SD - first test success
Feb 6, 2009, 06:23 PM
#79
I have already been very careful to avoid shared maps, which is why I made a new map for the MAP calibration. The only shared thing is the RPM axis for the RPM VE table - if you change this you also change the RPM axes on many other tables. However, there should be no need to change it since RPM is a minor effect and we have 500-8000 RPM.
excellent news though.
Feb 7, 2009, 01:16 AM
#81
Evolved Member
Thread Starter
I've killed two OBD II baro codes and one OBD II IAT code. So 3 down, 1 to go (P0100).
It might be useful if a US Evo 8 and US Evo 9 unplug their MAF and start their engine and post up the OBD II codes you get. Don't drive the car much on these settings because it will run rich, zero octane and open loop.
It might be useful if a US Evo 8 and US Evo 9 unplug their MAF and start their engine and post up the OBD II codes you get. Don't drive the car much on these settings because it will run rich, zero octane and open loop.
Feb 7, 2009, 05:35 AM
#83
I've killed two OBD II baro codes and one OBD II IAT code. So 3 down, 1 to go (P0100).
It might be useful if a US Evo 8 and US Evo 9 unplug their MAF and start their engine and post up the OBD II codes you get. Don't drive the car much on these settings because it will run rich, zero octane and open loop.
It might be useful if a US Evo 8 and US Evo 9 unplug their MAF and start their engine and post up the OBD II codes you get. Don't drive the car much on these settings because it will run rich, zero octane and open loop.
Feb 7, 2009, 10:32 AM
#84
Evolved Member
Thread Starter
All OBD codes now removed (P0100 MAF, baro is P22xx can't remember the xx, IAT can't remember the code), no CEL. Drives normally. My MAF and boost solenoid harness are tucked up near the intercooler spray reservoir, but to drive or log you wouldn't know (except baro shows 0 in logs and IAT is the same as MAT).
I did some testing of STFT at idle, 1250, 1500, 2000, 3000 RPM.
From this I concluded that idle at 40kPa it was correct, but free revving at 1500-3000 RPM had a positive trim at 30kPa. So I dropped the RPM VE by 10% at 500-1000 RPM, and increased the MAP VE by 10% at 40kPa to compensate. I increased the VE by 10% at 30kPa, and now free revving has negligible trims.
Electrical loads, return to idle, everything is like stock. I am very pleased.
Whilst I have been "naughty" and tested full load without wideband, I was watching narrowband and it was the usual 0.88 or 0.90 V, IDC and timing were usual, and it had my usual 0-3 knock sums. I will of course confirm the tune with my wideband (I shoot for WOT 11.5:1 AFR 93 PON stock turbo) before doing further full boost stuff, but I think it will need at most fine tuning, if at all. I have not changed my fuel and timing maps at all, and have only reflashed the VE once from my initial guess. It seems easy and forgiving to tune, just as we want it.
I don't think the OBD system is wildly different in the US cars, it is OBD routines I've tweaked to fix the baro and MAF errors. Hopefully the transfer will be easy.
I suppose next I should put together info on the changes and where they are in the code so others can help to find them... but I'm out this evening so it won't be tonight.
If in the meantime anyone has 88570008 and wants to look at my xml entries (note change endian if after 1.29a, I still use an old version).
Here is my ROM (SD, Tephra, DMA/live mapping):
http://banks.myzen.co.uk/88570008-v5-modsJB2SD2.zip
I did some testing of STFT at idle, 1250, 1500, 2000, 3000 RPM.
From this I concluded that idle at 40kPa it was correct, but free revving at 1500-3000 RPM had a positive trim at 30kPa. So I dropped the RPM VE by 10% at 500-1000 RPM, and increased the MAP VE by 10% at 40kPa to compensate. I increased the VE by 10% at 30kPa, and now free revving has negligible trims.
Electrical loads, return to idle, everything is like stock. I am very pleased.
Whilst I have been "naughty" and tested full load without wideband, I was watching narrowband and it was the usual 0.88 or 0.90 V, IDC and timing were usual, and it had my usual 0-3 knock sums. I will of course confirm the tune with my wideband (I shoot for WOT 11.5:1 AFR 93 PON stock turbo) before doing further full boost stuff, but I think it will need at most fine tuning, if at all. I have not changed my fuel and timing maps at all, and have only reflashed the VE once from my initial guess. It seems easy and forgiving to tune, just as we want it.
I don't think the OBD system is wildly different in the US cars, it is OBD routines I've tweaked to fix the baro and MAF errors. Hopefully the transfer will be easy.
I suppose next I should put together info on the changes and where they are in the code so others can help to find them... but I'm out this evening so it won't be tonight.
If in the meantime anyone has 88570008 and wants to look at my xml entries (note change endian if after 1.29a, I still use an old version).
Code:
<scaling name="MAP 16bit" units="kPa" toexpr="x*1.334" frexpr="x/1.334" format="%.0f" min="0" max="400" inc="1" storagetype="uint16" endian="little"/> <scaling name="MAP 8 bit" units="kPa" toexpr="x*1.334" frexpr="x/1.334" format="%.0f" min="0" max="400" inc="1" storagetype="uint8" endian="big"/> <scaling name="Percent (128)" units="%" toexpr="x/1.28" frexpr="x*1.28" format="%.0f" min="0" max="100" inc="1" storagetype="uint8" endian="little"/> <table name="SD NOP 1" category="Misc" address="12100" type="1D" level="1" scaling="Hex16"/> <table name="SD change to preptable" category="Misc" address="120d4" type="1D" level="1" scaling="Hex16"/> <table name="SD change to mov #0x73,r2" category="Misc" address="120da" type="1D" level="1" scaling="Hex16"/> <table name="SD NOP 2" category="Misc" address="263ec" type="1D" level="1" scaling="Hex16"/> <table name="SD NOP 3" category="Misc" address="265e2" type="1D" level="1" scaling="Hex16"/> <table name="SD MAP sensor VE and calibration" category="Misc" address="2a94" type="2D" level="1" scaling="MAP 8 bit"> <table name="X" address="29ee" type="X Axis" elements="8" scaling="MAP 16bit"/> </table> <table name="SD RPM VE" address="2a7c" type="2D" level="1" scaling="Percent (128)"> <table name="RPM" address="5cee" type="X Axis" elements="17" scaling="RPM"/> </table> <table name="SD Map calibration header 1" category="Misc" address="29e4" type="2D" level="1" scaling="Hex16"> <table name="X" type="X Axis" elements="5"/> </table> <table name="SD Map calibration header 2" category="Misc" address="2a92" type="1D" level="1" scaling="Hex16"/> <table name="SD Map calibration pointer" category="Misc" address="1225e" type="1D" level="1" scaling="Hex16"/> <table name="SD disable MAF error main" category="Misc" address="11f5e" type="1D" level="1" scaling="Hex16"/> <table name="SD MAT to IAT change to 0E" category="Misc" address="b215" type="1D" level="1" scaling="Hex8"/> <table name="SD disable baro error OBDII CEL" category="Misc" address="20ca6" type="1D" level="1" scaling="Hex16"/> <table name="SD disable MAF error OBDII CEL" category="Misc" address="208a4" type="1D" level="1" scaling="Hex16"/>
http://banks.myzen.co.uk/88570008-v5-modsJB2SD2.zip
Last edited by jcsbanks; Feb 7, 2009 at 10:37 AM.
Feb 7, 2009, 05:04 PM
#88
Evolved Member
Thread Starter
SD NOP1 stops the code trimming the calculated MAF variable which would limit the maximum load.
SD NOP2 and 3 remove conditional branches - the effect is that our calculated MAF variables replace the ones that would come from the MAF sensor. In between NOP2 and NOP3 there is some code that is only present in JDM 9 engines with MAP sensors - like mine (the code is used to limit the MAF reading on lift off from high loads). So the gap in bytes between NOP2 and NOP3 will be much less in your ROM.
If you want to send me your best disasm I can help you find the right spots or confirm yours. Nothing like seeing something on your "home" ROM to get a handle on it.
SD NOP2 and 3 remove conditional branches - the effect is that our calculated MAF variables replace the ones that would come from the MAF sensor. In between NOP2 and NOP3 there is some code that is only present in JDM 9 engines with MAP sensors - like mine (the code is used to limit the MAF reading on lift off from high loads). So the gap in bytes between NOP2 and NOP3 will be much less in your ROM.
If you want to send me your best disasm I can help you find the right spots or confirm yours. Nothing like seeing something on your "home" ROM to get a handle on it.
Feb 7, 2009, 05:05 PM
#89
Is it realistic to expect this to be distributed as a patch in a matter of weeks? I'm not a tech guru by any means but it seems like Banks has this 98% completed. I'm actually in the process of getting dangerously close to the breaking point of the honeycomb on the mas and this most certainly beats my only other option, Maf T Pro.
Last edited by RyuEvoIX; Feb 7, 2009 at 05:07 PM.