Tip-in knock or exhaust rattle?

[solidONE]

Here is a log I took after (re-re-re)installing a somewhat ill-fitting front pipe (j2 70mm) that doesn't have provisions for the stock doughnut gasket.

http://datazap.me/u/solidone/fp-no-o...data=1-5-11-12

As you can see, a stupid amount of knock at tip-in and upon throttle closure.

Here's another log I took after having really torqued down the catback to front pipe connection.

http://datazap.me/u/solidone/tip-kno...data=1-5-11-12

tip-in knock still present. I checked all the connections for leaks and was not able to locate any.

Now, I recall a long while ago there was some discussion about "phantom knock" on these boards. After having taken these logs, I'm suspicious of the vibrations in the exhaust system upon sudden load increase during low RPM operation being mistaken for engine knock by the ECU. What are your thoughts?

Note, this tune I'm using is a slightly less aggressive tune I dialed in last year with minimal knock on 91 octane with pretty much the same exact hardware save for some different gaskets.

Nearly same tune save for slightly more aggressive base timing and same hardware taken last year:
http://datazap.me/u/solidone/stg1-20...0&data=1-12-14

[steve99]

Its pretty minor rearly.

to check if its due to exhaust rattle/hitting chassis ect pull a degree out of the map in the whole rpm/load area see if the knock disapears, it it does its likely knock if it still their its likely vibration.

might be just crappy fuel

iam is solid at 1 , ecu is not concerned about it, on petrol (especially 91) anything less than 0.7 is pretty insignificant especially if its not dropping iam

notice when you IAT is lower the knock is much less if it is knock, maybe adjust the iat retard tables.

[Wayno]

0.7? Do you mean stupid amount small or stupid amount big?

[solidONE]

Originally Posted by Wayno

0.7? Do you mean stupid amount small or stupid amount big?

Stupid amount relative to my log from last year. Sorry, I fixed the links to the logs in the first post. @steve99

Note the first link was before I re-tightened the FB to CB connection, and the loads were pretty conservative yet it was knocking almost every time I open and close the throttle suddenly. And because of this I could not bring the advance multiplier up to 1. I will try pulling 1 entire degree globally and test again.

[Wayno]

One log per year isn't exactly a comprehensive data set.

[solidONE]

Originally Posted by Wayno

One log per year isn't exactly a comprehensive data set.

Okay... I don't think I said anything about "comprehensive data," but would you like to see all of them? I posted these examples so people dont have to look at multiple logs with similar result. I have about 50-100 in my 2 laptops and maybe 10-20 of different versions of this stage 1 tune on datazap before I went back to stage 1 this time from a stage 2 tune.

But here are all the ones with the exhaust related "knock" starting from 2 weeks ago. I have more in my laptop if you want me to email all of them to you if you want "comprehensive":
http://datazap.me/u/solidone/stg1-fp...=0&data=1-5-12

http://datazap.me/u/solidone/fp-no-o...=0&data=1-5-12

http://datazap.me/u/solidone/stock-e...=0&data=1-5-12

http://datazap.me/u/solidone/freeway...g=0&data=1-5-9

http://datazap.me/u/solidone/stock-b...g=0&data=1-5-9

http://datazap.me/u/solidone/stock-b...g=0&data=1-5-9

http://datazap.me/u/solidone/intake-...g=0&data=1-5-9

http://datazap.me/u/solidone/stock-v...=0&data=1-5-13

[Koa]

Wayno is the man, don't worry about it he was just pointing it out man. Don't need to get "passive aggressive" ;D

[Wayno]

I'm not looking through all those but if I had to guess I'd say there's not enough safety built in the fuel, so as soon as you hit a batch of fuel only slightly worse than usual, it knocks.

So here's an example from one log

3000 rpm
0.8 g/rev

Commanded AFR is logged as 14 and you haven't logged fuel sys status but it looks like it's in CL as there's STFT present.

The openflash tunes target 12.7 in OL and 13.7 in CL in that cell.

You could modify the CL fuel to follow open loop more closely, say around 13 instead of 14.

http://datazap.me/u/solidone/fp-no-o...5156&mark=5032

[solidONE]

Originally Posted by Wayno

I'm not looking through all those but if I had to guess I'd say there's not enough safety built in the fuel, so as soon as you hit a batch of fuel only slightly worse than usual, it knocks.

So here's an example from one log

3000 rpm
0.8 g/rev

Commanded AFR is logged as 14 and you haven't logged fuel sys status but it looks like it's in CL as there's STFT present.

The openflash tunes target 12.7 in OL and 13.7 in CL in that cell.

You could modify the CL fuel to follow open loop more closely, say around 13 instead of 14.

http://datazap.me/u/solidone/fp-no-o...5156&mark=5032

If you look at the logs in the first post, the one before re-tightening the the front pipe registered as much as -2 flkc, then the second log after having re-tightened the hardware it went down to -.65 max.

I don't think this has anything to do with the target fueling since by simply torquing down the hardware a dramatic effect on the amount of knock being registered. Also, this tune never used to have such a problem with tip-in knock when I ran it previously, even on winter blend gas. At most it will register more knock on the top end with +1.0 load with the shittier gas, but never this kind of tip in knock. I've added plenty of transient fuel and ignition retard to prevent that on this tune.

I think steve99's suggestion of pulling 1 degree is a good diagnostic. I will try that and see what happens before I do anything else to the hardware. Although this bring me to question some of the "knock" detected by the ecu, especially tip in knock, that's not truly knock but "phantom knock".


edit:
before re-tightening hardware: http://datazap.me/u/solidone/fp-no-o...5921-5035-5981

after re-tightening hardware: http://datazap.me/u/solidone/tip-kno...4212&mark=1127

[solidONE]

I pulled 1 degree of timing in all the fields on the ignition advance table that had a positive number then took a log.

http://datazap.me/u/solidone/kc-diag...mark=6977-2877

Very strange looking. It actually registered a little more flkc than previously, but only toward the end of the log. I suppose it safe to say that retarding the ignition advance 1 degree did not have too much effect on the amount of flkc being registered compared to previous logs.

I just ordered new 2.75" gaskets for the front pipe and will try readjusting the exhaust pieces to see of that changes anything.

[solidONE]

I reinstalled the stock front pipe then reloaded the original STG1 tune:
http://datazap.me/u/solidone/stock-f...6455-6663-2319

While the amount of tip in FLKC has reduced, there is an increased instance of KC being logged.
@Wayno On what table are you adding CL AFR targets? Both A and B load CL fueling compesation tables?

[steve99]

Originally Posted by solidONE

I reinstalled the stock front pipe then reloaded the original STG1 tune:
http://datazap.me/u/solidone/stock-f...6455-6663-2319

While the amount of tip in FLKC has reduced, there is an increased instance of KC being logged.
@Wayno On what table are you adding CL AFR targets? Both A and B load CL fueling compesation tables?

you need to do both A and B tables, the two tables are their because of the way the ecu software was compiled the same table is referenced twice, its effectivly the same table it just occurrs twice in the rom, happens with some other tables as well

the load limit tables ere same.



This is different to the base timing A and B tables which actually have different functions

The aparant increase in knock correction will be just because the ecu software caught the knock when it was operating in a different mode fine correction or coarse correction. FLKC correction and FBKC correction are using same knock sensor but the ecu is in a different correction mode, it only ever has one mode active at any one time, but in can switch rapidly.

romraider forum has a very good explaination of subaru knock control strategy its quite complex, but rearly just be concerned with the amount of correction and if it caused iam to drop.

Anything under about -0.7 correction is usually insignificant for either type of knock correction especially if its not frequent enough to cause iam to drop.

the ecu runs some tupe of algorythm on the knock corrections looking at magnitude and frequency and duration of the events and decides if it will drop iam.

usually if they are just a spike in the logs rarther than a bucket shape its less significant.

[solidONE]

Originally Posted by steve99

you need to do both A and B tables, the two tables are their because of the way the ecu software was compiled the same table is referenced twice, its effectivly the same table it just occurrs twice in the rom, happens with some other tables as well

the load limit tables ere same.



This is different to the base timing A and B tables which actually have different functions

The aparant increase in knock correction will be just because the ecu software caught the knock when it was operating in a different mode fine correction or coarse correction. FLKC correction and FBKC correction are using same knock sensor but the ecu is in a different correction mode, it only ever has one mode active at any one time, but in can switch rapidly.

romraider forum has a very good explaination of subaru knock control strategy its quite complex, but rearly just be concerned with the amount of correction and if it caused iam to drop.

Anything under about -0.7 correction is usually insignificant for either type of knock correction especially if its not frequent enough to cause iam to drop.

the ecu runs some tupe of algorythm on the knock corrections looking at magnitude and frequency and duration of the events and decides if it will drop iam.

usually if they are just a spike in the logs rarther than a bucket shape its less significant.

Got ya. I'm really just trying to figure out why I'm getting more knock than I was before while running this modified stage 1 tune last year. I suppose the fuel might possibly be worse than last year, though I'm more worried about the hardware such as the exhaust and possibly wear on engine components causing it to be more prone to knocking. The logs I posted in the first post and the last log after reinstalling the factory front pipe seem to point to some exhaust rattle causing or amplifying the effect of tip in knock even though there was no exhaust leaks found.

[solidONE]

Originally Posted by steve99

Its pretty minor rearly.

to check if its due to exhaust rattle/hitting chassis ect pull a degree out of the map in the whole rpm/load area see if the knock disapears, it it does its likely knock if it still their its likely vibration.

might be just crappy fuel

iam is solid at 1 , ecu is not concerned about it, on petrol (especially 91) anything less than 0.7 is pretty insignificant especially if its not dropping iam

notice when you IAT is lower the knock is much less if it is knock, maybe adjust the iat retard tables.

Originally Posted by Wayno

I'm not looking through all those but if I had to guess I'd say there's not enough safety built in the fuel, so as soon as you hit a batch of fuel only slightly worse than usual, it knocks.

So here's an example from one log

3000 rpm
0.8 g/rev

Commanded AFR is logged as 14 and you haven't logged fuel sys status but it looks like it's in CL as there's STFT present.

The openflash tunes target 12.7 in OL and 13.7 in CL in that cell.

You could modify the CL fuel to follow open loop more closely, say around 13 instead of 14.

http://datazap.me/u/solidone/fp-no-o...5156&mark=5032

I checked the original tune I was on and it already retarding as much as 1.5 to 2 degrees at 86* 104* IAT along with tip-in fuel enrichment and transient retard added. I went ahead and reduced the CL fueling targets to see if that has an effect on the knock. No bueno. I fabbed a doughnut gasket adapter for the front pipe and new 2.75" gaskets. will log and update with the original tune from last year and pics later.