least one important PID not available. In the E38, we have no charge temp PID to correct the air density for the VVE. We know that it's there and is used bu the OS ( unless it's disabled by the tuner ), but we have no access to check what value is being applied. I did try and reverse engineer a PID, but with two unknowns, it got too hard and I gave up.

The closest you can do is to use GM.AIRPERSEC and then set up a BEN with SAE.MAF. It's a lot easier to get the MAF calibration spot on, so do that first, then use the BEN to make adjustments to the VVE. It's not a lot different to just using a WB BEN though, so it's really a case of what seems to work best for you.

Another tip if you do want to "tweak" using the trims, is to watch the O2 voltages as well, especially when the engine is cold and the trims are not yet active. It will only tell you if the mixtures are rich or lean, but can help to pin point issues. There are still only 2 reasons that I use the trims for any corrections....


The tune is running CL and I don't want to drop back to OL just to make a couple of minor changes.
The is no WB data available and I really feel some VVE / MAF adjustment is needed to fix specific issues.


I know many say you can't / shouldn't tune without a WB, which is true when looking at the whole tune. You can however make small measured changes using the trims.

Ben, if you want to go through an excercise for this, let me know and I'll do something up for the next tuning day.

Simon.

[Dieselman]

Ben, no problem, I can help you with calc pids.

Thanks Joecar that would be good :rockon:




Ben, if you want to go through an excercise for this, let me know and I'll do something up for the next tuning day.

Simon.

That would be good Simon :) I wish I lived a bit closer to the city sometimes, it would make catching up a bit easier!!!

I don't know if I understand how the charge temp is calculated. Is it just calculating what the air temp entering the cylinder is???



Ben

[swingtan]

Basically, yes. The ECM uses IAT and ECT as start and end points for the charge temp calculations. Then using {B0179} and {B0180} it predicrs how much the air entering the cylinder will change from the IAT, toward the ECT. The idea is that the longer the air sits in the manifold, the close it will get to ECT. The ECM then uses this to sdjust the fueling for the motor, as warmer air is less dense and needs less fuel to maintain the same exhaust AFR.

Setting {B0179} to a flat value, will result in either a rich low air speed condition, or a lean high air speed condition. Sure, you can let the trims try and fix it, but that's not what we are about here. The only time I'd do this is if I wanted to actually see what the difference in fueling was like under similar conditions so I could set up the table correctly. You can probably understand then, that if you don't actually know the exact formula to use for calculating the charge temp, it's very difficult to try and correlate the MAF readings with GM.AIRPERSEC.

I just had a thought that the charge temp may be directly taken from the tables, but I don't think it is quite that simple.

Simon.

[Dieselman]

Thanks for the explanation Simon. I had a rough idea that was how it worked but you have cleared it up 100%

Maybe using STFT's would be ok to use when someone starts to do a VVE tune just to make sure they are not running too lean to start with ie if the trims are say +5 you could add 5% to the whole VVE table and go from there??


Cheers

Ben

[swingtan]

Yeah, you can. The basic idea is that, but in practice the trim percentage does not relate to a change in the VVE in a 1:1 format. It takes a bit of practice to judge the amounts.

[joecar]

+1 what Simon said, he knows E38/VVE :cheers:


I would use LTFT's since they are in effect numerically filtered STFT's (even if that means turning on CL/LTFT).

[Dieselman]

Yeah, you can. The basic idea is that, but in practice the trim percentage does not relate to a change in the VVE in a 1:1 format. It takes a bit of practice to judge the amounts.

Yeh ok that makes things a bit harder. Nothing is what it seems with the E38 :shock::shock:

Thanks for the explanations anyway :):)


+1 what Simon said, he knows E38/VVE :cheers:


I would use LTFT's since they are in effect numerically filtered STFT's (even if that means turning on CL/LTFT).


Ok thanks Joecar. I will just keep doing it the way Simon has outlined in his tutorial :cheers::cheers:

[swingtan]

Some more help on the VVE.

If you are making small to moderate changes to only one to three cells in a zone, you will find that the generation of the coefficients and VVE will pretty well revert the cells back to pre-change values. This is because the zone is seen as a whole to the coefficient generation and all cells influence the final shape of the zone. To get the values where you want, you need to over correct the cells you want to change and then blend into the cells around those ones. You also must remember to set up the zones so that common driving conditions fall as close to the centre of a zone as possible. For example, idle, cruise, acceleration and WOT should be areas that fall in the centre of zones ( as far as possible ).

Taking this into consideration, if you are trying to change cells at the edge of a cell, you need to blend into any neighboring zones as those will also affect the values of the cells you are trying to change.

If in doubt, over correct and blend....

Simon.

[Dieselman]

If you are making small to moderate changes to only one to three cells in a zone, you will find that the generation of the coefficients and VVE will pretty well revert the cells back to pre-change values


This is what has been happening to me I think. I make changes and nothing seems to happen then all of a sudden it moves the other way:doh2::doh2:


Very good tips Swingy