Knock event registration

[darthqwo]

I reset my ecu memory three days ago along with a tcu reset on my 2008 legacy 3.0r. I am using the btssm android app, which is giving me knock events once every five minutes on average usually on low rpms and low speeds. I reseated the airboxes to throttle and been meaning to clean the maf sensor while i find a maf cleaner spray.

 

I saved my timing charts through the days which initially showed a few negative values but now after four days are very negative. If someone knows how to read these charts, which ive uploaded in order, id appreciate your insight.

 

omg thanks.

[relative4]

Low fuel pressure.

[utc_pyro]

1) STOP USING CRAP GAS IN THE POOR THING! Give it at least midgrade if not premium, even NA Subaru motors are knock limited.

2) You appear to have an intake leak somewhere, it's adding quite a bit of fuel. You may want to try and clean the MAF as well.

[darthqwo]

Could it be a bad pcv valve; a couple weeks ago i halfway took it off to check for carbonation, and the rubber pcv hose seems bit calcified; will be replaced eventually.

 

Im getting a maf cleaner tomorrow. I can count the number of times i have gotten mid/premium with one hand but will give it a try next. How should i try to correct fuel pressure?

[relative4]

Pyro makes a very good point with the gas. The trims do not indicate a vacuum leak.

Follow the procedure in the FSM to test the fuel pressure. Do not get the Harbor Freight gauge, it's worthless. A decent gauge is $50 at AZ, etc.

[darthqwo]

The chart seems to normalise on its own; no idea if it is normal to see this kind of variation. The last chart immediately follows MAF sensor cleaning and a short drive; it seems to make a difference, but the fuel correction jumped immediately while merely idling. If it is helpful i have some more obd sensor readings after cleaning the maf sensor.

 

I am still testing fuel pressure.

[darthqwo]

Completed fuel pressure test; AZ has a loaner for this! The gauge read 52 psi during idle and during load. It is 1.5 psi above the rnage specififed in the manual, whihc only suggests replacing fuel pump or lines if below; would anyone suggest replacing a vacuum line--i am ot sure where the vacuum line is, as it is not a typical fuel pressure regulator layout--and picture attached is.

 

I also reseated the maf sensor and to had to delete the maf rubber gasket to do so; the numbers normalised further, but wondering what issues there still might be. The only other issue i have while driving is delayed/slow throttle response most noticable from a stop but seems to be present at any speed when reengaging the gas pedal. There is no delay when revving in neutral. While in drive i can usually step on the pedal for up to .5 sec and release without any increase in rpm or speed; if stomping on the pedal, almost flooring, it takes maybe .25 sec before any response. If doing this while in motion the car will jerk forward due to sudden gas pedal engagement and seems to replicate the shift shock that seems to be been learned into the tcu previously before i had reset it.

[utc_pyro]

Fuel pressure issues are extremely rare on these cars. They have turbine style pumps with long lifespans. Your fuel pressure is fine.

 

 

 

You have ether a MAF issue or an air leak. Look up how to built a boost leak tester or go to a shop that has a smoke machine.

 

 

 

That "gasket" (o-ring) is actually really important as it can screw up the readings. You'll want to replace it, or buy a new MAF that includes one. You don't need to buy a new MAF from Subaru, but it needs to be at least a Denso unit (who Subaru buys them from).

 

 

 

And again, if you plan on this car living long, please put 93 in the poor thing until you get the issues resolved. After that 89, not the cheap stuff.

[relative4]

As the problem is not fuel pressure, and the trims do not indicate a post-MAF vacuum leak, I agree with Pyro that the next place to look is at the MAF. Not sure how you "had to delete" the O-ring, but presumably it was stretched. When I've experienced that and didn't have ready access to a replacement, I cut out a short section and made the remainder just the right length to make a seal. Anyway, you cannot run properly without a seal there; if the old O-ring wasn't sealing properly, perhaps that was your problem. Unfortunately, the O-ring isn't easy to get by itself. I've heard part 806925100 will work; haven't tried it myself.

 

Another possibility is dirty injectors. If flow through the injectors is impeded, the expected symptoms would be consistent with the trims you're getting. And this would make sense given the long history of cheap gas. Have you run a couple tanks of fuel system cleaner through it yet?

[utc_pyro]

Again, I disagree about it not looking like a post MAF leak.

 

A) this isn’t a turbo car. You won’t see the carcteristic change to pulling fuel once you get into light boost like your’re expecting.

B) he’s not getting it into the high load/RPM ranges anyway.

 

So I think his “D” learning column can more or less be ignored.

 

We also have issues of in knocking like hell. To get this bad he’s running crap gas for sure, but there is also something else making his MAF read too low and thus apply too much timing. There should be no way to see THAT much timing pulled when IAM is 0 otherwise.

 

If it was clogged injectors they wouldn’t clog this evenly. You’d have a few that got unlucky first. Thus you’d probably see misfires as well, which his BTSSM screen shot isn’t showing. And again, the fuel systems on these cars are quite good. We have two fuel filters, with great filtration on the second one. We also have filters in the injectors.

 

I’d look for post MAF pre throttle body leaks first as the trims are flat ish. Then PCV system leaks (OP stated he messed with it), then general post throttle leaks.

[relative4]

Well, certainly can't go wrong with a smoke test.

[darthqwo]

Thanks for the feedback! Misfires being rare is very true; attached is screenshot with a longer run and higher speed range.

 

So, i will reattach the gasket. Despite fuel pressure being good is it possible fuel filter can be the culprit yet?

 

Somehwere it is said seafoam put in the gas tank can clean the fuel injectors; so in the meanwhile i will use one bottle here and the other bottle to clean the intake area. I am wondering if i can snort this up the brake vacuum line (by #1), which idea comes from

.

 

Researching delay throttle makes vacuum lines post MAF somewhere suspicious from, say, clog built up, say from poor gas over the years? Looking at the diagram the vacuum lines seem to be the connections at C, D, E and the pressure assembly filter marked by 4. The only other things remaining are the pcv valve and the brake vacuum line.

[VespaGTS]

Smoke that car! Get new plugs, clean the MAF as well and try again with better gas. Premium is all I run in my car, but I don't have to drive far.

[darthqwo]

Not sure if it is a proper smoke test, but the seafoam method generates a lot of smoke, some of which seemed to emerge from engine pass side, though this could be almost anywhere still including possibly from undercarriage exhaust. Just a hint to anyone wanting to do this; after snorting the seaform, the engine can be let to idle during heat soak instead of being off, and the smoke does not blow out unless revved.

 

I ordered new pcv valve/hose, map sensor filter/hoses (prev. ind.), intake manifold throttle gasket. The hoses are being replaced as they are calcified, but i noticed i never reclamped the pcv hose after playing with it the first time, and this immediately changed some of the numbers; see attached.

 

After removing the caked carbon off the inner plate of the throttle, the timing numbers have all gone to 0s; see attached #2. The amount of cake that i removed is probaly the culprit behind the poor throttle response; the pedal now commands much improved control of the electric-motor throttle to the point there is no more jerky response from transmission. I will have to drive around some more to see if this result is real.

[relative4]

Wow, nice work. Now the vacuum leak is clear.

[Perscitus]

OP - you need to understand what you are looking at more.

The fact that the in-memory learned timing values have gone to zero has as much to do with RC deciding to bump your IAM/DAM up to 0.5 as anything you did to cause that to happen.

 

 

 

LV snapshots are useful but nowhere near as useful as targeted logs with a. high polling rate b. right selection of fields (varies depending on what you are doing, trying to track down).

 

 

Read more about RC, FKC, FLKC, MAF transfer function, its effect on calculated Load, Timing... have a look at Learned Timing, Base Timing, KCA table, etc.

 

 

Log, re-read, try to interpret logs, re-read, interpret again, rinse/repeat until you understand what you are looking at and how the ECU and sensors respond to driving cycles.

 

 

https://legacygt.com/forums/showthread.php/everything-you-wanted-know-subarus-knock-prevention-strategy-254139.html

[darthqwo]

Thanks for the link; it was nowhere to be found! I was wondering if something was wrong with my voltage or my ecu partially reset, as this did happen before; during the first few drives it changed a few numbers from 0 but would revert back to all 0's and seems to still happen. But since then my timing chart generally repopulates with IAM no longer at 0.0

 

I definitely need to take better directed snapshots; but these are some more recents ones.

[relative4]

Wow, your IAM is really in the basement. Have you smoke tested yet?

[BMCracer]

I'm watching with much interest as I have a 2007 facelift 3.0 / 5EAT with very low IAM and lots of knock events. My LV charts looks like yours, except my Learn% 2nd row usually has negative values in there.

[darthqwo]

I'm watching with much interest as I have a 2007 facelift 3.0 / 5EAT with very low IAM and lots of knock events. My LV charts looks like yours, except my Learn% 2nd row usually has negative values in there.

 

Can i ask what kind of mileage you're getting on the gallon? im just clueless, but negative numbers in the fuel learn would indicate good fuel economy?

 

I am still trying to diagnose why my fuel learn values are so high; the fuel consumption certainly matches, as i am getting between 15-18 mpg overall. This is with new spark plugs, good ignition coils, and new fuel injectors; in addition most vacuum hoses are new.

 

Its been a few months since i updated my story here, but im inclined to think that 87 vs 93 octane fuel is not the cause of knock events; rather, the higher grade octane seems to increase knock events, causing iam to hover between .3 and .6 iam; the immediate response with a full tank of regular is to go back to 0.0 - 0.03 range and register less knocks.

 

For spark plugs i have gone with ngk lfr7aix, whichchoice might be partially responsible for the fuel numbers. Before the new spark plugs install i did a a cold compression test that had yielded 182-195 psi on all cylinders with the exception of cylinder 1 on forwards passenger spot, where it is around 140. This is a possible cause of the fuel issue, but since the new spark plugs, the most common cylinder for misfires has been cylinder 4.

 

any insihgt friends?

[BoozeRS05]

When you see large variation between the cylinders that's an issue, for example 182-190-195-140. I don't remember what the percent value is, but larger than 15%(?) variation means you have a problem.

 

Have you leakdown tested? It's probably time before you keep throwing money at it.

[BMCracer]

Can i ask what kind of mileage you're getting on the gallon? im just clueless, but negative numbers in the fuel learn would indicate good fuel economy?

 

I'm averaging around 12.2litres per 100km ( 23 mpg ).

 

My car has just turned over 150,000km, and I don't know if / when it had a set of plugs, so I got a set of standard factory plugs in my last order from Partsouq. So I'll fit them one day soon and see if that changes anything.

[darthqwo]

When you see large variation between the cylinders that's an issue, for example 182-190-195-140. I don't remember what the percent value is, but larger than 15%(?) variation means you have a problem.

 

Have you leakdown tested? It's probably time before you keep throwing money at it.

 

The manual says 7% or so difference is the tolerance range. Anyway, any power loss while running the car is not noticeable.

 

So, for the leakdown I need to get an air compressor?

 

I'm averaging around 12.2litres per 100km ( 23 mpg ).

 

My car has just turned over 150,000km, and I don't know if / when it had a set of plugs, so I got a set of standard factory plugs in my last order from Partsouq. So I'll fit them one day soon and see if that changes anything.

 

Can i ask what you get as highway economy?

 

 

Anyway, there seems to be a difference between fuel efficiency and fuel consumption, and low compression would cause poor efficiency for sure, but the fuel consumption is dynamically programmed into the ecu in my case if anybody has seen my fuel trims. Given this, I wonder if the low compression in the single cylinder is not the cause of my excessive fuel consumption in bank 1. Im basically running out of things to test, the the only things left are to look at are possibily clogged catalytic converts and oxygen sensors; but the former would require a new exhaust manifold.

[BMCracer]

Can i ask what you get as highway economy?

 

Dunno cos I hardly ever get to run on the highway or open road

[08 LGT]

IAM at anything other than 1.000 is a sign something is wrong. It dropping that far low means something is very off - the car is severely retarding timing to compensate for detonation/knock. The IAM is the ECU's last and most aggressive line of defense (adjusting FKC and FLKC being the first two).

 

You should take your car to a Subaru shop if possible (or a trusted shop if a Subie-specific one isn't available) and have someone with more experience do a legit smoke test. I had similar issues with wacky fuel trims and a similarly low IAM (it dropped to 0.000) that were caused by a vacuum/boost leak in the intake manifold.

 

Something else to consider looking at is the upstream 02 sensor, as a malfunctioning one can also cause nutty fuel trim adjustments.