OEM fuel mapping, AFR/Knock learning [Archive] - SaabCentral Forums

: OEM fuel mapping, AFR/Knock learning


lowoffset
7th July 2014, 07:36 PM
I'm curious if the AFR's and fuel trimming is a common "problem" on our ECU/OEM fuel mapping. Trying to see if anyone else has experienced this. Usually modern ECU's (SAAB being a forefront runner) have AFR Learning (AFR Learning A,B,C,D), knock even learning, etc as the engine learns of knocking during certain RPM's , loads, and so on.

What I'm finding is that when I'm at WOT in 4th, overtaking someone, the car feels like the fuel mapping has flat spots. I know turbos don't have a linear "feel", and that is not what I'm inquiring about. Is this the car's ecu trimming down timing/fuel mapping as a result of a previous learned knock event? That is how I'm seeing it, but not sure. I've owned plenty of Turbo cars and what I'm experiencing is the "tune" not having correct fuel/timing (I've also done a wee bit of open source tuning on my subies), leading me to believe this could possibly be learned knock events, thus, like said earlier, reduced timing/fuel mapping at certain loads/RPM's.

Also wondering if anyone else has had flat spot experiences during the higher rpms.

I haven't fully reset the ecu yet. I will be doing a stage 1 tune later in the year, but still learning how the ECU works.

Saabohème
7th July 2014, 09:27 PM
Start with Analyzing Trionic 8 with T8Suite (http://socalsaab.com/suites/Trionic8.pdf), rev 1.03, © Dilemma 2011. It contains a lot of useful explanatory information about how T8 works.
If you want to dig deeper about ion current sensing there are lots and lots of tech documents available on the web. A brief sampling:
An Analytic Model for Cylinder Pressure in a Four Stroke SI Engine (http://tikkari.net/Public/cylinder_pressure/EriAnd_2002_SAE.pdf)
An Ion-Sense Engine Fine-Tuner (http://www.ieeecss.org/CSM/library/1998/oct1998/02-EngineFineTuner.pdf)
Further Development of an Ion current Combustion Control System (ICCS) (http://www.diaelec.co.jp/eng/randd/pdf/t0103.pdf)
Fuzzy Control of Spark Advance by Ion Current Sensing (http://www.sid.ir/en/VEWSSID/J_pdf/102720050202.pdf)
Real-Time Digital Signal Processing of Ionization Current for Engine Diagnostic & Control (http://papers.sae.org/2003-01-1119/) (I can only find the abstract now, but I have the full paper in pdf if you want it)
Spark Advance Modeling & Control (http://www.vehicular.isy.liu.se/en/Publications/PhD/99_PhD_580_LE.pdf)
A Chemical Kinetics Model of Current Signatures in an Ionization Sensor (http://web.cse.ohio-state.edu/~saday/894/spark9.doc) NB: This link will automatically download a Word doc to your computer if you click it.
Spark Advance Control by Ion Sensing (SparkAdvance Control by IonSensing)
Ion-Sense-Based Real-Time Combustion Sensing for Closed Loop Engine Control (http://delphi.com/pdf/techpapers/2013-01-0354.pdf)
Delphi Ionization Current Sensing Ignition Subsystem (http://delphi.com/shared/pdf/ppd/pwrtrn/ionization-current-sensing-ignition-susbystem.pdf)

lowoffset
7th July 2014, 09:32 PM
Start with Analyzing Trionic 8 with T8Suite (http://socalsaab.com/suites/Trionic8.pdf), rev 1.03, © Dilemma 2011. It contains a lot of useful explanatory information about how T8 works.
If you want to dig deeper about ion current sensing there are lots and lots of tech documents available on the web. A brief sampling:
An Analytic Model for Cylinder Pressure in a Four Stroke SI Engine (http://tikkari.net/Public/cylinder_pressure/EriAnd_2002_SAE.pdf)
An Ion-Sense Engine Fine-Tuner (http://www.ieeecss.org/CSM/library/1998/oct1998/02-EngineFineTuner.pdf)
Further Development of an Ion current Combustion Control System (ICCS) (http://www.diaelec.co.jp/eng/randd/pdf/t0103.pdf)
Fuzzy Control of Spark Advance by Ion Current Sensing (http://www.sid.ir/en/VEWSSID/J_pdf/102720050202.pdf)
Real-Time Digital Signal Processing of Ionization Current for Engine Diagnostic & Control (http://papers.sae.org/2003-01-1119/) (I can only find the abstract now, but I have the full paper in pdf if you want it)
Spark Advance Modeling & Control (http://www.vehicular.isy.liu.se/en/Publications/PhD/99_PhD_580_LE.pdf)
A Chemical Kinetics Model of Current Signatures in an Ionization Sensor (http://web.cse.ohio-state.edu/~saday/894/spark9.doc)
Spark Advance Control by Ion Sensing (SparkAdvance Control by IonSensing)
Ion-Sense-Based Real-Time Combustion Sensing for Closed Loop Engine Control (http://delphi.com/pdf/techpapers/2013-01-0354.pdf)
Delphi Ionization Current Sensing Ignition Subsystem (http://delphi.com/shared/pdf/ppd/pwrtrn/ionization-current-sensing-ignition-susbystem.pdf)

Thanks! I will definitely read those over this week.

However, do you know off hand of the "flat spots" I'm having?

Saabohème
7th July 2014, 10:02 PM
Could be too many things to figure out this way.
Could simply be the unholy delay built into the stock throttle by wire pedal sensor. Could be plugs and/or coils (and you should always check the gaps and general overall conditions of your plugs at the first sign of anything being different or wrong about how your car is running). Could even be gas (not flatulence) but if you got some cruddy gas, it could be that, as these cars are very sensitive to the fuel used. The Ohio State paper regarding Chemical Kinetics gives some compelling reasons why this might be so, inasmuch as that paper investigates the numerous processes that produce the ions and determines that A/F ratio plays a critical role in the quantities of the various types of ions that will be produced during the combustion and pressure cycles. Since Trionic retards timing first and then starts adding additional fuel in the presence of knock, the A/F ratios will get progressively richer when using insufficient octane fuel or contaminated fuel.

Diggs
7th July 2014, 10:06 PM
I'd highly recommend you go and get your car tuned by John while he is in the area.

https://www.saabcentral.com/forums/showthread.php?t=402538

lowoffset
7th July 2014, 10:08 PM
Could be too many things to figure out this way.
Could simply be the unholy delay built into the stock throttle by wire pedal sensor. Could be plugs and/or coils (and you should always check the gaps and general overall conditions of your plugs at the first sign of anything being different or wrong about how your car is running). Could even be gas (not flatulence) but if you got some cruddy gas, it could be that, as these cars are very sensitive to the fuel used. The Ohio State paper regarding Chemical Kinetics gives some compelling reasons why this might be so, inasmuch as that paper investigates the numerous processes that produce the ions and determines that A/F ratio plays a critical role in the quantities of the various types of ions that will be produced during the combustion and pressure cycles. Since Trionic retards timing first and then starts adding additional fuel in the presence of knock, the A/F ratios will get progressively richer when using insufficient octane fuel or contaminated fuel.

Good info.

When I bought the car, the PO replaced all 4 spark plugs, so it can't be that. I'm in California so we have piss water 91, which I have always been using in the car.

When I get close to empty I will add a few gallons of 100 octane and see how WOT feels. I will also reset the ECU before my next outing.

Saabohème
7th July 2014, 10:10 PM
Before you do ANYTHING else, pull the plugs. Make sure they are the correct plug specced for your car and that the gap does not exceed 1 mm.
2005 9-3SS owner's manuals are here (http://justgivemethedamnmanual.com/saab/saab-9-3-owners-manuals/).

lowoffset
7th July 2014, 10:24 PM
Before you do ANYTHING else, pull the plugs. Make sure they are the correct plug specced for your car and that the gap does not exceed 1 mm.
2005 9-3SS owner's manuals are here (http://justgivemethedamnmanual.com/saab/saab-9-3-owners-manuals/).

Will do. The PO brought it to a SAAB shop in san diego, but I will double check their work.

I talked with John at JZW and may bring the car down to him to get a stage 1 :)

He mentioned it may be the Ignition timing pulling and adding, so my speculations were correct. Just didn't know how our ECU's worked.

Thanks for all the responses!

Saabohème
7th July 2014, 10:39 PM
Yes and no. Regardless of adaptation (learning), (i.e. whether you have been driving it like you stole it or whether you have been the little old lady from Pasadena puttering off to church), WHENEVER Trionic detects knock, the FIRST thing it does is start retarding timing. At a certain point, it will also begin to add additional fuel if the knocking persists. If timing has been retarded (to a total of 15 degrees IIRC) and additional fueling has failed to cure the knocking, Trionic begins dialing back boost. Again, these protective measures are employed regardless of how you have adapted the ECU to your driving style. I think the main parameter affected by adaptation is the torque request per pedal sensor input, which will also necessarily affect air mass. It usually only takes 1 to 2 WOT runs in 3rd through 4th gear to learn that thang a new lesson.

lowoffset
7th July 2014, 10:44 PM
Yes and no. Regardless of adaptation (learning), (i.e. whether you have been driving it like you stole it or whether you have been the little old lady from Pasadena puttering off to church), WHENEVER Trionic detects knock, the FIRST thing it does is start retarding timing. At a certain point, it will also begin to add additional fuel if the knocking persists. If timing has been retarded (to a total of 15 degrees IIRC) and additional fueling has failed to cure the knocking, Trionic begins dialing back boost. Again, these protective measures are employed regardless of how you have adapted the ECU to your driving style. I think the main parameter affected by adaptation is the torque request per pedal sensor input, which will also necessarily affect air mass. It usually only takes 1 to 2 WOT runs in 3rd through 4th gear to learn that thang a new lesson.

This is what I speculated, but not as in depth in my posts. You nailed it spot on.

Subaru's ECU's are very similar, except for the fact if it see's increase in power/airflow/etc. (like a down pipe), it doesn't compensate to try and bring the power back into the stock power "threshold". Also Subaru ECU's do not compensate for lower octane gas (like as in a low octane map the T8 has...just read about that in the link provided) which is pretty damn awesome.

Diggs
7th July 2014, 10:52 PM
These things are quite complex, almost to a fault.

Complex car + limited market = very few tuners

Saabohème
7th July 2014, 11:01 PM
This is what I speculated, but not as in depth in my posts. You nailed it spot on.

Subaru's ECU's are very similar, except for the fact if it see's increase in power/airflow/etc. (like a down pipe), it doesn't compensate to try and bring the power back into the stock power "threshold". Also Subaru ECU's do not compensate for lower octane gas (like as in a low octane map the T8 has...just read about that in the link provided) which is pretty damn awesome.OK. I guess I was confused by this statement:
Is this the car's ecu trimming down timing/fuel mapping as a result of a previous learned knock event?I interpreted that to mean you were speculating that how the ECU was adapted affected how Trionic employed its protective defaults, and so I was emphasizing that Trionic will respond to all knocking following the standard algorithms, without regard to adaptation. I guess it can be stated this way: Trionic will NOT anticipate knock based upon previous events, but it will aggressively follow its preprogrammed protocols to eliminate knock when it is detected. In addition to that, when knock has been quelled, Trionic will begin dialing back in the more aggressive parameters that existed prior to the knock to maximize performance. That is why some knock count is tolerated by the programming - to let the ECU try to walk the tightrope.

Deli-Llama
8th July 2014, 09:55 AM
I would suggest buying a Bluetooth OBD adapter and running an app on your smartphone to log some data. I was amazed to find the stock tune would continually adjust the throttle position on WOT. Same fuel with VTuner stage 0 tune just kept the throttle wide open.

lowoffset
10th July 2014, 12:07 PM
I would suggest buying a Bluetooth OBD adapter and running an app on your smartphone to log some data. I was amazed to find the stock tune would continually adjust the throttle position on WOT. Same fuel with VTuner stage 0 tune just kept the throttle wide open.


Maybe this is what I'm feeling. It does feel like the tb is slightly closing/opening. I will check out a bt adapter

Ric-Raf
12th July 2014, 03:12 PM
I have experienced the same "flat spots" that you are talking about and was curious myself as to whether it is the adaptive tuning or not. I bought my '08 9-3 a couple months ago and it has recently seemed flat compared to what it did initially. I do drive it pretty mellow for run around trips so I had figured that it was adapting to that style of driving rather than WOT pulls throught the gears. Always use 93, but it doesnt seem as strong on the low end and feels like the efficiency of the turbo is getting surpassed at the higher RPMs. I havent looked at knock history which I need to do. What bluetooth/apps have you guys had good experiences with for logging?

Deli-Llama
12th July 2014, 09:41 PM
I use the OBDLink LX adapter which includes Windows and Android software. I also purchased Torque Pro for Android which has more advanced configuration for dashboards but they all display the same information.

lowoffset
12th July 2014, 11:20 PM
Is AFR, AFR learning, and knock sensors on the torque app?

edusaab
13th July 2014, 05:52 AM
About "flat spot", do you refer that you hit the pedal, the car goes full boost, but for some reason after going full boost, it seems like the TCS acts, cuts the boost for 1 seconds more or less, like it detects wheelspin, and then goes full boost again.

This happen to me sometimes, in third gear, since i tuned the car. I've been said by Maptun that this happens to some highly tuned Saabs. If you disconect the TCS then it's gone, it never happens.

I tried to trick the TCS sensors, adding the module with bigger brakes, as I have the an AP racing brake kit, 330mm, it's seems it's been reduced but it still happens sometimes. as I have LSD installed i simply disconnect the TCS always.

Regards

Deli-Llama
13th July 2014, 09:12 AM
Is AFR, AFR learning, and knock sensors on the torque app?

Unfortunately the answer is not really, yes (if you mean short fuel trim and long fuel trim) and no. The OBD port seems to only supply "standard" or SAE PIDs for the 2.0T.

That being said, you can get the realtime voltage of both O2 sensors but they aren't wideband. You can get a snapshot of the misfire count of each cylinder but not knock.

The realtime value for #1 cylinder ignition timing is possible as is throttle position, MAF, MAP, catalyst temp, coolant temp and inlet temp.

Of you are interested, I can dump the entire list of supported PIDs.

Deli-Llama
13th July 2014, 06:53 PM
Diagnostic Report

Created by OBDwiz - OCTech, LLC www.obdsoftware.net

Date: 7/6/2014 5:55:36 PM

VIN: XXXXXXXXXXXXXXXXX
Manufacturer: Saab
Model: 9-3
Year: 2008

Monitor Status Report

Name Continuous Available Complete
Misfire Yes Yes Yes
Fuel System Yes Yes Yes
Components Yes Yes Yes
Catalyst No Yes Yes
Heated Catalyst No No No
Evap System No No No
Secondary Air System No No No
AC Refrigerant No No No
Oxygen Sensor No Yes Yes
Oxygen Sensor Heater No Yes Yes
EGR System No No No

MIL Off
Number of Stored Codes: 0
US State: CT

This vehicle is ready for emissions testing.

Trouble Code Report

There are no pending, stored, or permanent diagnostic trouble codes (DTCs).

Service $01 - Powertrain Diagnostic Data

PID Description Value Units
SAE 0x03 Fuel system status Unknown
SAE 0x04 Calculated load value 0 %
SAE 0x05 Engine coolant temperature 75 C
SAE 0x06 Short term fuel % trim - Bank 1 0 %
SAE 0x07 Long term fuel % trim - Bank 1 -3.91 %
SAE 0x0B Intake manifold absolute pressure 102 kPa
SAE 0x0C Engine RPM 0 RPM
SAE 0x0D Vehicle speed 0 km/h
SAE 0x0E Ignition timing advance for #1 cylinder 0
SAE 0x0F Intake air temperature 51 C
SAE 0x10 Mass air flow rate 0 g/s
SAE 0x11 Absolute throttle position 34.51 %
SAE 0x13 Oxygen sensors present Bank 1: Sensor 1, Sensor 2, Bank 2: None
SAE 0x14 O2 voltage (Bank 1, Sensor 1) 0.28 V
SAE 0x14 Short term fuel trim (Bank 1, Sensor 1) -1 %
SAE 0x15 O2 voltage (Bank 1, Sensor 2) 0.26 V
SAE 0x15 Short term fuel trim (Bank 1, Sensor 2) -1 %
SAE 0x1C OBD requirements to which vehicle or engine is certified EOBD
SAE 0x1F Time since engine start 0 sec
SAE 0x21 Distance traveled while MIL is activated 0 km
SAE 0x2E Commanded evaporative purge 0 %
SAE 0x2F Fuel level input 56.86 %
SAE 0x30 Number of warm-ups since DTCs cleared 26
SAE 0x31 Distance traveled since DTCs cleared 239 km
SAE 0x33 Barometric pressure 100 kPa
SAE 0x3C Catalyst temperature (Bank 1 Sensor 1) 75 C
SAE 0x42 Control module voltage 12.3 V
SAE 0x43 Absolute load value 0 %
SAE 0x44 Fuel/Air commanded equivalence ratio 2
SAE 0x45 Relative throttle position 32.94 %
SAE 0x47 Absolute throttle position B 33.73 %
SAE 0x49 Accelerator pedal position D 19.22 %
SAE 0x4A Accelerator pedal position E 19.22 %
SAE 0x4C Commanded throttle actuator control 19.22 %
Custom 0x00 Input voltage read by the scan tool 13.4 V

Service $02 - Freeze Frame Report

Freeze Frame data is not available

Service $05 - Oxygen Sensors

Sensor Available
Bank 1 - Sensor 1 Yes
Bank 1 - Sensor 2 Yes
Bank 1 - Sensor 3 No
Bank 1 - Sensor 4 No
Bank 2 - Sensor 1 No
Bank 2 - Sensor 2 No
Bank 2 - Sensor 3 No
Bank 2 - Sensor 4 No

Service $06 - On-Board Monitoring

Component Description Value Minimum Maximum Units Result
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $01 - Rich to lean sensor threshold voltage (constant) 0.6 0.6 0.6 V Pass
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $02 - Lean to rich sensor threshold voltage (constant) 0.3 0.3 0.3 V Pass
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $07 - Minimum sensor voltage for test cycle (calculated) 0.1 0.1 65.535 V Pass
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $08 - Maximum sensor voltage for test cycle (calculated) 0.889 0 1.199 V Pass
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $82 - Manufacturer Defined -1 -25 25 Pass
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $80 - Manufacturer Defined 98 0 220 counts Pass
$01 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 1 TID $81 - Manufacturer Defined 4 4 10 counts Pass
$02 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 2 TID $07 - Minimum sensor voltage for test cycle (calculated) 0.142 0.1 65.535 V Pass
$02 - Exhaust Gas Sensor Monitor Bank 1 – Sensor 2 TID $08 - Maximum sensor voltage for test cycle (calculated) 0.734 0 1.2 V Pass
$21 - Catalyst Monitor Bank 1 TID $80 - Manufacturer Defined 1.575 0.901 10 sec Pass
$81 - Fuel System Monitor Bank 1 TID $81 - Manufacturer Defined -0.82 -25 25 % Pass
$81 - Fuel System Monitor Bank 1 TID $82 - Manufacturer Defined -0.02 -25 25 % Pass
$A2 - Misfire Cylinder 1 Data TID $0B - EWMA (Exponential Weighted Moving Average) misfire counts for last ten (10) driving cycles 0 0 65535 counts Pass
$A2 - Misfire Cylinder 1 Data TID $0C - Misfire counts for last/current driving cycles (calculated, rounded to an integer value) 0 0 65535 counts Pass
$A3 - Misfire Cylinder 2 Data TID $0B - EWMA (Exponential Weighted Moving Average) misfire counts for last ten (10) driving cycles 0 0 65535 counts Pass
$A3 - Misfire Cylinder 2 Data TID $0C - Misfire counts for last/current driving cycles (calculated, rounded to an integer value) 0 0 65535 counts Pass
$A4 - Misfire Cylinder 3 Data TID $0B - EWMA (Exponential Weighted Moving Average) misfire counts for last ten (10) driving cycles 0 0 65535 counts Pass
$A4 - Misfire Cylinder 3 Data TID $0C - Misfire counts for last/current driving cycles (calculated, rounded to an integer value) 1 0 65535 counts Pass
$A5 - Misfire Cylinder 4 Data TID $0B - EWMA (Exponential Weighted Moving Average) misfire counts for last ten (10) driving cycles 0 0 65535 counts Pass
$A5 - Misfire Cylinder 4 Data TID $0C - Misfire counts for last/current driving cycles (calculated, rounded to an integer value) 0 0 65535 counts Pass

Service $09 - Vehicle Information

General Information
Description Value
Vehicle Identification Number XXXXXXXXXXXXXXXXX
Calibration ID - $7E8 XXXXXXXX
Calibration Verification Number - $7E8 XXXXXXXX

Deli-Llama
13th July 2014, 06:54 PM
Diagnostic Report

Created by OBDwiz - OCTech, LLC www.obdsoftware.net (http://www.obdsoftware.net)

Date: 7/6/2014 5:55:36 PM


That was from the PC based software OBDwiz but you'll get the same information from Torque Pro, just displayed differently.

lowoffset
30th November 2014, 03:15 PM
I would suggest buying a Bluetooth OBD adapter and running an app on your smartphone to log some data. I was amazed to find the stock tune would continually adjust the throttle position on WOT. Same fuel with VTuner stage 0 tune just kept the throttle wide open.


I just realized what you said here...4 month later.

Anyway, will a tune by JZW or BCB have this affect also? Preventing the TB from "fluctuating" in turn keeping it open in reference to pedal position? I hate this fluctuating TB...

nairua
12th December 2014, 03:37 PM
Unfortunately the answer is not really, yes (if you mean short fuel trim and long fuel trim) and no. The OBD port seems to only supply "standard" or SAE PIDs for the 2.0T.

That being said, you can get the realtime voltage of both O2 sensors but they aren't wideband. You can get a snapshot of the misfire count of each cylinder but not knock.

The realtime value for #1 cylinder ignition timing is possible as is throttle position, MAF, MAP, catalyst temp, coolant temp and inlet temp.

Of you are interested, I can dump the entire list of supported PIDs.

Just get a wideband afr sensor and EGT/boost if you go on the DYI tuning road.

I have PLX icm PCMScan. Can't get T8S working with realtime logging. (lawicel)