I've been banging my head on this a little while and I wanted to see what the collective input is on my current tuning situation.

I'm trying tune out knock and clean up my fuel maps for California 91 octane. I've got it almost all sorted out except for the intersection of 200ish load and 5000 rpm. My wideband reads 12.3 to 11.8 while reving through, but my fuel maps, both low and high octane have that section at 11.3 (high) and 10.5 (low). Obviously I'm not hitting my target.

Things I've tried:
- ridiculously rich targets in both maps but I can't seem to get the AFRs much lower
- ramp up the enrichment from 100 load on to 200 from 12.2 to 11.1, no change
- retard the timing across the same range

If you care to look, I've attached logs of five 3rd gear pulls I ran tonight. Feedback is welcome on other aspects that aren't exactly on topic.

I leapt to the conclusion that it may be time to get an after-market fuel pump. Would you agree?

 

a simple test would be to make the settings in the a/f map to where the car SHOULD be pig rich.. you mentioned that this cannot be done. Log the run to verify that it did not get richer. Now, lower the boost considerably and see if the a/f gets pig rich.

if it does then you do have a fuel pressure problem. clogged injectors, filter, kinked fuel line, fuel pump or fuel pressure regulator issue etc..

 

The fuel maps at high loads are not absolute fuel targets. 11.3 is set too lean. Your logs show 11.9 AFR at 5000 on 91 octane.

Try setting the fuel map .5 to 1 full point richer there and see how the car reacts. You're not running out of injector because your IPW's seem fine.

 

Thanks for the clarification.

I read in a different thread that if you correctly scale the injector sizing, the fuel maps will read more "true". I've dropped my own (OEM injectors/OEM pump) from 503cc to 464cc and managed to get with .2 or so of actual AFRs in the maps.

Since that process will likely change my LTFTs for the worse, do you know of a way to index the fuel maps' readouts to be based on injector scaling? Ideally, it would be a user-inputted value because of the likely large variations arising from fuel makeup, octane, and I'm sure many other reasons.

 

Well no matter how you scale it your logs still show AFRs that are too lean IMHO.

Once you get more comfortable with tuning fuel maps, you will notice the actual numbers don't really matter, think of it as a +/- system. For me, the numbers only come into play for smoothing.

 

Hi

IMO you are running too much ignition timing as well. 7 at 4600 is to much for 91. The JDM timing map is set at that for what would be about 96 octain in the US.

MB

 

wow, how can you guys read those logs?* ugh.

could you be getting a misfire in that region? it will show as lean on the WB.

I sometimes think that when the ecu pulls a lot of timing it acts like a misfire on the afr log.


* is that log display a result of file transfer? your original display is easier to read?

 

This is a very good point. However, misses don't show up the same way on all widebands. It really depends on how the controller handles infinity. Some widebands will show even a single miss as a distinct spike. Others act as you describe.

For tuning and diagnosis purposes it is probably good to know how your controller responds to misses and to gas pressure changes. Since different lambda meters respond differently to these cases.

Razorlab's point is also very good. Although the numbers can look official, the AFR is largely a guess on the ecu's part. So, you have your actual (logged), your desired, and the guess (ECU table). Use the first two to tweak the third and don't worry too much about the specific values.

-jjf