Intake Calibration is good but AFR learning moves a lot

I thought I did a fairly decent job of calibrating my intake early this year, especially focusing at the 40-60g/s range, as that impacts the higher flow ranges in open loop.

 

Please see the MAF check that I just graphed from about 2 hours of highway driving.

 

To me my intake scale looks pretty good, yet my AFR learning D is currently -5.47%, it moves around quite a bit, is usually negative, and worries me when it gets to -5.47 or -7%, sometimes it stays near 0% for a few weeks. What am I doing wrong?

 

I always fill up a the same gas station, so the fuel should not be changing too much.

 

thanks for the help,

Jeremy

intake cal Jeremy Butts 10-6-09.pdf

what is your D setting? +50g/sec

what is your D setting? +50g/sec

 

Coming from a guy who learned to tune on a WRX.

 

Most LGTs and STIs that I have seen have the "D" range at 40+ g/s.

what is your D setting? +50g/sec

 

I use Cobb Street Tuner and do not have the option of changing the D range. I believe it is 40+ g/s.

 

Jeremy

Coming from a guy who learned to tune on a WRX.

 

, a mighty fast 2.0L WRX.....

 

I've changed mine to +50g/sec on D, it keeps the D range zeroed out. I've also changed A to <10g/sec to keep it consistent.

 

If your D is floating around that much how is the rest of the ranges doing?

With Street Tuner you can change the range, it is in the base map, requires a new base map to be flashed to the car. (just highlight the 40 and change it to a 50 if you want)

 

IMO, I like the stock settings (5.6, 10, 40) I want D to change based on fuel I use. Low idle is below 5.6, and higher idle is between 5.6 and 10. If I blow a vacuum hose range A will change quickly to +15, but range B does not need to change quite as much to accommodate a leak.

 

At any rate, you do not need to change anything to get your car working great.

You know I had trouble stabilizing my AFR Learning ranges too. I have modded stock injectors with an AVO intake. What I did was change the stock latencies. Now all trims are consistently within 3% with D at about .3%..(set at 40+ g)

, a mighty fast 2.0L WRX.....

 

I've changed mine to +50g/sec on D, it keeps the D range zeroed out. I've also changed A to <10g/sec to keep it consistent.

 

If your D is floating around that much how is the rest of the ranges doing?

 

The rest of the ranges are within 2%, including idle (unless idling for an extended time and it creeps).

 

I have a K&N Typhoon, wrx fitment 20G with FMIC, and 650cc injectors, 2005 model.

 

Jeremy

Wandering trim makes me suspect that the MAF error at (for example) 42 g/s is much different from the MAF error at (for example) 60 g/s. So depending on what flow rate the car last spent time in, you'll get a different fuel trim.

Wandering trim makes me suspect that the MAF error at (for example) 42 g/s is much different from the MAF error at (for example) 60 g/s. So depending on what flow rate the car last spent time in, you'll get a different fuel trim.

 

My thoughts too, except that over time the AFR learning D always seems to trend negative, even though the MAF scale looks to be +-2% max.

Can this be temp induced error or has it been off all summer? Could be the temp compensation table.

I can really see the comp table affecting D range that much.

Can this be temp induced error or has it been off all summer? Could be the temp compensation table.

 

Its been varying all summer.

 

Jeremy

So it stays negative after 200 miles or so of driving??...AFR learning D..?

I can really see the comp table affecting D range that much.

 

I think you meant to write "can't really see..."..and I agree, attached is the comp table. At that low g/sec there is very little adjustment.

 

 

Maf Intake temp comp.csv

So it stays negative after 200 miles or so of driving??...AFR learning D..?

 

Yes, and yes.

 

I reset ECU and it stays near zero for a couple days, it's stable enough that I am happy to remove the laptop from the car and continue with life while not worrying about my tune, then a month later I check AFR learning D and its negative. When it gets to -5% or -7% its worrysome...

 

I drive about 75 miles/day.

Jeremy

I think I know what the problem is....

 

AF learning D only changes when you're running over 40 g/s and (critically) you're in closed loop mode. That's a pretty thin line, because unless you're going uphill, that much airflow will accelerate you into the open loop mode within seconds.

 

My guess is that your 40+ maf scaling is indeed 5-7% off, and it just takes a while before you spend enough time in the 40+ range for the fuel trim to show it.

 

I set the D trim by driving up a mountain pass. If you have one nearby (or even just a big long hill), set up RomRaider with MAF and OL/CL status prominently displayed and you'll see what I mean about the fine line between 40 g/s and open loop.

I think I know what the problem is....

 

AF learning D only changes when you're running over 40 g/s and (critically) you're in closed loop mode. That's a pretty thin line, because unless you're going uphill, that much airflow will accelerate you into the open loop mode within seconds.

 

My guess is that your 40+ maf scaling is indeed 5-7% off, and it just takes a while before you spend enough time in the 40+ range for the fuel trim to show it.

 

I set the D trim by driving up a mountain pass. If you have one nearby (or even just a big long hill), set up RomRaider with MAF and OL/CL status prominently displayed and you'll see what I mean about the fine line between 40 g/s and open loop.

 

+100

 

I put OL/CL status and MAFg/s on the logger dashboard and drive above 40 and make sure the OL/CL status stays on 8 (I think 8 is CL, and 10 is OL)

If he is scaling MAF with the RR tool, he will see it there as well (when you interpolate). Wouldn't it have been noticed earlier?

That depends how many data points he collects in the 2.6-3.0 volt range.

I think I know what the problem is....

 

AF learning D only changes when you're running over 40 g/s and (critically) you're in closed loop mode. That's a pretty thin line, because unless you're going uphill, that much airflow will accelerate you into the open loop mode within seconds.

 

My guess is that your 40+ maf scaling is indeed 5-7% off, and it just takes a while before you spend enough time in the 40+ range for the fuel trim to show it.

 

I set the D trim by driving up a mountain pass. If you have one nearby (or even just a big long hill), set up RomRaider with MAF and OL/CL status prominently displayed and you'll see what I mean about the fine line between 40 g/s and open loop.

 

NSFW, thanks for your explanation

 

I log with Romlogger

 

I do use long uphill sections to tune the D range, I also have found that logging while loaded down with the bikes on the rack creates enough drag that I stay around 45g/sec cruising at 75 mph. Add in an uphill and I am easily at 55g/sec just to maintain speed. I do watch g/sec and OL/CL while logging.

 

I look for 200+ data points in order to feel confident in the trim at every MAF voltage.

 

FYI - the D range just switched to -2% yesterday.

 

Jeremy

I mentioned before I had similar problems stabilizing D like you. I would say the scaling is off but in your case I don't think that's happening. Just as a sanity check, could you check the vacuum lines and connections post turbo outlet to TB? Check for leaks etc.

OK, so much for that theory...

OK, so much for that theory...

 

Man don't say that. It was your post on RR that made me decide to tweak around with my latencies to get my D range down to +0.3. Only reason my idle trim A floats up to 5% is because of the heatsoak with the AVO intake without a shield.

I will continue to log and tweak the intake calibration, if required.

 

I'm beginning to think there may be variable amounts of ethanol in my fuel, even though the gas station owner (mobil) claims no ethanol. I will probably switch stations for a months and see if that stabilizes things.

 

Jeremy