335i boost leak challenge.

[Herbmiester 908]

I bought my 335i last September and so far I have only done a few thousand km's in it. It has a persistent boost leak. So far I have done the following, new intercooler and piping, new vacuum hoses and checked and replaced the boost control solenoids.  I have had smoke tests and vacuum tests done, the computer reset and it still persists. Now I will admit it could be the charge pipe, I have removed it and inspected it and I can't see any cracks but to be safe I have an Evolution Raceworks Charge pipe coming.   I will also replace the OEM plastic diverter valves with some Forge Diverter valves. The OEM diverters are far from tight in the OEM charge pipe. With all that said I think the real culprit might be the leaking valve cover gasket. I had a conversation with a guy on the US E90 forums who went through a similar scenario and the valve cover gasket was the culprit. BTW my car is just about to hit 80,000kms; hopefully one day BMW will sort out their gasket materials. I will post an update next week with the results. I am hoping this is it so I can finally give it an MHD tune and extract all the potential the engine has. 

 

[Matth5 471]

That is a tough one. Good luck.

No chance it's failing turbos/wastegates?

[Herbmiester 908]

I can't say for sure it's not the turbos but two mechanics didn't seem to think so. No waste gate rattle and at 80,000 ks there shouldn't be. Everything I have read and discussed with people here and in the US says charge pipe or gasket. Fingers crossed.

 

 

 

[M3AN 4015]

Since you've had smoke and vacuum tests done I was also thinking wastegates. Might be worth testing them to affirmatively rule them out.

Do you have a mechanical boost gauge?

[BreakMyWindow 1870]

I suspect the arrangement of the Crank case ventilation system requires that the valve cover is leak free. 

[Herbmiester 908]

This is from a post by Rob Beck of RB turbo in the US. I run one of his replacement PCV valves

 

The other "PCV" you reference is a check valve that only limits airflow in low load conditions for the PCV in the valve cover to vacuum up impurities in the crankcase. This is accomplished from vacuum in the intake post throttle blade, through individual channels from each intake port in the cylinder head casting up through the valve cover gasket and into the chambers of the valve cover and finally to the PCV valve. Under high loads the air charge has the same path, but closes the PCV and uses the Bank 2 turbo to help relieve crankcase pressures as you mentioned. But the point is that if the valve cover gasket is leaky or whatever, boost could leak out of it under high loads and it could be a vacuum leak at low loads. The valve cover gasket is the seal for the channels. Personally if I ever pulled one off I would use some RTV Gray or equivalent around each of those 6 channels between the head mating surface and the gasket, but that is just me.

[jin108 25]

You can easily test the wastegates by applying vacuum directly to them and see if they hold. You mentioned the diverter valves are loose that is highly likely to be the issue. Maybe hold off on getting the forge diverter valves you can test whether the oem valves themselves are leaking by pulling them out and applying vacuum to them directly. If they dont leak then its the connection to the oem CP and the new CP should fix it as will have rubber hoses that should seal nicely. 

[Herbmiester 908]

I bought a second hand set of Forge DV,s a while back as the oem ones have a sketchy reputation. I should have mentioned I checked the wastegates with a vacuum pump and they held seemed to be OK but I can't establish the actual seal unless I remove them. It could be the actuator arm becoming sloppy but in theory that doest usually happen at such low kms.

[Herbmiester 908]

The leak has to be fixed so if the fault code doesn't show after its fixed it should be a fairly good indication. If it doesn't solve the problem then at least I have dealt with a pesky maintenance issue. It is possible that it is a combination of issues, Rob Beck did say that the N54 can be a difficult engine to diagnose as there are a number of potential leak points including the valve cover itself which cracks.  

I would say that the Saab experience may or may not have relevance here as its a completely different engine and if I have learnt one thing about tuning engines over the last 30 or so years it is that principles are all well and good but the devil is in the detail.

Edited March 8, 2017 by Herbmiester

[jin108 25]

45 minutes ago, Herbmiester said:

I bought a second hand set of Forge DV,s a while back as the oem ones have a sketchy reputation. I should have mentioned I checked the wastegates with a vacuum pump and they held seemed to be OK but I can't establish the actual seal unless I remove them. It could be the actuator arm becoming sloppy but in theory that doest usually happen at such low kms.

If you used a vacuum gauge should have been pretty clear whether there was a leak or not. If there is then tighten the actuator arm but then it should be apparent if theres a problem as the wastegates would rattle if they were loose. 

Ive never heard of a boost leak caused by the VCG but in cases of boost leaks CP and DVs are the first ones to rule out. 

[polley 916]

So how have be decided there is a boost leak? Is boost low or you have a fault code that indicates a boost leak?

[Herbmiester 908]

The 30ff code shows that there is a difference between actual boost and what the ECU wants to see as boost. Common reasons are , leaking vacuum lines, faulty wastegate solenoids, leaking diverter valves, leaking intercooler, cracked charge pipe and wastegate rattle caused by wear in the wastegate arm pivot. As mentioned I have been through the obvious stuff, two separate workshops have had a go as well and it persists. So now we are down to the less common reasons. And while Ron may have his theories I am inclined to listen to a guy who has specific experience with the N54 platform. The reason I am not convinced it's the charge pipe as it didn't leak on the smoke test and concensus on the US e90 form is that it usually handgrenades when it goes rather than Crack.  However the valve cover cracks and there is a direct correlation between the 30ff code and leaking/cracked valve covers. So we will see what happens when this is fixed.

[Herbmiester 908]

1 hour ago, jin108 said:

If you used a vacuum gauge should have been pretty clear whether there was a leak or not. If there is then tighten the actuator arm but then it should be apparent if theres a problem as the wastegates would rattle if they were loose. 

Ive never heard of a boost leak caused by the VCG but in cases of boost leaks CP and DVs are the first ones to rule out. 

The vacuum gauge showed they closed and there was no leak down over time. If you Google waste gate rattle you will see how the pivot arm wears and even tightening up the arm will not solve the issue. That said at 80ks there should not be enough wear to cause this.

[jin108 25]

2 hours ago, Herbmiester said:

The vacuum gauge showed they closed and there was no leak down over time. If you Google waste gate rattle you will see how the pivot arm wears and even tightening up the arm will not solve the issue. That said at 80ks there should not be enough wear to cause this.

I think many would disagree with you that tightening up the arm would not solve the issue. In my case the rear WG was rattling I tightened it up and no rattle for about a year now. It is the only fix short of replacing them. If you run out of thread to adjust then the only fix is new WGs. It can happen at 80ks if its been driven hard but since you've tested it doesnt sound like thats the issue. Smoke test wont show a vacuum leak on loose DVs if it only happens under boost so its the easy thing to fix before changing VCG. 

[Herbmiester 908]

No wastegate rattle so that's why I am focussing on other issues. I will get someone else to have a look at it if the gasket and charge pipe fixes fail. 

You have probably seen this video but I will post it for those that have not. 

https://www.youtube.com/watch?v=VQD6tITYH7k

[Herbmiester 908]

I went to pick the car up today after the valve cover gasket was replaced. I took the car down the road bringing it up to temp so I could do a pedal to the floor pull and you guessed it, the check engine light came on. But this time it was cylinder 6 injector issue. Back she went to the shop, hopefully its an easy fix.  

[Matth5 471]

Damn you are having the worst luck with yours :/

Hopefully that's the last one. 

[Olaf 3235]

yes!  hopefully final steps before establishing a solid platform!

[Herbmiester 908]

Update, the shop never put on the injector connection correctly. Duh! Anyway the 30FF check engine light issue is improved but not solved. In the past I would get on to a longish straight and from about 50kph floor it, as the revs rose and the boost built it would hit around 4500 rpm and the check light would come on. In the past it would do it every time. Now it does it around 50% of the time, I guess that's an improvement. USPS shows the ERW charge pipe is in NZ and should be here tomorrow, I will get it on asap and fit the original diverter valves, check, then fit the forge diverter valves. The journey to 400 MHD enhanced horsepower is proving to be a long and trying one. This car reminds me of a Redhead I once dated, great to look at, was great when new but went downhill rapidly. 

Edited March 15, 2017 by Herbmiester
punctuation

[KwS 2372]

Do you know how much boost you are actually getting and what the ecu is expecting (and have you confirmed with an independent gauge)? Could it be something as simple as a faulty pressure sensor telling the ECU its seeing less boost than it is?

[Matth5 471]

MHD can log that if you've got the cables, you do have to pay though for the logging feature. About $50 I think. That's how I found my issue, mine would almost never trigger a fault code, MHD devs pointed it out when I shared my logs online...

If you need some free logging, you might be able to make do with Torque, but it won't show Target boost I think. 

[Herbmiester 908]

My cables have just arrived so I will log with MHD good reminder i nearly forgot about logging.  

[Herbmiester 908]

Here is a log of a pull from low revs to 6472 rpm. Green is predicted boost, yellow is actual boost and the grey line is rpm. It's a big leak. 

[Herbmiester 908]

This one is run with the check engine light on, actual and targeted boost are both way down.

[Herbmiester 908]

So the left hand scale is RPM and the right hand is PSI, what I am reading on graph one is a 1psi average difference between what the computer wants and what is actually being delivered. 1 psi may be enough to trigger the 30ff code. I will check this on the yank forums.