Recurring P0172 on Stage 3

I've been having a problem for more than the past month that I can't seem to diagnose. I got the car back from the tuner and it ran fine. However, after a couple weeks of driving it threw a p0172 "rich" code. I cleared the code and reset only for the code to come back. Since then, I've cleaned the MAF/IAT sensors, cleaned the air filter, checked for exhaust leaks preturbo, checked vacuum lines and fmic couplers. Nothing obvious that I'm seeing.

 

When the ecu gets reset, the afrs balance out at around 14.7 or so and corrections tighten up to normal after a few miles of driving. A couple days later, the code comes back. Corrections are maxed out by this time. Interestingly, it only consistently runs rich at idle. Someone take a look at the logs and LV and let me know if I should let the tuner have it back first or diagnose and fix a mechanical problem first. I don't want the tuner to get it back only to mask something that's causing this and I don't want to throw parts at it blindly. Current mod list is here.

 

http://legacygt.com/forums/showpost.php?p=3654317&postcount=1

 

The following is the latest learning view. The first log is a cruising log. The second 2 are 3rd gear WOT logs. I have some knock that can easily be addressed, but I don't think the AFRs look out of line. The tune is not completely finished, but when I got the car back, the maf had been scaled and corrections were spot on. Any advice would be appreciated.

romraiderlog_20121103_191758.csv

romraiderlog_20121103_192706.csv

romraiderlog_20121103_192836.csv

What intake do you have?

Have you changed it since the tune?

Please post the low end of your MAF scale, like 0-35 g/s. That way we can compare it to a known-good scaling.

 

My guess is that the MAF scaling just needs to be adjusted. Leaks would cause it to run lean, not rich, and I can't think of anything else... Maybe someone else will have better ideas though.

Thanks for the reply.

 

I've got the AEM intake and it has not been changed since tuning and it was scaled.

 

Here is a datalog of my vehicle idling and a screenshot of the maf sensor scaling. If there's anything else that might help, I pulled the ROM from the ECU so it's readily accessible.

romraiderlog_20121104_134304.csv

What injectors do you have, and what are they scaled to?

 

My AEM scaling has values about 10% higher than yours, so switching to mine would actually make this problem worse. But I had aftermarket injectors so it's not an apples-to-apples comparison... I'd forgotten about that.

My injectors are the DW sidefeed 1100cc. They are scaled to 1085cc and their static flow is rated at 1214cc.

There's nothing obviously wrong with your MAF scaling in that area, but I still think that it's worth revising it to see if that fixes the problem. Subtract 10% to everything from 0-35 g/s and smooth the transition from 35-45 or so. Drive it like that for a day or three, check the AF Learning values periodically, see what they do. I can't think of a mechanical issue that would cause this, so the MAF scaling seems like the most likely culprit.

 

If that fixes the problem, then great. If not, then you can just go back to your current tune. It would only take a few minutes to make the change, and a couple/few days to test it, depending on how much you drive.

 

And at the same time, I'd also suggest smoothing out that zig-zag shown in the middle of the MAF curve in that screenshot. I doubt that it's really ideal.

if everything was fine, and a learning view and/or corrections/learning was checked during in-car cruising then maybe you have stuck injector(s).

 

at first glance I immediately thought injector scaling/latencies being off since it's severely rich across the ranges(except for OL but use a wideband to confirm).

There's nothing obviously wrong with your MAF scaling in that area, but I still think that it's worth revising it to see if that fixes the problem. Subtract 10% to everything from 0-35 g/s and smooth the transition from 35-45 or so. Drive it like that for a day or three, check the AF Learning values periodically, see what they do. I can't think of a mechanical issue that would cause this, so the MAF scaling seems like the most likely culprit.

 

If that fixes the problem, then great. If not, then you can just go back to your current tune. It would only take a few minutes to make the change, and a couple/few days to test it, depending on how much you drive.

 

And at the same time, I'd also suggest smoothing out that zig-zag shown in the middle of the MAF curve in that screenshot. I doubt that it's really ideal.

 

So decrease the voltage in that range to lean the fueling out?

 

Could a faulty fuel pressure regulator or, like ripemeat stated, a stuck open fuel injector cause a rich condition only at idle?

I'd leave the voltages untouched and just raise the MAF values.

 

If you have confirmation that the AFR is on target in WOT pulls, that would suggest that the problem is not a stuck injector.

Thanks for your advice. I flashed the ECU with increased values and initially, it ran rich with corrections maxed and after some learning, it found it's way back with corrections well within the proper ranges. We'll see if it stays that way after a few days of driving and I'll keep my eye on it. Now might be a good time to install that AFR guage that's been sitting in my basement for a few years.

after a flash the ecu will command fuel pump duty cycle to be 100% for 10-15 seconds and the afr's will be pig rich. the pump duty will drop down and the afr's will swing lean and then quickly revert close to 14.7. log a/f correction 1 and a/f learning 1, you'll see the a/f corrections (immediate ecu correction) move around at idle..... this will be averaged into the learning 1, when you see the learning 1 move, the correction should get closer to zero.

Car threw the code again so I reset. Then, like clockwork, a few miles after the reset, code came back as I watched the afrs dip to the 12 range at idle. when I got home I noticed that the afrs were back to normal but the light was still on. Ran a couple errands logging the whole way and finally the light turned off so we'll see what happens.

sounds like the cyl is loading up with fuel at idle. if you dont want to mess with the injector latencies use the romraider maf tool....it works pretty well....you'll have to do it a couple times to get it right because i think you're pretty far off in terms of scaling. do about 10 min of idling, 30 min of around town driving, and 20 min of highway driving all with gradual changes in throttle, don't punch it. vary your speeds and loads. before you do this, you need to open your maf scaling table up in ecu editor..... it will not change the maf scale for you unless you open the table up first- and then open logger and start the maf tool. Works pretty well, the longer you log using the tool the more accurate it will adjust. the more you use the tool the better off you are. Don't sweat over +/-6% in the a/f trims.

Thanks for the tip. Since I have done no tuning on this, I'm going to have to read into how to use that tool. Doesn't sound too difficult. I think my tuner has used that to set the maf set so I'll have to ask him. I'm beginning to think I should learn more of the basics of tuning myself. Once I get this figured out and a couple other spots, hopefully it will be about time to turn up the boost.

google A subie noobie tuning guide. its a good read for someone starting off learning about subaru ecu's and knock control etc. the romraider tool is easy to use and its only intended for closed loop scaling (I.e. out of boost). you can do that safely without a wideband, the open loop fueling(~>40-50g/s) you'll need a wideband and have to manually scale the maf scale

Problem persists. Continually running pig rich only at idle. 11-12 range is about where it settles. Car ran without the cel for 2 weeks, now it only takes a day before it throws the code after a reset. A local tuner friend of mine (who's actually not tuning my car) thinks it might just be the nature of these big sidefeeds. He reports that he's tuned a vehicle with big sidefeeds that idled rich no matter what until they dropped in a set of ID1000s. I'm not ditching these injectors just yet, though. I've had them on the car over a year so I'm taking it to another friend (the top subie tech in the area) to eliminate mechanical issues that might be causing this. I'll keep you posted. Thanks for the suggestions and if no problems are found, I'll revisit this thread looking for solutions to this.

 

One other thing I've read is that these big injectors perform better when running e85 through them. Anyone attest to that?

if its only at idle you can still get closer than 11-12. start scaling the maf range 0-5.6g/s down by a few percent and try again, I betchya you'll find a good improvement. You'll end up at a point where you get close and then change it a tiny more and now you're a little lean, you might not get it perfect.

Start scaling it down? If I pull what it's reading for air, wouldn't it richen the mixture?

 

One of the strange things about this is that the first 1 or 2 driving cycles after a reset, the AFRs will find themselves in the 14-15 range no problem. Then, after a couple cool downs and warm ups, it'll be back at the 11-12 range. I've got access to a set of PE850cc injectors that i could try out at no charge to myself to determine if this is an injector issue so that may be coming up.

 

Otherwise, what else in the actual tune should I look at. I noticed that tip-in enrichment at 0 for some reason was set to .524 so I zeroed that out, but that made no difference. Any warm-up enrichment tables or anything like that I should look at?

it takes a while for the lv trims to settle because they are LTFT (long term fuel trims). you need to scale the maf range 0-5.60 DOWN if idle is your problem. the ecu reads maf voltage and based on the maf scale table asserts x g/s of air. then it figures a certain amount of fuel in relation to the g/s that is set in the table. therefore if you lower the g/s in the range of 0-5.60, the ecu will account less fuel, therefore your ecu will not have to keep pulling fuel in order to reach 14.7.

Well, I think I finally have this one figured out and I can move forward.

 

When studying my ROM and comparing my issues with others I've searched out, I noticed that my latency values looked a little high. I found a multiplier value of .64 that was being used when tuning OS. I simple multiplied the flow sheet latency values by that figure and plugged them in. Car has not been idling rich. Usually after a couple driving cylces it would richen to 13, then 12 and eventually down to 11 range at idle. That hasn't happened.

 

Now I'm thinking my maf scaling could use some attention. My corrections are +15-20 so I need to figure out what I need to change to accomodate. Back to the drawing board.

high latency values would mean your fuel trims would be lean. because latency is dead time between injector fires.

Higher than correct latency values tell the ECU the injectors are slower than they actually are so the ECU starts injecting fuel sooner than needed. Therefore, the mixture is rich. It's been 5 days since I made the change to the injector latency and the car is still stoich at idle. No cel either. That's the only thing that has changed from the map that originally ran that learned to run rich at idle. Before, corrections maxed out towards the negative, and now we're still in the +umpteen range in corrections. I still have to fix that. With improper latencies, I'll probably have to go back to a known proper maf scaling and work from there. Injector scaling may also need attention.