Can't shake the knock...

[cynner]

My car had been getting fairly persistent knock, always around 0.9 load, 1800-2000rpm, and generally around 0psi or just under. Usually the knock is between 1.4 and 2.8, but sometimes much higher.

I've been troubleshooting for the last couple of months... with much help from Matt (Kido) analysing the logs. So far I have done the following:
– car has been checked out for vaccuum leaks (a couple fixed)
- car has been checked out for exhaust leaks (a couple fixed)
- injectors have been replaced
- fuel pressure has been tested, and fuel pump visually inspected (o-rings re-seated)
- WG duty cycle set to zero for testing
– less timing in tune (made it worse)

None of the above has made any difference.

Anyway, Matt is pretty certain it is the wastegate actuator causing the flap to rattle (phantom knock). I have adjusted the pre-load of the wastegate actuator and this has definitely changed the knock. Now it happens at around 1.5 to 2.0 load, 2800rpm - 3500rpm, 5 to 10psi (not ideal... I'm avoiding boost until I figure out what is going on). The knock is also less than before (usually around 1.4 instead) I guess tightening the wastegate causes it to stay shut longer, and now it is opening at higher boosts. So I suspect maybe the noise happens when the wastegate is opening.

Does that sound like the correct theory? Or would have tightening the preload of the wastegate simply changed the dynamics of the boost tables...

Just trying to work out if I should be spending money on a new wastegate actuator or not

Here's a recent log file of when the knock occurs... http://www.datazap.me/u/cynner/logs-20- ... oom=77-119

Thoughts? Thanks!
Kim

[cynner]

Oh, and the latest LV (with WG set to zero).

[alexeiwoody]

Oh, and the latest LV (with WG set to zero).

1.4 degrees of knock, especially if in varying areas (not always in the same place) - can be just noise and are pretty common.

That LV and your log look fine. Is that FLKC you're logging? It seems the ECU realised it was just noise, as it added the timing back in (the event didn't repeat itself).

My thoughts on the wastegate: trying to be as scientific as possible here...the WG is nowhere near the knock sensor. It would have to be majorly faulty to make enough noise (if it's even possible) - yet your boost curve is smooth and solid.

I can't comment on changing pre-load affecting the knock...was it loose to begin with? It could well affect your boost if you've pre-loaded too far? Have just the right amount of tension and make sure it's not loose.





I actually had a similar issue for a long time, where I would get crappy LV corrections in first trim range (like your 8, mine was 5-15 even after rescaling maf), also had a heap of random knock and it seemed to be affected by the weather (happened more on hot days). Not sure what it was - could've been my turbo inlet pipe not sitting right on the turbo...hard to say. Also my up-pipe had a rattly heat shield at the time. Plenty of suspects.

I think if it's just minor knock and if it's in random places - you should be ok. IAM drops on the other hand, would be something clearly wrong somewhere (but again, could be just noise). Is it real cylinder knock, or is it just noise - that's the question you should find the answer to

All the best.

[dr20t]

From memory you had it confirmed with knock ears that it was real knock?

If that's so, then its not wastegate chatter or other noise

Knock is knock - and audibly via knock ears cannot be mistaken for anything else. Once you've heard knock on knock ears you will know what I mean. It sounds like popping / cracking (think of something being zapped in the microwave). That can't be wastegate chatter if knock ears have been used to say its knock

If your car is knocking, then its something else. If pulling timing makes it worse, then it could be preignition from hot spots in the combustion chamber. Or carbon build up around the valve bowl, spark plug, quench area.

Its mechanical related by the sounds of it, and unfortunately this will need diagnosis by deduction almost

Really feel for you but all the best with it and let us know how it goes

Mick

[cynner]

Yep preload set exactly the same as that video.

My mechanic had heard real knock through knock ears, however Matt suspected that feedback knock correction was picking up some kind of noise, causing it to pull timing, which would then cause pre-detonation (and that was what the mechanic had heard). In the log files I would see -1.4 knock, then it would spiral out of control going up to -11 until I took my foot off the throttle. This was always under light load up a slight hill at atmosphere.

Since then Matt changes the knock sensitivity settings so that FBKC only listens at higher loads, and FLCK now listens at lower loads. Apparently FLCK is more resistant to engine noise or something? This has definitely helped.

Matt has dropped off the radar over the last week unfortuantely (I fear I've exhausted the "free re-tunes" limit)... so I'm at a bit of a dead end as far as troubleshooting or getting a final tune file). I may have to get a road tune by a local tuner to see if he has trouble with knock... (he tunes quite differently to Matt... I think he tweaks less tables - a simper tune with less wizardry, but I think not as torquey and higher boost.

Will keep you guys updated anyway... if it doesn't go above -1.4 maybe i'll just live with it...

[dr20t]

Makes sense about real knock being caused as a result of pulled timing from false knock.

Fbkc and flkc both reference same knock sensor inthe same location though. Flkc isn't more sensitive so to speak but more broad in application.


Just be mindful of trying to tune around what is potentially a real mechanical issue

Mick

[KiDo_Tuning]

Timing has had 6 degrees more and 12 degrees less has seen no change(less timing makes more noise) along with various combinations of AVCS, the video of the WG flap rod in the exhaust housing is an audible clack clack(youtube it up if possible cynner).

It is in closed loop, minimal AF correction or learning present so with AVCS and timing both have no change

I put a local 08 on the dyno today, tapped the exhaust housing with a screw driver... knock sensor went off

[cynner]

Here's a couple of videos...

This is me wobbling the wastegate actuator arm... the clicking you can hear is inside the actuator itself: https://www.youtube.com/watch?v=DnZ7yEvTsL8

Then here is with the actuator arm removed... I can rattle the wastegate flap inside the turbo: https://www.youtube.com/watch?v=r8ur9lFCYrI – I believe this is normal, I've had a look at a couple of other turbos and they also do this...

I have disconnected the actuator from the vacuum line altogether (and clamped shut the hose) and still got the phantom knock – so I'm not sure if that rules anything out?

[bigBADbenny]

One way to get WG pressure is just to connect the comp outlet hose direct to the actuator, eg by folding over & clamping the line to the ebcs. But that would allow it to rattle open and closed on boost.
So how about disconnecting the actuator shaft from the WG lever and just leaving it open?
Then log for knock...?

Also you have the ebcs bracket? With the one you supplied me I eventually did a 360 degree soft mount for the ebcs and zip tied that to the bracket. Previously I could hear the Pierburg rattling away on boost.

Also I went through the engine bay tapping and twanging everything in sight and used ds tape and dense foam rubber to quell any rogue vibration and noises.

Same for all the zorst shields down under

[cynner]

Disconnecting the actuator shaft would make the lever flap about like crazy wouldn't it?

I'll try isolating the boost solenoid even further. I have rubber grommets on it, but maybe need to do even more.

[bigBADbenny]

I forgot to mention the wiring it open part
But yeah I don't remember seeing any actuators with that much movement...
I'll test out my spare...

[alexeiwoody]

Disconnecting the actuator shaft would make the lever flap about like crazy wouldn't it?

I'll try isolating the boost solenoid even further. I have rubber grommets on it, but maybe need to do even more.

Hey Kim, is your actuator arm straight or does it bend near the end? I'll see if I can find my old actuator off a similar turbo. It has a slightly bent arm - hopefully same as yours?

[cynner]

Yep it has a bend in it. Ah awesome, let me know if you find it

[dr20t]

I've actually got one too if alexei can't find his

[cynner]

Have been speaking to some other people too... Michael South said "Put the wg preload back to where it was, knock in high rpm is from turbo back pressure pre turbine, that can blow your engine".


He suspects it is more likely cam gears... suggested "hold 1500 rpm and see if the cam advance is stable or mores more than 1 degs". Will give that a go...

dr20t and Alexi - would the actuators you have suit a blouch td05 18g? I'm told the blouch's have something unusual as far as how their actuators hook up or something?