[matt88-8]

I downloaded the btSsm app on my phone hoping to get more info than torque could get me. During my first drive with it i noticed a little icon at the top, that apparently is a knock indicator.

The car seems to drive fine, I cannot feel anything strange. it only knocks at higher throttle inputs, if i drive gentle it seems fine.

Car is a 07 with the 2.5 N/A engine.

KNOCK DETECTED at 2017-09-23 06:35:03 PM
AFR: 15.7
Boost (psi): 0.00
Coolant (degC): 87
FKC (deg): -1.05
FLKC (deg): 0.00
Injector PW (ms): 12.0
Intake (degC): 33
Load (g/rev): 1.07
MAF (g/s): 40.01
Oil (degC): 92
RPM: 2254
Speed (mph): 17.4
Throttle plate (%): 96.9
Timing (deg): 19.0


Is this something I should be immediately concerned about?? I can try different gas, but it will take a few days to empty the tank. I'm a bit freaked out because i just did my head gaskets about 500km ago, but again, I haven't used this app before and nothing feels different. no dash light, fuel milage is the same etc..

What do you wise people here think??

 

[jwrezz]

What fuel do you run? I know my H6 really prefers the 93. My wife's name 4 Forester is fine with 87. Haven't ever tried to see if it's mileage improves with higher octane.

Sent from my Nexus 6 using Tapatalk

 

[1 Lucky Texan]

07 is not all that old but, many older subes get new knock sensors - they are cheap and easy to try.

also, maybe you got some bad gas?

 

[plain OM]

The -1.05 degree correction is the minimum the knock correction can apply, so even a minor knock will result in this change.

Disable the knock notification, or set the alarm/trigger point (FKC) to something like -4 degrees. (The default is 0 degrees, so any timing correction triggers a report.)

I've been dealing with this for over six years now. http://www.subaruoutback.org/forums...bo-what-about-iam-what-value-do-you-have.html

You should be good with good 87, but might find it a bit better moving to 91. Again, been there, done that.

Does the band at the top of the screen show a button for LV? If so, select that. A table will appear on the screen. It can be saved, and sent/transferred to your home computer etc as a .jpg file. Share it here. The FLKC part of the table will be more telling than a single knock event.

(BtSSM works similarly to Romraider, and for full functionality, requires that the ECU be "defined". Your 07 might have the same ECUID as mine, and if so, the definition is available. Also, the default list of parameters in BtSSM is not complete, and has parameters not applicable to a non-turbo, such as Boost. You can add parameters ("custom fields") which will expand the number significantly. I'm using BtSSM as well as a "permanent" dashboard: http://www.subaruoutback.org/forums...e/249641-bluetooth-alternative-romraider.html)

 

[matt88-8]

I will get the learned values for you.
It's Petro-can gas, tends to be good stuff. I will try something else next time.
Is the -1.05 going to change much? My butt dyno says no, but maybe it's always been like this...

Sent from my SM-G903W using Tapatalk

 

[plain OM]

Is the -1.05 going to change much?

No. It's one degree, at that moment, in 19 degrees of total advance at that same moment. FKC is a correction applied at the time when knock is detected. The rest of the time FKC is zero.

 

[matt88-8]

Lots more info from today's drive. I used full throttle several times hoping to get complete data. Seems like the knock is worse than I thought. Strange that I don't feel anything.


KNOCK DETECTED at 2017-09-24 10:15:54 AM
AFR: 13.1
Boost (psi): -0.40
Coolant (degC): 87
FKC (deg): -1.40
FLKC (deg): 0.00
Injector PW (ms): 15.6
Intake (degC): 30
Load (g/rev): 1.14
MAF (g/s): 89.51
Oil (degC): 92
RPM: 4720
Speed (mph): 49.7
Throttle plate (%): 100.0
Timing (deg): 20.0


KNOCK DETECTED at 2017-09-24 10:15:28 AM
AFR: 13.8
Boost (psi): -0.40
Coolant (degC): 86
FKC (deg): -4.04
FLKC (deg): 0.00
Injector PW (ms): 14.8
Intake (degC): 29
Load (g/rev): 1.16
MAF (g/s): 86.50
Oil (degC): 92
RPM: 4489
Speed (mph): 49.7
Throttle plate (%): 100.0
Timing (deg): 17.0


KNOCK DETECTED at 2017-09-24 10:15:28 AM
AFR: 13.9
Boost (psi): -0.40
Coolant (degC): 87
FKC (deg): -1.26
FLKC (deg): 0.00
Injector PW (ms): 14.8
Intake (degC): 29
Load (g/rev): 1.17
MAF (g/s): 84.93
Oil (degC): 92
RPM: 4365
Speed (mph): 48.5
Throttle plate (%): 100.0
Timing (deg): 18.0


KNOCK DETECTED at 2017-09-24 10:15:27 AM
AFR: 14.0
Boost (psi): -0.40
Coolant (degC): 87
FKC (deg): -1.25
FLKC (deg): -5.00
Injector PW (ms): 14.3
Intake (degC): 29
Load (g/rev): 1.13
MAF (g/s): 79.57
Oil (degC): 92
RPM: 4241
Speed (mph): 46.0
Throttle plate (%): 100.0
Timing (deg): 15.5


KNOCK DETECTED at 2017-09-24 10:15:26 AM
AFR: 14.1
Boost (psi): -0.40
Coolant (degC): 87
FKC (deg): -2.10
FLKC (deg): -5.00
Injector PW (ms): 13.8
Intake (degC): 30
Load (g/rev): 1.10
MAF (g/s): 72.03
Oil (degC): 92
RPM: 3926
Speed (mph): 41.0
Throttle plate (%): 100.0
Timing (deg): 15.5


KNOCK DETECTED at 2017-09-24 10:15:25 AM
AFR: 14.1
Boost (psi): -0.40
Coolant (degC): 87
FKC (deg): -1.05
FLKC (deg): -5.00
Injector PW (ms): 13.8
Intake (degC): 30
Load (g/rev): 1.12
MAF (g/s): 72.04
Oil (degC): 92
RPM: 3855
Speed (mph): 39.8
Throttle plate (%): 100.0
Timing (deg): 17

Sent from my SM-G903W using Tapatalk

 

[plain OM]

Lots more info from today's drive. I used full throttle several times hoping to get complete data. Seems like the knock is worse than I thought. Strange that I don't feel anything. . . . .

First off, "yay!"; finally another 07 apparently using the same ECU/CALID as mine (E2UG001B). If you've looked at my linked "IAM" thread, this has been a goal for a long time. Side benefit -- we can compare data for identical engine management systems.

Also, and relevant to my IAM thread, I notice that the A/F Learning in the LV tables is similar to mine; i.e., the A/F Learning steps up with the increasing MAF ranges. So it appears this might not necessarily be a fault as I had thought; it could very well be related to the calibration in the E2UG001B ROM tables. (Or, we have the same fault and should win the lottery this week!)

All your knock event records are at relatively higher engine loads, i.e., above 1 g/rev. I have knock events recorded on my car with engine loads far lower. I've since turned off the reports because of the clutter, but the "knock sum" parameter still increments up, often while cruising at light loads. After six years of the same with no change or apparent deterioration, I don't disregard this, but I also no longer worry much about them -- as long as there isn't a significant change in the pattern . . .

Also interesting that, as in my case, there doesn't seem to be outward signs of knock, e.g., the classic metal knock sound from the engine. I've never heard it, whereas there are posts where others with similar engines have clearly heard the knock.

As for the FLKC, with the exception of the -5.00 and -4.65 degree corrections in the higher load, 3600 - 6400 rpm range, the pattern is not unlike what I have as well. (A number of less than 1.0 corrections, perhaps one or two in the 1 to 2 range, and even some positive ones.

In regard to the two high corrections, looking at the knock logs, they appear in the first (earlier time) few, but not in the last three, where FLKC is 0.00 in essentially the same rpm/engine load combinations as the earlier reports.

The FLKC is very dynamic -- the ECU is constantly testing to see what level of timing works without knock in each "cell" and will move the correction value up or down. As we see in your tables, from one minute to the next, the content of a particular cell can change. Based on where I am now in my quest for answers, I wouldn't worry about it unless those higher levels of negative FLKC show up a lot in normal driving.

Did you change the cats and sensors as part of the HG work? If so, is there a difference in the data that you had recorded earlier? (BtSSM can be used for this, if it has, or you set up, the necessary "fields".)

 

[matt88-8]

I'm glad this Ecu is the same. After skimming your thread I believed it was.
I did change the Cat while doing the head gasket, but ended up leaving the sensors. I can retake the cat related data in torque again. Or romraider or btSsm. (All three work of my odblink mx!!)

The high knock levels where at full throttle, which is something I rarely use.
I do detect a strange rattle at high throttle low rpm with the windows down, but it sounds tinny, like a heat shield I didn't tighten properly.
I will be looking into this, and perhaps checking the knock sensors connection also. Maybe replace it if its cheap. These lil 2.5s don't have power to spare while towing, I hate the thought of losing some to something as silly as timing from a faulty sensor.

Sent from my SM-G903W using Tapatalk

 

[plain OM]

If you're willing, have a look at the cat-related data. (Your other thread started with that P0420.)

Romraider probably displays more variable parameters and switches than the basic complement in BtSSM (although this can be added to with "custom fields") and it's quick to select parameters. (BtSSM on a phone or tablet, however, has the advantage for monitoring/display while driving, and if all the parameters are added, is equal if not more than Romraider.)

Interesting about the odblink mx working with all three. With Romraider, have you tried running it on CAN? Just wondering, because the BtSSM manual suggests an ECU patch is needed for this on the 2007 and I would imagine it's the same for RR. The CAN data is reported to flow faster.

I did notice that all your knock reports were at WOT, so some knock (being corrected) isn't surprising. The rattle at high throttle could be shields. In this regard, there have been suggestions shield rattle can result in the knock sensor picking up the sounds that are then interpreted by the ECU as engine knock! It would be good to verify if the sound is from the shields, especially if the sound is new since the exhaust parts were replaced.

Earlier generations could be fitted with quite inexpensive generic knock sensors from places like eBay, but I haven't seen any for ours, which have a different part number. The Subaru part is over $100, whether from a dealer or, say, Amazon.ca.

 

[matt88-8]

I will get the data again. I will use torque, so all is equal on the recording part. Never tried the CAN side.
BtSsm would be fine if it only had tcm!

Sent from my SM-G903W using Tapatalk

 

[plain OM]

. . . BtSsm would be fine if it only had tcm!

It does. Works with my 4AT. Like for the ECU, I had add additional fields in BtSSM to get everything I wanted that is available from the TCU. Note: BtSSM, RR and even the Subaru Select Monitor, can monitor only one control unit at a time, i.e., either the ECU or the TCU. But switching between the two is fairly simple in both RR and BtSSM.

"To try TCU monitoring, first you need to already have a working profile for your car's ECU (virtually all current users have already done this). Then press "Car", press your current profile, press "New", change the dropdown from "ECU" to "TCU", give it a meaningful name (perhaps with "TCU" in it), and press OK. It should then communicate with your automatic transmission's TCU. You will notice that when you go into "Fields", the list is much smaller and specific to a generic TCU. Otherwise BtSsm should work pretty much the same way regarding gauges, logging, etc." [quoted from the developer, hieche, at http://legacygt.com/forums/showthread.php/btssm-bluetooth-adapter-app-229709.html ]

Might find some of the supplied fields are for the 5AT and don't work with the 4AT. I disabled those and added others.

 

[matt88-8]

best news ever! Now I can watch temps while towing. Perfect. I will run the cat related tests and get back to you

Sent from my SM-G903W using Tapatalk

 

[RoughDiamond]

The FKC is not something to be TERRIBLY concerned about. The ECU's threshold for what it defines as "Knock" is pretty low. Loose heatshields, brackets, stiff suspension, socket rattling around on top of the engine, etc, can be detected. I'll get 30+ "Knock Sum" counts on a typical around town drive with zero FLKC. (this is usually caused by the ECU adding timing with the FKC values)

The number you like to see near/at zero is the Fine Learning Knock Correction (FLKC). This is where the ECU has determined and confirmed knock events at this load range and RPM and will automatically adjust the timing when entering this zone (even if no knock is detected). That -5.00 is not a good number. However, if having recently changed things like head gaskets, valve lash, plugs, exhaust, etc, without resetting the ECU (and therefore the learned knock tables), it will take a while for the values to come back to a reasonable number.

But like plain OM said, I'd set the "knock" notification higher so you don't panic every time it goes off. The same advice goes for the turbo guys and their AccessPorts: put it in the glovebox and leave it there.

 

[matt88-8]

I had the battery out for a week, so I'm sure the Ecu has been reset

Sent from my SM-G903W using Tapatalk

 

[l88m22vette]

I'd consider throwing in 93/91 octane and some Techron system cleaner, and maybe try some black magic: http://www.subaruforester.org/vbulletin/f113/how-idle-re-learn-after-battery-disconnect-154161/

best news ever! Now I can watch temps while towing. Perfect. I will run the cat related tests and get back to you

x2, I was going to get a trans temp gauge but can save the real estate, methinks its time for a tablet install

 

[matt88-8]

That's quite a song and dance to reset a Ecu. But I can give it a try. Thanks for the link. I'm not pleased with the -5.00 But am at odds as to the cause. I will try different gas. Maybe some high octane, but it won't ever have a steady diet of it. Is it possible something is off with the timing? Nothing sounds different.

One off thing I noted was the the left head gasket had been changed in the past. I sent them both out to the machine shop, assuming they would make them the same. Is it possible on bank now has higher compression than the other?

Sent from my SM-G903W using Tapatalk

 

[plain OM]

. . .I'm not pleased with the -5.00 . . . Is it possible something is off with the timing? Nothing sounds different. . . .

Has the -5.00 FLKC (e.g., in the LV table) come back?

 

[matt88-8]

I haven't reset the ecu yet. I've gotta find time to pull the cable

Sent from my SM-G903W using Tapatalk

 

[plain OM]

I haven't reset the ecu yet. I've gotta find time to pull the cable

Sent from my SM-G903W using Tapatalk

Why reset the ECU?

The procedure in the linked Forester thread is to get the ECU to relearn idle settings after it has been reset.

But if it is necessary to reset the ECU, it doesn't require disconnecting any cables. You can use BtSSM. Whenever codes are cleared, the ECU is actually being reset. When the BtSSM code reader is used, even if there's no codes, there should be a button to reset the ECU. Do this with the ignition On, but engine Off. If you then request a LV table, the IAM will have been reset to 0.5, there will be no A/F Learning corrections, and any FLKC cells will be zeroed. Whe the engine is started it will begin to relearn, and after some time the LV table will start showing data depending on the driving conditions.

Incidentally, after resetting the ECU, turn the ignition Off, and leave it there for a couple of minutes. The ECU goes through a shut down routine during the minute or so after the ignition is turned off, and, in my experience, the reset needs that shut down to complete. When the ignition is turned on again, and BtSSM connects, the new LV table can be requested and will be seen to have changed.