NewCelica.org Forum banner

Tuning out Fuel Trims on the Emanage Ultimate

67K views 108 replies 13 participants last post by  marsrock7 
#1 · (Edited)
***Note -- this is a work in progress -- I will add to and edit this thread as needed. If anything isn't clear, just let me know -- it is admittedly a bit rough, but should be very helpful for those tuning with the EMU****

Street Tuning the E-manage Ultimate (EMU) to Minimize Long Term Fuel Trims (LTFT’s)

Please note that this thread is NOT intended to be a guide for tuning the EMU to achieve maximum safe power. That goal is best achieved on a dyno with the help of a skilled tuner.

However, the steps described in this thread will greatly facilitate your ability to extract optimal safe power out of your car with the EMU.


The Problem with Piggyback Controllers

The major problem with piggyback controllers is that they are always fighting the ECU in closed loop (partial throttle) for control of the fuel injectors (for a detailed explanation of why this is the case please read this thread http://www.newcelica.org/forums/showthread.php?t=84901).

If a given load cell in a piggyback map commands the injectors (either by spoofing air flow signals or direct control) to inject a quantity of fuel that will lead to a (narrowband measured) AFR that is greater or lower than the ECU is programmed to allow, the ECU will retaliate by trimming (adding or subtracting) fuel from those load cells.

The high fuel trims have two undesirable effects (1) they can royally screw up drivability, and (2) they can interfere with any changes made to your WOT tune.

This is because the ECU will determine (by means that are not entirely clear) an appropriate LTFT to apply to the cells that are activated in open loop (WOT).

If this number is relatively high (for sake of argument, let’s say greater than 9%) it may adversely affect the power (run too rich) or safety (run too lean) of your tune. Therefore, it would be desirable to minimize or eliminate all of the LTFT’s BEFORE tuning for optimal power.


More on LTFT’s

Contrary to what some of you may think, there isn’t a single LTFT in closed loop, but MANY. The Toyota ECU creates LTFT’s for small blocks of load cells in its map – I’ve found some evidence that there may be individual LTFT’s for individual cells! (Probably, they are eventually averaged so that you have less variation in the LTFT’s across blocks of cells, but this is all top secret ECU voodoo that few people know about in great detail).

Although you cannot see the Toyota ECU’s map like you can on the EMU (and there are several, not merely a single map), you CAN record all of the LTFT’s in real time while driving, and identify which cell on your EMU map is associated with each recorded LTFT. Then you can make changes to the corresponding EMU cell to minimize or eliminate the trim.

I’ll show you how to accomplish this task step-by-step.

Ideally, when you are done, all LTFT’s will be close to zero (for our purposes, let’s say < 10%, but hopefully closer to %5 or less)*, and the LTFT carried into WOT will be almost zero.


Before you begin, you will need the following items:

1. Laptop and most current version of EMU software.

2. An OBDII scan tool capable of logging ECU data from the OBDII port. I use the Auterra tool with my Palm pilot – for more details go here http://www.auterraweb.com/dynokitsnohotsync.html

3. Ideally, a wideband AFR gauge (although this is not abosolutely necessary to tune out fuel trims, you'll need it for power tuning and for monitoring the car).

4. Considerable patience

5. Go to a professional tuner if you don't know what you're doing. Err on the side of caution. Basically, if what I say makes no sense to you, just pass on these suggestions to your tuner...


Once you have the prerequisite items, here’s what to do before you begin:

1. Go to the parameter menu. Make sure the stock injector value is correct. If you had/have the yellow injectors that value should be 310 and if brown it should be 330. If you have new injectors of a different size, enter in the value for the new injectors you are using.

As to the inj lag time, if you have a value in there that is working for you, just leave it alone for now. If not, this will take some trial and error to set correctly, but I can tell you that the car won't idle properly if the value is too high (idles overly rich) or too low (idles overly lean). High fuel trims at idle are a sign that the inj lag isn't set quite right... As far as I can tell, when warm, it should idle around 14.7 - 15.1 AFR (or a tad leaner) and between 13.0 - 14.0 AFR when cold. More on this later...

2. Zero out the airflow map, at least through 2K in low load. Worthless. All it's doing is throwing things off. I would only mess with this if you are having idle problems due to your intake. Best to leave this zeroed though, IMO [**BUT, SEE NOTE BELOW].

If you are working with an airflow map that has adjustments to the high load cells, I recommend you leave them as they are for now, as they are probably keeping the MAF from maxing out.

I will have additional tuning suggestions added later for what to do if the MAF is maxing out and how to tell when this is happening (hint, if you are logging inj input duty cycles close to 100%, the ECU is not happy with you and you need to adjust airflow voltage readings in the affected cells and or use the boost limiter to clamp the airflow readings!).

3. Datalog your MAF voltage readings and re-scale your Y axis on all the relevant maps based on your voltage. Evenly spaced break points going up by .2 volts is a good starting point, but this will take some experimentation. The range is going to be 1 - 5v.

Likewise, scaling the x axis (RPM's) will take a little trial and error to find the right breakpoints, but evenly spaced intervals, though a good starting point, is probably not going remain consistent throughout the map.

Just be sure to scale all your load-based maps the same way.


Recording and tuning out Long Term Fuel Trims (LTFT)

1. Zero out all the low load (closed loop) cells in the inj map, if you haven't already done so. Again, make sure the airflow adjustment map low load cells are also zeroed out as well [**SEE NOTE BELOW]

2. Reset the ECU.

3. Start the car; get it up to operating temp.

Hook up your laptop and data log RPM and airflow voltage on the EMU. You may have the map trace on if you wish to help focus on specific load cells while driving.

Hook up the OBD II datalogger and record RPM and LTFT bank 1 -- these are the only two ECU variables you care about right now.

4. Drive around for at least 30 minutes (drving around longer is better -- see "HINT" below). Try to hit as many load points -- all in part throttle, closed loop, just crusing around -- as possible on the emanage fuel map. While doing this, record LTFT and RPM on your scantool.

At the same time, have the lap top recording RPM and, if you wish airflow voltage, (really, all you need is RPM). You may have the map trace on if you wish, but it's not necessary.

Note that the LTFT’s that are recorded on the OBD II reader change as you hit different load cells on the EMU map – this is because there are separate LTFT's associated with either individual cells or at least a cluster of cells in the ECU's fuel map.

BTW, also note that the ECU cuts fuel when decelerating (if you have a wideband, you will see the gauge indicate a severely lean condition in decel) -- therefore, don't worry too much about part throttle load cells activated when decelerating -- basically, you are temporarily in open loop and it doesn't seem to affect things much if at all. You can probably leave those particular cells (high RPM, lowest load), zeroed out.

HINT: If you have the time, it's better to drive for several hours (or even wait several days) so that LTFT's stabilize and THEN datalog. When datalogging, it's also worthwhile to record the Short Term Fuel Trims (STFT).

Eventually, the STFTs will minimize in value and should fluctuate back and forth between small positve and negative values.

Once the STFTs have minimized (don't be alarmed by an occasional large one), you know that the LTFTs should be stable, and time to make corrections.

The biggest factor that will affect this is the weather -- the best time to do this is when it's right around the average weather temperature for the time of year that you are tuning (i.e., don't pick the hottest day or coldest night of the year! Ideally, you'd do it on a day with low to moderate humidity and about 75* F.

Either way, you probably need to tweak things once or twice a year -- it's good to check the fuel trims at least once every couple of months).


5. Go home. Bring in the lap top and the OBD II datalogger to review the recorded data from both.

When you review the OBD II data, you should play it back as a graph displaying both recorded values (i.e., RPM and LTFT) at the same time.

It will look something like this:



In reviewing the OBD II data, take note of LTFT's and RPM on the scantool recording.

Pause the play back, and you can select individual data points.

Your goal is to match this OBD II data to the data recorded on EMU in order to locate cells that need tuning on your EMU inj adj map.

By eyeballing the shape of the recorded RPM curve from the OBD II data log, you will be able to easily locate the same RPM/load region on the emanage map that corresponds to the one datalogged with a given LTFT on your scantool.

Then it's simply a matter of selecting the datapoint on the EMU data log that matches the one you've selected on the OBD II datalog.

Observe (Click on thumbnail for larger images)…



Click on thumbnail for larger images.

If you click on an RPM data point in the recorded emanage graph that matches the same spot on your OBDII datalog, you will be able to figure out what LTFT was associated with the corresponding load cell in the EMU inj adj map.



Go look at your fuel map on the emanage. The data point that you clicked on in the EMU datalog will automatically highlight the corresponding cell in your inj adj map.



If the logged LTFT on the OBD scantool that matches the selected datapoint on the EMU datalog graph/inj adj map was negative, subtract some fuel from that cell. If positive, add some fuel to that cell.

If you find you get a wide variety of LTFT's in that same cell, you may need to add another break point in the fuel map for greater resolution.

6. Make small changes to your EMU map, flash it to the EMU, and drive around as you did before, once again recording with the laptop and OBD II datalogger as previosuly described.

Repeat these procedures until all LTFT's are in the single digits (as close to 0.0 as possible).

7. Now comes the scary part: Reset ECU and do it all over again -- you need to do this to help ensure that the LTFT's will always be low even if the ECU is reset -- remember, the ECU is always adjusting itself, so you want your tune to allow that to happen without high LTFT's being created unexpectedlly.

If your tune was good, after driving around for a little while, the LTFT's should stabilizie, and be very close to where they were previously (i.e., low single digits).

Be sure to give it a few hours of driving to ensure that all the LTFT’s have stabilized.

It will always vary a little bit due to ambient temp, but ideally, it will remain in low single digits or close to zero under all operating conditions -- that's the goal!

***NOTE: What's most important is that the LTFT that is carried into open loop is close to zero (mine is < 1%). If you have a few cells where the LTFT is a bit high, but drivability is fine and the WOT (open loop) LTFT is low, you're golden.

To determine your LTFT in WOT (open loop): Floor it and datalog the LTFT. Whatever that number is should remain fairly constant once LTFT's in the rest of the map have stablized.***

This whole process can take hours or even DAYS, and it is one of the main reasons why folks generally hate working with piggybacks.

That said, if you can endure it, eventually you can get the car running very smoothly off and on boost.

*Note on fuel trims. Keep in mind that the fuel trims correspond to a change in inj duty cycle that moves the AFR to a value the ECU will accept.

HOWEVER -- it should also be noted that if you are using larger than stock injectors, the actual effect it has on your output duty cycle will be less than the ECU claims.

In other words, if you are running injectors that are 50% larger than stock, the actual effect of the duty cycle change commanded by the ECU will be less (not exactly 50% less, though...).

The ECU has no idea you are using larger injectors; it's merely additng or subtracting fuel until it sees a narrowband measured AFR voltage it likes. The amount reported reflects the amount of correction ECU "thinks" it has made based on it's mistaken assumption of stock injectors.

Anyway, I'm pointing this out because in the afforementioned example, a recorded LTFT of 5% would probably really be only create a 2-3% correction.

Using the AFR target Map to maintain desired AFR's in WOT
Coming soon! Short version is, by creating a closed loop map on the cells activated in WOT, you will be able to more easily keep your AFR's where you want them even when there are dramatic changes in weather conditions and even if the (small) fuel trim that gets applied to WOT changes slightly.

Further Improving Drivability: Tuning Injector Acceleration Enrichment

Quick inputs to the throttle will result in an abrupt change in the AFR as a large mass of air is ingested. Unfortunately, this may create a momentary lean condition and hesitation off idle. This is corrected by injecting extra quantity of fuel for a few milliseconds. You may also tune this feature with the EMU.

If you are getting hesitation off idle, use the map trace on the injector acceleration enrichment map, and take note of which cells are activated when it hesitates. You'll most likely need to add or subtract fuel to those cells.

Drive around for a while, somewhere where you can stop and start, applying different degrees of throttle to take off -- a big, deserted parking lot would be a good place to do this. This is tedious and will take a while to dial in, but you can get the car to have excellent throttle response with trial and error.

How to Use the Emanage Rev limit Cut Setting

For details, please see this thread http://www.newcelica.org/forums/showthread.php?t=275482

NOTE ON AIRFLOW CORRECTION
If your intake tube has an inside diameter that is larger than the stock 2.5", you might try making a global correction to the airflow map (at least in map cells activated in part throttle). The global correction should be approximately equal to the difference in the diameter of the intake tube from the stock diameter.

Example: The stock tube has a 2.5" inside diameter (i.e., 63.5 mm). If you are using an intake with a 3" inside diameter (i.e., 76.2), that's a difference of 20%. Therefore, you would make a correction of 20% in all of the cells of the airflow map, or at least those that are activated in part throttle (in WOT, especially if under boost, you may or may not want to do this -- I'd determine that empirically).

If you need to make this correction, it should be done BEFORE tuning out any fuel trims or tuning for drivability or power.
 
See less See more
4
#2 ·
Awesome!!!

I was planning on making a Greddy supercharger FAQ that should be sticky worthy when I come back from Florida on Wednesday. I'm definetally going to include this. Too bad I'm here in Apopko, by Orlando, and not where you are jlit. :( I would've definetally liked to see your car. Drives to go ANYWHERE down here are longggg man. The only incentives to living down here is the fact that you don't have emissions, and that the roads and weather are WAY better to own supercars versus NY.

Since yesturday (12/30) was my birthday, matadorgts came through to a North East meet, and I got to get a ride in his car :D . He posted that he upgraded to a 10 psi pulley recently, so when I met up with him, he said he came up with the new tune HIMSELF! Keep up the good work jlit :thumbup: I can't wait to go back and learn all about the emanage from him while he's here in NY, and not busy with his business trips again.
 
#3 · (Edited)
Eventually I need to expand on this to talk about setting injector lag times (I think I figured out a good way to do it), monitoring knock, tuning airflow adjustment to keep the MAF from maxing out, using the AFR target map, etc.

I also hope some of the more experienced EMU and EM Blue tuners will offer useful suggestions to be added to this thread.

I also welcome corrections or suggestions to anything I've posted from those with more experience in engine management -- again, this is a work in progress ;)
 
#4 ·
Hi!

First of all, I wish to all of you a happy new year.
Second, this is a very nice post from jlitman. As always, his advice is knowledgeable and precisely to the point. I am very eager to see all of his updates, especially those related to knock monitoring.

Moreover, some discussion regarding the VVT tune when operating with superchargers would be of much interest but this is not directly related to this thread...

Keep up the good work!

G.
 
#5 ·
Thanks :)

I'm still trying to decipher my knock signal output -- it reports a percentage of knock activity and the highest I've ever logged is about 55% (usually 0% to low teens) -- but is that high or low? Is something close to 100% activity = to knock? :confused:

I never see any activity that makes me think the car is knocking (i.e., a huge spike that is way greater than the rest of the recorded activity) and there's been no evidence of pinging since the lower comp pistons and the W/I has been fixed (it was clogged at one point).

What I really need is to see what other people are logging to compare and contrast -- it would be really nice to see what a stock 2ZZ does for comparison.

I'll see if I can make some sense of a recent datalog and will post the screen. I'll have to see if I can output the data into an excel file or something... maybe I can compute mean knock activity in a given load region and see what the SD is :shrugs:
 
#6 ·
Hi!

Good work must always be credited jlitman!

As far as value deciphering is concerned, maybe you should log ignition timing along with the knock activity value. The main goal should be to relate ignition timing alteration with knock activity.

One procedure that comes to mind is the following:

- Program EMU setting in order to be on the "safe side" knock-wise. (Alternatively you could use some race-gas or modify W/I settings).
- Reset ECU
- Drive around in order to log ignition timing in order to write-down the "stock" ignition timing (when the ECU does not need to pull timing due to knock)
- Reset ECU
- Program EMU the way you had it programmed BEFORE logging stock ignition timing.
- Drive around in order to log ignition timing. This way, the knock values you mentioned (like the 55% value you had) will eventually appear and you will be able to see if the ECU pulls timing.

I think that having this data will enable you to check when the ECU "thinks" that knock is dangerous for the motor (and thus will pull timing and add fuel to protect the engine). In other words, you will be able to "guess" the ECU threshold for pulling timing.

I hope my suggestions help.

Best regards,

G.
 
#7 · (Edited)
I need to have the EMU wired into the crank sensor -- I'm not convinced that what I'm seeing on the scantool is the actual ignition timing -- that is, I think the ECU is unaware that I'm adding or subtracting timing because the EMU commands the coilpacks directly. I saw no evidence of timing being pulled the day my piston went kablooie, though clearly I added 1 degree too many...

Anyway, so far I've seen no relationship between logged ign timing and any knock I've logged, so my guess is it's all pretty much below the threshold, especially with my water injection -- and I'm not turning it off just to try and induce some knock for the purposes of an experiment.

If you are interested in messing with it, go for it. I'm sticking to offering advice on tuning fuel trims for now.

But in case you were wondering...



No evidence of timing being pulled at the one moderately large spike. I'm guessing noise has to be closer to 100% before the ECU freaks out -- someone who knows better correct me if I'm wrong...

Anyway, this is a second gear pull on the road.
 
#8 ·
Hi!

Nice graph you have there. It is quite useful as to what someone would expect to see when logging knock activity.

Please, by no means am I suggesting to "induce" more knock than you already have. I just recommended "safer" settings and ECU reset in order to check for ignition advance timing adjustments compared to your current settings. Maybe if you share how you managed to log knock activity with the stock sensor, I could help by applying the thoughts of my previous post to my own car... ;)

Since it's known that superchargers induce further noise to the engine and considering the shape of the graph, maybe some data processing would be advisable in order to obtain more meaningful data. It seems that knock activity gets higher when RPM gets higher and that there is more or less a certain oscillation amplitude; these are some facts that can be used in order to "normalize" knock sensor values...

Keep up the good work!

G.
 
#9 · (Edited)
^^^ Yep -- the S/C definitely creates more noise, so a lot of the logged activity is a reflection of that.

I already did this on the dyno -- adding timing up top had no noticeable effect on recorded knock levels. Still read between 0 and 60% activity in the same places.

I think I really need to see data from another car to compare and contrast or be able to measure true ignition timing, assuming that's not what I'm getting from the ECU.

I had my E5 (Knock sensor) pin from the ECU wired into the EMU harness -- can't recall which pin gets it tho'. Check the EMU pinout, which can be found on mohdparts (and probably on here as well).

Settings are detailed in another thread by me on knock sensor activity -- see here http://www.newcelica.org/forums/showthread.php?t=261787

I'll take a close look at the ignition timing that I data logged in relation to that, but I'm pretty sure there was no evidence of timing being pulled -- It just added it in a linear fashion up to (a reported) 22 or so degrees of advance.

The only place I ever see timing drop a bit is at 4.5K, but it always does that even when 0% knock is logged, so I think that's just part of the ECU's factory tune :shrugs:

BTW, that one spike is right around lift cross over, so I bet it's just noise.

EDIT: I'll do some more data logging, but I just looked at the OBD II data associated with the above graph and there is NO relationship between timing and the knock sensor activity I recorded. I think all of that activity must be below the threshold.

EDIT: I think anything > 80% activity will result in a degree or two being momentarily pulled by the ECU.
 
#11 ·
I gotta lotta stuff to learn but it's something i like doin' especially on cars. I'll probably get the pc version for the scantool though since i have a laptop for the emanage blue. Good job though on being very detailed on all the info. and including visuals.
 
#13 ·
I mentioned a couple of times to individuals here and else where that a much cooler way to tune LTFT and STFT was on its way, well, now its here. I've pasted in the full announcement below, but de-emphasised the less relevant bits.

I worked on the design of the software with Graham from GLM software on this, talking him through what I was after and why, so I know it turns out usable, useful results. Actually, even the early beta I tried kicked ass :) I will be buying a copy of the software (that was the deal - He wrote exactly what I wanted and I would finally pay for a copy of the software) and would recommend it any one that was wanting to do some tuning.

Also, you can get a trial version that lasts just long enough to be useful ;)

Avendit.

Dear OBD 2007 User,

We are pleased to announce the release of build 1.0.13.93 for OBD 2007 for PC. This
is a free upgrade to all registered users, available from the version check menu
option within OBD 2007.

This build of OBD 2007 introduces support for the Elm 327 v1.3 chip. The Elm 327
v1.3 is due to go into public beta July 1st, with a release in August. This build is
primarily for those of you who have applied to Elm Electronics to beta test the new
chip. We will have a final release when the v1.3 chip is shipped in August.

Features of the Elm 327 v1.3 chip

Speed

The Elm 327 v1.3 chip provides many new exciting features. The one feature that is
likely to impress most is the increase in speed when OBD 2007 is operating in
continuous read mode. For CAN vehicles the increase in speed is extremely impressive
� double the speed of the old 1.2 chip. In fact double the speed is quite a
conservative claim as we have seen speeds as high as 98 pids per sec when running at
234.4k baud. However speeds of approximately 40 pids per sec are quite normal for
either a USB scan tool or a serial scan tool. The speed increase for non CAN
vehicles is not as significant.

Buffer Full Errors

The other major feature which results in the elimination of buffer full errors is
the ability to control whether spaces are included in the output of the responses
from the OBD port. By turning off spaces, the internal buffer of the chip is
effectively increased, which eliminates the chance of a buffer full error. This
means that scan tools can now run at lower bauds without hitting the dreaded buffer
full error. This is very important as vehicle manufacturers are now starting to
exploit the CAN bus by increasing the amount of information coming from the bus.

Since we last wrote, OBD 2007 has been steadily evolving with new features added
with each release. The following is a summary of the highlights since our last
newsletter in March.

Build 1.0.13.8

In April we released build 1.0.13.8, this build included the new Auto Discovery
mechanism for the Com port connecting your scan tool. Instead of having to manually
select the correct Com port for your scan tool; OBD 2007 now does this
automatically, although the option is still there for experienced users to manually
select the appropriate port.

In the same build we enhanced the Dashboard tab by adding support for two more
gauges, Intake Manifold Absolute Pressure (MAP) and Fuel Rail Pressure.

Build 1.0.13.9

In late May we released build 1.0.13.9 which introduced a new module called OBD 2007
Mapping.

OBD 2007�s Mapping module allows you to examine data from your vehicle in a three
dimensional format. The Mapping tab is the one module of OBD 2007 that does not
require a connection to your vehicle. You are not examining live data, but data that
was pre recorded in an earlier OBD 2007 connection session.

When you are viewing live data of the pids of your vehicle in a normal OBD 2007
connection session, you have the option to record the live data to an OBD 2007 log
file. At the completion of your logging session, you will find a text file in the
OBD 2007 program folder that is date and time stamped for that session � for
example OBDLog10052008 153422.txt.

This text file contains the output of all the Pid information that was collected
during your connection session. The text file, which is viewable with Window�s
Notepad consists of individual pid descriptions, pid values and a timestamp of when
the event occurred, one line of information per pid. As with earlier versions of OBD
2007, this data can be imported into an Excel spreadsheet and manipulated within
Excel. With our new OBD 2007 Mapping module we now do that work for you and present
the data as a 3D map, plus a graphical representation of the individual pids,
graphed over the same period of time.

The mapping module window is made up of three components, a grid displaying the
actual pid values imported into the OBD 2007 database table, a mapping grid showing
the results of the data manipulation and three graphs of the individual pids that
were chosen to be mapped. The graphs all share the same X Axis (time) and therefore
each pid value can be compared at a particular point in time. The mapping grid shows
the average values of the pid that was selected to be mapped based on one other
userand selectable pid and engine RPM. The mapping grid therefore displays a 3D view
of the data over the complete range of engine operating speeds for that particular
OBD 2007 connection session.

This is a particularly useful tool for studying trends � a couple of examples
would be fuel maps of Short Term and Long Term fuel trims or Ignition timing maps.
The data in the cells can be coloured based on user definable settings to indicate
data, in and out of the ideal range.
Build 1.0.13.92

Earlier this month we released build 1.0.13.92. In this release we added support for
a very impressive scan tool OBDPro. OBDPro scan tool supports all OBD II protocols
like the ELM 327. It comes in two versions, serial and USB and is very attractively
priced. OBDPro is not based on the Elm 327 chip, but does emulate all the necessary
AT commands of the Elm 327 chip, and therefore works seamlessly with OBD 2007. For
more information about OBDPro please visit their website www.obdpros.com.

For full details about all these releases, please see our news items page at
www.glmsoftware.com.

Evaluation downloads

All our evaluation versions are fully functional after activating the license with
our server and operate for a period of 7 days.

GLM Software support

We endeavour to answer your technical support questions as quickly as possible.
Please always include as many details as possible, including make, model and year of
the vehicle and protocol if known, when writing to support@glmsoftware.com and do
not forget to include the log files that are always produced by an OBD 2007
connection session. These log files are invaluable to us in diagnosing a problem
that you may be experiencing. If your OBD 2007 folder also contains an
OBD2007Errors.txt, please make sure you send that as well.

Remember if you do not let us know something is wrong, we cannot fix it.

Regards
Graham McKechnie
GLM Software
www.glmsoftware.com
 
#14 ·
So long as you can match a LTFT to a specific cell in the EMU inj map, you're golden -- that's all you need to dial out excessive fuel trims.

I don't see the point trying to tune out STFT's -- they change way too quickly, and each LTFT (a more stable value) will be based on a set of corresponding STFT's anyway.
 
#15 ·
This is essentially a streamlining of the process I use in excel. The beauty is the averaging function - you go for a drive, or just leave it logging while you do normal driving, come home, and the averaging gives you a number for each cell that you can add/subtract to your emange map. Dials you in beautifully quickly :) Also, you only need your ODBII to be logging, which simplifies the process and makes you more likely to do it more often.

My only complication is that I tune against MAP rather than MAF, but that's my problem :) Just needs one last translation for to let me make larger corrections on the MAP additional injection map, and smaller ones go directly into the airflow adjustment map (which is always vs MAF on the blue any way). But this isn't a problem most people will have.

Avendit
 
#16 ·
Can anyone give a detailed explanation of how this faster method is accomplished? Also, how can this be combined with the target AFR feature on the EMU? I just ordered an EMU, plug n play harness, and AEM UEGO, so I'd like to get the LTFT's dialed out as well as use the target AFR feature to keep a tune through the different seasons and temps.
 
#19 ·
picture, thousand words etc:



30 minutes logging, seconds to generate. Take the displayed numbers and add them to the appropriate cell in the emange, either on airflow adjustment vs throttle position or do some rough guess work/log other stuff to do it on another map.

Avendit
 
#21 ·
Cool. That screen shot helps a lot. Ok here are a few more questions...

Zero out all the low load (closed loop) cells in the inj map, if you haven't already done so. Again, make sure the airflow adjustment map low load cells are also zeroed out as well.
How do we know which exact cells these are?

If the logged LTFT on the OBD scantool that matches the selected datapoint on the EMU datalog graph/inj adj map was negative, subtract some fuel from that cell. If positive, add some fuel to that cell.
Can the target AFR feature be used here rather than adjusting the inj adj map? If so, how? Wouldn't using a target AFR in closed loop eliminate the need to re-tune out LTFT's every couple months as intake temps changed?

Also, I was reading the EMU operating manual and saw what they call the "Auxiliary Output Setting." Supposedly this map is used to "Control VTEC or for O2 Sensor Adapter (to control the O2 sensor feedback feature to force the system to go in to open loop)." I know the EMU can't change our lift engagement point without the V-manage, but can this feature still put our ECU into open loop? If so, couldn't we just use target AFR's all day long?
 
#22 ·
Gravel if its any help, I have this from when I was doing it all manually:



Yellower bits are 'mostly' closed loop on my turbo'd 140. The closer a number is to 2.0 the more likely that cell is to be in closed loop. The problem is that cells can be either depending on other factors.

Perhaps we should move this to another thread rather than cluttering up the stickie?
 
#23 ·
and Automaton, some work has been done in your suggested direction on the MX5 forums with good success I believe, alhtough I think that was targeted towards getting the blue to have full control the whole time rather than target mapping. Same principals should work tho.
 
#24 ·
D'oh. Re-read the operating manual. Target A/F Ratio should not be used in closed loop.

So, now that I think about it... when using Target A/F Ratio, can't we just stay out of the stock ECU's way in closed loop and ignore whatever LTFT's it makes? If the Target A/F Ratio feature works well, it should have no problem hitting target A/F ratios in open loop where the stock ECU is running off of a fixed map. Right?
 
#25 ·
Pretty much anything below 3.2 volts and up to about 3200 RPM is (usually) closed loop. You can see it switch on the scantool.

I wouldn't use the target map in part throttle -- it will screw with too many driveabiity issues (e.g., accel enrichment). Also, it's not as sophisticated as the ECU's closed loop -- it's all a single pre-set STFT.

You can't always be in open loop -- the ECU will give a CEL.

I like the little table you have there, but how do you match that back up to a load cell on the EMU? Is the TPS scaled to match EMU scaling and is your map scaled that way? This is why I visually find them on the RPM curve.

Also note that LTFT's in open loop are based on a calculated value dervied from the LTFT's in closed loop.

One other point is that you have to give it a little time for the LTFT's to stabliize, which will be relfected in STFT's for the corresponding cell that switch back and forth rapidly between very small positve and small negative values -- looks like /\/\/\/\/\/\/\/\/\/\ when graphed; the LTFT, once stablilized, will be a pretty much flat line cutting through the middle of it.
 
#34 ·
Pretty much anything below 3.2 volts and up to about 3200 RPM is (usually) closed loop. You can see it switch on the scantool.
I thought I was getting there on my LTFTs, but the ECU is fighting me on some important cells. I think I am actually in closed loop in boost! At low throttle, I can creep into boost without the ECU noticing - at about 30% throttle, I am at 0.2kg/cm2 of boost (about 1/3 way) at a mere 2500 rpm and 2.8v MAF output, and I am pinging as my AFR's >15!

So, at this point do I actually start lying to the ECU about the airflow? :shrugs:
 
#26 ·
Just found a bitching new tool for logging OBD data. Its called the Mongoose interface from Drew Technologies. I got one for work and it allows you to use the actual toyota scantool software on a normal laptop. So far it seems full function just like the new techstream scantool, and it also worked on a Lotus this weekend using the generic OBD2

The cool thing is that the Toyota software lets you graph and save fuel trim, rpm etc. which means you can tune solo much more easily. It also means you can read the full datastream, run all the active tests, clear codes, save data etc. Its pricey ($500) but well worth it considering that the factory scan tool that it duplicates the function of costs $8500.
 
#27 · (Edited)
That's really sweet -- I'll have to check it out!

EDIT: Whoa -- cool, no bulky OBDII connector -- it's all contained in that little cable! Damn, I want one!

EDIT II: Looks like it uses the PMSscan software -- I wonder if that will work with my current connector... I might just give it a try. They have freebe versions...

Also, Avendit, either I missed something or you just updated the pic -- that software output looks great.

But keep in mind that 15% TPS on Toyota ECU is probably less than 5% on the EMU log (10% is usually 0%)... you'll still need to take a look at either the RPM curve or something else scaled the exact same way to correctly identify the corresponding load cell in the EMU maps...
 
#28 ·
Yea, I'm aware of the slight miss-alignment in the OBD data. I just do rough mental calculations to get it right, and as you can see, most of the tune is just a big block. I've decided to wait for it to get warmer before touching any of it, unless any bad spots develop over the next while.

Avendit
PS check out http://www.obdkey.com/ if you want small/easy connection. works with every bit of software I have tired so far.
 
#30 ·
2. Zero out the airflow map, at least through 2K in low load. Worthless. All it's doing is throwing things off. I would only mess with this if you are having idle problems due to your intake. Best to leave this zeroed though, IMO.
When running an aftermarket intake, the larger diameter pipe causes the MAF to report less air to the stock ECU than what is actually going in. Any reason why we wouldn't want to bump up the airflow map a little (even in low load) to compensate?
 
#31 ·
So long as the idle is stable, the rest can be tweaked by adjusting inj correction directly in the inj map (or inj accel map for tip-in adjustments) on the basis of fuel trims in closed loop and wideband AFR in open loop (except for decel, though you might need to add fuel for a softer landing to idle if you need to prevent a stalling condition -- unlikely with a wider diameter intake, tho').

If you can't get a stable idle otherwise, I think it's fine to try and make adjustments with the airflow map, but it globally commands the ECU to read all associated look-up tables in the wrong place, meaning it will potentially create lots of little drivability quirks.

Furthermore, you'll then need to make additional corrections to compensate for changes in fueling associated with the rescaled MAF voltage (i.e., fuel trims) with the inj map anyway (and possibly timing in the ign map, which the ECU won't like).

That's why I feel the airflow map should only be used to make corrections when maxing out the MAF sensor -- it's not a precise tuning adjustment; more like a hatchet than a scalpel.
 
#32 ·
Here are some screenshots from the Toyota software and Mongoose cable in graphing mode. You can also just do a datalist, or use a gauge or bargraph mode or whatever. :



 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top