How to find an elusive cylinder head crack?

I'm not quite sure I follow what the status of the head is...

This indicates that the head is installed, and the engine is fully assembled.

This indicates that the head is off the engine --- or was at some point for the valves to be removed....correct?

+++++++++++++

When you had the valves out for the dye (penetrant test, I assume?), did you also remove the valve seats?

If you are SURE that it is NOT the head gasket, and you are SURE that it is leaking coolant (NOT oil), your test results point to one possible scenario ---- the crack is hiding in the recess behind a valve seat. This would explain why the dye test did not find it --- if you did not have the seats removed when you performed the test. It would also explain why the pressure test did not induce the leak --- perhaps the valve in the affected seat was fully closed during the static test, and the pressure exerted by the valve was enough to effectively seal the seat against the crack. And, it also explains why the crack leaks while cranking the engine with the starter as the valve would then be open periodically and releasing the pressure against the seat enough to allow the leak.

I've seen a lot of cylinder heads with cracks in the valve seat recesses, but never one where the crack was completely hidden by the seat --- part of the crack was always visible beyond the seat margins. However, I suppose it is possible that you could have a crack which is completely covered by the seat. Or, there could be a small casting defect, such as a hole or a porosity, in that location.

The way I see it is that you have to remove the head to fix the problem anyway --- regardless of where the leak is. So, I wouldn't waste time doing any further "tests" or inspection (with or without a borescope) with the head on the car. I'd pull the head, take out the valves AND the seats, then repeat the dye penetrant test being sure to include the valve seat recesses. If nothing shows, heat up the head ---- toss it in an oven for an hour or so at about 350 F ---- this will help to open up any cracks. Then repeat the dye test --- hopefully it will pick up the "enhanced" crack(s).

 

Ooops, I just now noticed the photos So, my questions about your dye test are moot then

The test was done with the valve seats IN ---- take them OUT, and re-do the test

 

FTR, ultrasonic testing methods are not really applicable for this type of part --- too heavy and too much material thickness. Ultrasound is used effectively on lighter, thinner cross section pieces or laminated materials to detect "air spaces" such as voids (bond line) or porosities.

And, you are correct in what you have heard about X-ray exam of bulky Aluminum pieces --- standard X-ray equipment and methods are not effective. In the Aerospace world, we use CT scanners to exam such parts. Although the frequency is still in the X-ray region, the intensity level is higher, and the ability to take "slices" makes it possible to penetrate bulky pieces.

 

Well, if you are bored and curious......have a look underneath the valve seats sometime. It would be very interesting to determine the failure just for the sake of posterity

Inquiring minds want to know

And, if it turns out to be a casting flaw (I am highly suspicious of that), maybe you can squeeze a little "lemon" juice out of the folks in Alfa's warranty department ?

 

I agree with them that it would be highly unlikely......but, stranger things have happened.

To quote an old college buddy of mine who is now a Pathologist / Medical Examiner, "I've never seen that before, but it doesn't surprise me." ---- he says that often, by the way .

 

If you haven't already tried, and you have an airport nearby, call some of the aviation maintenance shops and ask where they get their NDT done and will do non aviation work. I would recommend looking for a shop that can do Zyglo inspections (http://www.magnaflux.com/ProductOverview/Penetrant/tabid/96/Default.aspx).

Zyglo can detect finer flaws that are sometimes missed with Dyecheck (Spotcheck). And as has been previously mentioned, knocking the valve seats out will allow a thorough inspection in that area as well.

I was in aviation maintenance for quit a while and Zyglo was preferred over Dyecheck for non-ferrous assemblies.

Hope this helps.

 

you may need to etch the surface first to open the crack as well. Especially if you do something like pull a valve seat. Aluminum smears fairly easily and that can hide a crack from the dye penetrant.

In the aerospace world, we frequently have to etch before penetrant inspection especially if we are removing a crack and want to be sure we got it all.

Etching only removes .0015/.002" from the surface but that is enough to expose the crack.

sand or grind, then etch, then dye penetrant inspect.

Typical etchant for aluminum is sodium hydroxide or sodium nitrate mixed with water, you can also add corn starch to thicken it so that it wont run.

Use it luke warm, let it sit for a minute or two and then rinse it with a bunch of water to neutralize it. If you get a big reaction (lots of foaming), its too strong and just rinse it off right away with water.

Normal mix for aerospace etching use is about 6 oz of sodium hydroxide per gallon of water and will etch about 5-15 mils/hour at 100-120 degrees F (.005/.015" deep etch in a hour).

Enough to brighten up the surface a bit is usually enough to allow good penetrant inspection.

Sodium hydroxide is more commonly known as lye. Crystal drano used to be soidum hydoxide as well.

If you do go to an aircraft joint they will know how to do the etching and will have fancy controlled solutions. The above is the garage version.

btw, a dilute solution of lye and water is a good way to remove stains from bare aluminum... it will also wreck any anodizing so dont use it for any aluminum thats colored or clear coated.

 

Davidoloan, that's a nasty story you got there. I can totally sympathise with you and I understand you would like Alfa to st*ff it.

But wouldn't you rather just chuck the old cylinder head down the bin (or through the showroom window of the Alfa dealer ;-), and be done with it? This story proves again that the experience of a car is very much dependent not just on its reliability but so much on the dealer. This is also why I do all car maintenance myself, I just had too many bad experiences. It's extremely difficult to find a good garage that gives you your money's worth.

One way or the other, enjoy the 166, hope it is done being a trouble child. Mine has not been without issues but it's a high mileage example by now and I use it as a car for the bad periods in the year - winter and rain. Apart from the interior heating unit which can leak and the rather weak suspension, they're so much better than you'd expect. Alfa tried to go premium with this model and I think you can feel that in almost every aspect with this car.
Despite Alfa's reputation, people I know with BMWs or Mercs have certainly not had less things wrong with their vehicles. Only the friends with Japanese products seem to have lesser issues... but that's a step I'm not willing to take ;-)

Let us know if you find anything on the cyl head!

 

I stick to my original opinion. Pressurize the coolant passages with air at fairly high pressure and leave it at least overnight in an area with constant temperature. If the pressure is unchanged, I don't believe there is a crack (or a porous casting, which is more likely anyway because cracks aren't this hard to find). I have found areas of porousity in old Alfa heads (which had pretty poor quality castings) using this method. I can't see that the crack (or whatever) is going to close up to the point where there is NO leak over an extended period of time with 50 psi air.

 

Of course. And if you can't prove the fault you can still chuck it through the showroom window.

I love BMWs too, but I tend to find the new ones a bit boring inside, unless you really spec them up. I like the Bangle styled E60 5 series, I think that design ages very nicely. With that car, Bangle got it right imho.
As per the Top Gear video, the interior is by far the best thing about the 166, I fell completely in love with it and have been for the past 5 years. You say it's not a driver's car, but I disagree. The twin spark engine is too small for it to make it a driver's car. Try a 3.0 V6 with a Bilstein chassis, and you'll look different at that statement. It's still fwd of course though.
The E32? For the cash you have to fork out these days for a 1 series you can get one of those AND get it completely restored complete with that 24v engine you wish for!
I have a low-mileage E9 as well that I like very much, compared to most modern cars it's even light and small - that was the other way round when it first came out. Would be good to have a 24v in that as well though, come to think of it.

{smirk} haha, well in my experience that's a bit different ! Until now I've enjoyed the 166 very much, even though the two are in a totally different league. I use it only for winter driving, hauling stuff etc. A real workhorse. For me it has proven to be an excellent daily driver with little issues - despite that it's becoming a high mileage car by now. I'll run it until it becomes cumbersome. May not be far off, but I'll see. Hope she'll last a little longer as I just blew my savings on the 456!