heinzboehmer's 2002 Topaz 6MT Coupe

[George Hill]

Never seen a bad ICV myself, I have seen a bad aftermarket one though, fixed with a used genuine one.

Keep us posted if it was the ICV.

[Obioban]

The valve part of my ICV gets a soak in carb cleaner any time I have the intake manifold off.

[heinzboehmer]

Got an unexpected EML light and limp mode yesterday. Here's a parsed down version of the DTCs I pulled with karter16's excellent tool:

• Active Codes
  • Error 0x88 - Idle-speed controller
    • Frequency: 1
    • Description: The DME was unable to regulate the engine to the specified idle-speed target. If no electrical fault for the idle actuator/ZWD is present, possible causes include contamination of the ZWD, blocked or disconnected hoses around the ZWD/idle-air distributor, or a throttle valve stuck open.
    • First Occurrence
      • Engine Speed: 1920 1/min
      • Relative Filling: 9.6 %
      • Idle Speed Controller Target Speed: 832 1/min
      • Adaptation Value For Loss Torque: 35.6 Nm
  • Error 0x90 - Lambda controller, bank 1
    • Frequency: 2
    • Description: The lambda control has exceeded its plausibility limits. If no electrical faults are stored, possible causes include throttle valves not closing correctly, unmetered air, intake leaks, idle-control actuator mechanical faults, or leaking/sticking fuel injectors.
    • First Occurrence
      • Engine Speed: 3040 1/min
      • Relative Filling: 25.6 %
      • Lambda Control Factor Bank 1: 0.695
      • Adaptation Factor (Multiplicative) Bank 1: 1
    • Second Occurrence
      • Engine Speed: 2320 1/min
      • Relative Filling: 23.2 %
      • Lambda Control Factor Bank 1: 0.695
      • Adaptation Factor (Multiplicative) Bank 1: 0.992
  • Error 0x91 - Lambda controller, bank 2
    • Frequency: 2
    • Description: The lambda control has exceeded its plausibility limits. If no electrical faults are stored, possible causes include throttle valves not closing correctly, unmetered air, intake leaks, idle-control actuator mechanical faults, or leaking/sticking fuel injectors.
    • First Occurrence
      • Engine Speed: 2360 1/min
      • Relative Filling: 36 %
      • Lambda Control Factor Bank 2: 0.695
      • Adaptation Factor (Multiplicative) Bank 2: 0.969
    • Second Occurrence
      • Engine Speed: 2880 1/min
      • Relative Filling: 27.2 %
      • Lambda Control Factor Bank 2: 0.695
      • Adaptation Factor (Multiplicative) Bank 2: 0.969
• Shadow Codes
  • Error 0x78 - Throttle valve potentiometer comparison
    • Frequency: 0
    • Description: Electrical fault present. Electronic throttle limp-home mode is active.
    • (No occurrences)
  • Error 0xDF - Intake manifold pressure
    • Frequency: 0
    • (No occurrences)

Active codes make me think something is up with the ICV. The rubber hoses in the idle path are all brand new and the valve itself is clean, but it could be failing electrically. The shadow codes are tripping me out though. Frequencies are 0, yet they're still reported and triggered limp mode? TPSs are also all pretty new (maybe 10k mi old?).

Cleared codes, drove the car around and only saw 0x88 come back (no limp mode). For now, gonna treat the shadow codes as red herrings and focus on the ICV.

On a related note, I've noticed that I've been much more clunky when pulling away from a stop lately. I thought I was just getting dumber and forgetting how to drive stick, but now I'm starting to think this is a symptom of a failing ICV. I do realize there is heavy confirmation bias involved in that last bit, but fortunately there's an easy way to prove out the theory.

Genuine part is on its way and will get swapped in soon. Price on it wasn't too crazy, so decided to just throw parts at the problem instead of spending a bunch of time diagnosing further. ICV lives right in one of the hottest parts of the engine bay and has been there for 24 years/150k mi at this point. Don't think it hurts to refresh it.

Once the new part is validated to work correctly, I'll take the old one apart to see what I find inside.

[heinzboehmer]

I will be doing this with all my washers from now on.

Yep! I think I will too.

[maupineda]

Coming back to this.

Have been driving the car around and keeping an eye on this leak. I've yet to see it leak any more and have not smelled any oil since I first noticed it.

I only noticed the smell after a track day, so I'm starting to suspect that the second half of the last sentence in the quote above is incorrect.

I'm guessing that what actually happened was that the tensioner had a small leak and the oil got flung around due to the g forces on track. Some made it onto (or close) to the exhaust manifolds (hence smell), some made it onto the tensioner boss on the head and some made it all the way to the water pump pulley (hence the oil trail near the belts).

So, plan is to replace the tensioner crush washer with an annealed one (thanks maupineda for the suggestion!) and go from there. Maybe this will end up being an easy fix?

This does really work, I was having a persistent leak when I replace the rack and annealed the washer and now is bone dry. I for sake of it, test it and you can bend them with the fingers after annealing so it does make the metal soft for a nice and easy ‘crush’

p.S. thanks for putting data to this! Spot on! The annealing also relieves stresses and at a much higher temp they will see in operation, so they won’t ‘anneal’ further which will prevent leaks overtime. My tensioner has been dry for the past two years.

you can also run a candle flame and the carbon residue goes off as you anneasl so you can use that as a gage, but seems the sharpie method is best. They will turn a bit orange and melt quickly after certain melt point is reached. I will be doing this with all my washers from now on.

[Ramage]

I'm not sure I would torque when warm. Wouldn't want to risk overtorquing and pulling out the soft aluminum threads in the head when everything cools back down. Retorque after a full hot-cool cycle (or multiple) seems reasonable though.

Good call, the thread engagement depth of the tensioner is somewhere between 14mm and 17mm making it already a risk of stripping. Edited my comment.

I've always kept the heat shields installed on my car, but I bet that taking them off doesn't help with this issue.

Adding to this theory is that the last time you smelled oil and saw seepage was after a track day which even with the heat shield in place would bring the temps around the tensioner up to extreme levels.

[heinzboehmer]

It might be worth adding a step to bring the car fully up to temp and then retorquing it to spec while warm, just to see if it settled. Only issue with this is that area is around the temperature of the 9th level of hell after the car is at temp.

I'm not sure I would torque when warm. Wouldn't want to risk overtorquing and pulling out the soft aluminum threads in the head when everything cools back down. Retorque after a full hot-cool cycle (or multiple) seems reasonable though.

Another thing your post made me think of, does the Exhaust Manifold Heat Shield (11627831352) make enough of a difference on cars that still have it installed and do not have this leak? I know for myself I only started having issues with the timing chain tensioner after I installed catless headers and did not reinstall the heat shield. Likely this is just a coincidence, but figured I would add another point of data.

I've always kept the heat shields installed on my car, but I bet that taking them off doesn't help with this issue.

[Ramage]

Great write-up!!

The hours I have spent fighting with this stupid leak... I do think you might be on to something about the engine's heat gradually "annealing" the washer. It might be worth adding a step to bring the car fully up to temp (maybe and few times) and then retorquing it to spec after it cool, just to see if it settled. Edited.

Another thing your post made me think of, does the Exhaust Manifold Heat Shield (11627831352) make enough of a difference on cars that still have it installed and do not have this leak? I know for myself I only started having issues with the timing chain tensioner after I installed catless headers and did not reinstall the heat shield. Likely this is just a coincidence, but figured I would add another point of data.

[heinzboehmer]

Swapped out the crush washer for an annealed one and did a bit of science while I was at it. I had 15+ of these things in the stash, so more than enough to mess around with.

First, let's document the annealing process.

I hung the crush washer from a bent paper clip and used a propane torch to heat it up (was afraid MAPP would pretty much immediately melt it), then quickly quenched it in some room temp water.

My first attempt involved trying to judge temperature by eye and resulted in this:



So, for my next attempt, I scribbled with a sharpie on one side of the part (as can be seen above) and hit it with the torch from the non-sharpied side. Once all the sharpie disappeared (i.e. part had heated all the way through), I quenched it in the water.

Supposedly, sharpie ink disappears right around the annealing temp for aluminum, which is why it can be used as a temp indicator. Can't remember where I learned this trick, but thank you brain for stashing that bit of info away.

This method worked significantly more reliably.

Subjectively, a stock crush washer felt harder than the annealed one when scratched with a metal pick. The annealed one also felt way more "clay-like" when torquing the tensioner, almost as if it was "smearing" more. The difference was subtle and it's hard to draw any real conclusions from a subjective account, so, I decided to collect some data.

Here's an overview of the experiment:

1. Measure thickness of crush washer pre-torquing.
2. Torque tensioner along with above crush washer on spare engine to factory spec. Spare engine is used here so that torque wrench is on straight and the procedure is repeatable.
3. Measure thickness of crush washer post-torquing.

I also found a crush washer stuck on the head of the spare engine, so I measured that one + the leaky one that came off the car for some extra data.

Some givens/assumptions before diving into the data:

• I assessed four different "versions" of the same part:
  • A new stock crush washer (1)
  • A new crush washer that I annealed (2)
  • The crush washer that came off the car (3)
  • The crush washer that I found on the spare engine (4)
• (1), (2) and (3) were all purchased at the same time. I placed an order for ~20 a couple years back, so it's highly likely that they were all produced at the same time, in the same factory, from the same material, etc.
• (4) has a completely unknown history. It's highly unlikely that it's from the same batch as the others.
• (3) was torqued by me following the BMW spec.
• (4) was torqued by someone else to some unknown value.
• Pre-torque thickness of both (3) and (4) are unknown. Since (3) is likely from the same batch as (1) and (2), we can assume that its pre-torque thickness was similar to that of those.
• For max scienceness, I prepared three samples of (1) and three samples of (2), then ran the experiment on all of them.

Data:




From this, we learn the following:

• (1) crushes by a small, but very consistent amount.
• (2) crushes by a larger, less consistent amount. I suspect the lack of consistency comes from the annealing process. Doing so with a torch is not super repeatable, which likely leaves the hardness all over the place. Nonetheless, they still crush more than (1), implying they're softer.
• The post-torque thickness of both (3) and (4) are closer to (2) than to (1), even though they were not annealed prior to torquing.

The first two observations confirm that the annealing works, but I think that the last observation is by far the most interesting one.

The fact that the used crush washers are closer to the annealed ones makes me wonder if this part is gradually annealing from engine heat cycles, causing the joint to lose clamping force over time. This is a bit of a logical leap, since we don't have the pre-torque data, but this could explain why this part seems to develop a leak over time, rather than leaking immediately. I'm for sure gonna be checking the torque on the tensioner after some time.

Here's a post-torque comparison of an annealed (left) vs stock (right) crush washer, as well. It's subtle, but you can tell that the annealed part crushed more:



Super long post for a super simple job, but I thought this was pretty interesting!

[heinzboehmer]

Hopefully you will be luckier than I was. Wet in the same spots, but my leak turned out to be the VANOS gasket.

I hope so too!

I've swapped that gasket out too many times now chasing leaks. Think what finally solved it for me was a timesert on the upper right corner of the head. Guessing the original threads were slowly failing and causing the bolt to lose clamping force.

I just poked around there with a mirror and it looks dry. Seems like that gasket isn't the cause of the leak, but it's fairly hard to tell without removing more stuff.

[D-O]

Coming back to this.

Have been driving the car around and keeping an eye on this leak. I've yet to see it leak any more and have not smelled any oil since I first noticed it.

I only noticed the smell after a track day, so I'm starting to suspect that the second half of the last sentence in the quote above is incorrect.

I'm guessing that what actually happened was that the tensioner had a small leak and the oil got flung around due to the g forces on track. Some made it onto (or close) to the exhaust manifolds (hence smell), some made it onto the tensioner boss on the head and some made it all the way to the water pump pulley (hence the oil trail near the belts).

So, plan is to replace the tensioner crush washer with an annealed one (thanks maupineda for the suggestion!) and go from there. Maybe this will end up being an easy fix?

Hopefully you will be luckier than I was. Wet in the same spots, but my leak turned out to be the VANOS gasket.

[heinzboehmer]

Unfortunately, when parking back in said garage, I noticed a faint oil smell. Popped the hood and immediately saw this:


...

Timing chain tensioner is drenched:



The boss on the head that it threads into is drenched as well though, so it seems like the oil is coming from higher up.

Coming back to this.

Have been driving the car around and keeping an eye on this leak. I've yet to see it leak any more and have not smelled any oil since I first noticed it.

I only noticed the smell after a track day, so I'm starting to suspect that the second half of the last sentence in the quote above is incorrect.

I'm guessing that what actually happened was that the tensioner had a small leak and the oil got flung around due to the g forces on track. Some made it onto (or close) to the exhaust manifolds (hence smell), some made it onto the tensioner boss on the head and some made it all the way to the water pump pulley (hence the oil trail near the belts).

So, plan is to replace the tensioner crush washer with an annealed one (thanks maupineda for the suggestion!) and go from there. Maybe this will end up being an easy fix?

[George Hill]

My Z4 braces just showed up

And the brace looks sooo good, but that snorkel being cerakoted vs CF is the chefs kiss IMO.

[heinzboehmer]

Both brace and snorkel turned out amazing with the cerakote!

The snorkel might need a sticker delete

Yeah, sticker is pretty unsightly. I asked for it to get masked cause it has manufacturing info on it, buut all of that came off when unmasking.

Which means that if I do remove it, I have to figure out how to blend the overspray with the raw carbon under the sticker. Soooo I've just been ignoring it.

[Slideways]

Both brace and snorkel turned out amazing with the cerakote!

The snorkel might need a sticker delete