|
I also have an ADM 04 Liberty GT 5EAT with 110kms on the clock and I've been following this forum for a long time waiting for another one to pop up. I've been saving my first post for when I could actually add some value and I hope this does so. That way I can feel slightly less guilty about using all the good advice on the forum (Colby, Freon, Mickeyd and Merchgod are the standouts)!
The car isn't standard, with a Hyperflow TMIC, TD04HLA ($600AU from EBAY), gutted downpipe and at one time in the past an ECUTEK tune. All tuning on the current configuration was done by me (though as I said I owe a lot to others on this forum) and I will try and hit the highlights rather than the entire saga (which includes bench flashing the ECU after a stupid moment).
The key thing that I hope helps is that after being frustrated with not having access to "extended" logging parameters such as knock sum and other 2 byte engineering parameters, I went in search of the latest base map that supported them. I ended up with an A2WD012G ROM which is a later version than the "tuned by STI" ROM and I copied all the JDM tables over to it as my starting point. And luckily it worked no problems and has made logging far easier.
Next I borrowed a WBO2 for a while and was able to confirm that the standard AFR sensor is very, very close and after some minor re-scaling has an excellent correlation with the WBO2. This aligned with other information posted in this forum quite some time ago and has resulted in being able to use the standard sensor down to 11 to 1. However, scaling the AFR highlighted that the fuelling was way out, with 5%-10% difference between open loop fueling and actual AFR.
After some closed loop logging to scale the injectors (which if I recall resulted in very little change) I decided to rescale the high flow portion of the MAF. This took a very, very long time with lots of logging and tweaking of ROMs. It turns out that the standard MAF appears very optimistic (353max vs 318max), assuming my injector scaling correct, and results in much richer AFR than target. In conjunction with the stupidly rich standard fuel targets (9.8 I think) the resulting actual AFRs would have been in the 8s! I now have actual AFRs that track the map to within a couple of % but am still running essentially the same as the standard JDM.
As far as other changes to the map go, the AVCS I think is almost identical to what Mickeyd came up with on another JDM liberty (which worked exceptionally well), I have zeroed CL/OL delay, increased temperature compensations, set peak boost at 18.5psi and adjusted WDGC tables to suit. Ignition timing has of course been using Mickeyd's great spreadsheet.
Whilst the car is definitely stronger at the top end (far stronger than the ECUTEK tune), what I really like is how quickly it now spools and accelerates at low rpm and part throttle. Acceleration on the highway that would originally have had the gearbox unlock and downshift can now be had without it even unlocking. I often hit loads of 2.6g/s which would be even higher with standard MAF scaling, compared to less than 2.3g/s standard.
What's wrong with the tune?
I think the peak boost is a bit high: after 17.5 psi the increases in mass airflow were negligible and I'm not really sure what the bottom end of these engines is (rumoured to be forged psitons and semi-closed deck).
The WGDC is very agressive: I see 95% a lot and I can often hear the solenoid at slow speeds and part throttle.
I ocassionaly get very minor compressor surge: this happens at boost below wastegate spring pressure so is pretty hard to tune out.
I have the odd knock in mid-load region: it appears very ocassionally in Learning Views but I'll be damned if I can actually catch it happening.
Not withstanding these issues, I've been driving the car with this tune for nearly 12mths and I just can't get motivated to do the final few changes to address them.
I have attached my ROM for what its worth, any feedback or questions welcome.
I hope my first post wasn't to long and painful to read: I promise the next will be shorter!
Cheers,
Sean
|
