The Engine Comes off Again

[gevertex]

Hey all. I had a near miss with my engine this past week. I was doing some training with my instructor for IFR currency (I like to get with an instructor every couple months to make sure I am on top of my game) and at some point during the flight I noticed that #3 EGT was consistently ~100 degrees cooler than the rest. I'll give some more details about this later, but this was different than what I observed on prior flights. EGTs were usually within 50 degrees of each other. I submitted the engine logs to Savvy for analysis. They recommended mixture sweeps be done at WOT with reduced RPM (to limit power) to eliminate induction leaks as a contributing factor. Luckily I didn't do this. I instead flew the airplane 15 minutes to the next airport for it's annual inspection. Once they started the inspection they immediately found 5/8 cylinder studs sheered off on the #3 cylinder. As in #3 was only being held on by 3 studs. I obviously feel really stupid for flying the airplane to the annual at all, as based upon the engine data it's clear this failure was in full swing on my prior flight.

Currently prepping to get the engine sent off for an IRAN.

I'm avoiding placing blame about this until the engine is torn down and inspected. Lots of factors here.

Edited May 24, 2024 by gevertex

[EricJ]

Ouch.   Glad that got found!

Was that cylinder taken off or replaced recently?

[gevertex]

Engine monitor picture. % power shows higher than reality there. POH shows that MAP as a full inch below 65% power. I am probably closer to 60% power in that photo.

[gevertex]

2 hours ago, EricJ said:

Ouch.   Glad that got found!

Was that cylinder taken off or replaced recently?

Not recently, but engine has 1.5 years + 87 hours tach SMOH.

[Greg Ellis]

There have been recent articles on cylinders exiting their engines due to the use of a sealant or gasket maker between the mating surfaces of the cylinder and the engine.  
here is one such article.  I am not saying that is what happened here but I would check for sure and ask the folks doing the annual if they saw anything other than a cylinder base O ring.  Because it could be on your other cylinders.

https://resources.savvyaviation.com/wp-content/uploads/articles_eaa/EAA_2013-04_a-little-dab-ll-do-ya-in.pdf

[hammdo]

How on earth does that happen - over-torqued or fatigued studs? - wow…

-Don

[RoundTwo]

3 minutes ago, hammdo said:

How on earth does that happen - over-torqued or fatigued studs? - wow…

-Don

My understanding is that a loose cylinder will vibrate and subsequently fatigue the studs to failure.

[Fly Boomer]

17 minutes ago, hammdo said:

How on earth does that happen - over-torqued or fatigued studs? - wow…

-Don

As pointed out by @Greg Ellis, sealant allows the cylinder to "wobble" a little when there is 1,000 PSI at the time of combustion, and it eventually destroys the integrity of the studs.  Could have been something else, but this one comes up quite a bit.

[hammdo]

If not gasket material on the o ring, then an improperly torqued nut would be the cause I’d assume. When we did my #4 cylinder, I purchased all new nuts, etc. and the studs were cleaned and oiled by the AP to ensure proper torque. No gasket material on the o-ring. I wanted to be certain new hardware was used and studs cleaned properly, oiled, and torqued. My AP totally agreed which really helps…

That picture is scary to look at… glad nothing happened…

-Don

[IvanP]

Had the same thing happen on my E (IO360) about 15 years ago on our way back from Mexico. Subsequent inspection also found a cracked case. My engine was about mid-time so I ended up replacing it. Not a pleasant experience.  

[gevertex]

1 hour ago, IvanP said:

Had the same thing happen on my E (IO360) about 15 years ago on our way back from Mexico. Subsequent inspection also found a cracked case. My engine was about mid-time so I ended up replacing it. Not a pleasant experience.  

We will find out when they inspect it. They already told me I will need a new crankcase based on the number of fixes mine has had. 3x total and 3800 hours on the case. Luckily my chosen shop has as new one in stock. It's an IO-360-A3B6D.

[gevertex]

1 hour ago, hammdo said:

If not gasket material on the o ring, then an improperly torqued nut would be the cause I’d assume. When we did my #4 cylinder, I purchased all new nuts, etc. and the studs were cleaned and oiled by the AP to ensure proper torque. No gasket material on the o-ring. I wanted to be certain new hardware was used and studs cleaned properly, oiled, and torqued. My AP totally agreed which really helps…

That picture is scary to look at… glad nothing happened…

-Don

Needless to say I was a little shook up by it. There is someone watching over me up there for sure.

[gevertex]

2 hours ago, Greg Ellis said:

There have been recent articles on cylinders exiting their engines due to the use of a sealant or gasket maker between the mating surfaces of the cylinder and the engine.  
here is one such article.  I am not saying that is what happened here but I would check for sure and ask the folks doing the annual if they saw anything other than a cylinder base O ring.  Because it could be on your other cylinders.

https://resources.savvyaviation.com/wp-content/uploads/articles_eaa/EAA_2013-04_a-little-dab-ll-do-ya-in.pdf

I heard about that, really unfortunate. I hope thats not the case here.

[Jsno]

Even paint under the nut can cause failure.  I cover the area around the holes when I paint, and if the cylinder comes with paint around the holes I strip it in the area prior to installing the nuts to get a proper torque that will hold. The area around your cylinder bolt hole looks like it had paint.

[gevertex]

15 minutes ago, Jsno said:

Even paint under the nut can cause failure.  I cover the area around the holes when I paint, and if the cylinder comes with paint around the holes I strip it in the area prior to installing the nuts to get a proper torque that will hold. The area around your cylinder bolt hole looks like it had paint.

I think you are right. Might even have more than one coat of paint. You can see in this photo it looks like one layer is flaking off and another is still under it. Perhaps a new coat was even applied over a layer of flaking paint. 

[PT20J]

Anything that relieves the preload on the studs can cause this.

What’s more interesting is how this is related to the EGT drop. Or, perhaps, the two events have unrelated causes. It would be valuable to learn the details. 

[gevertex]

1 hour ago, PT20J said:

Anything that relieves the preload on the studs can cause this.

What’s more interesting is how this is related to the EGT drop. Or, perhaps, the two events have unrelated causes. It would be valuable to learn the details. 

Given it's #3 with the EGTs and #3 with the studs, I'd wager they are related causes. But perhaps one caused the other.

[gevertex]

Another thing my mechanic mentioned was the color of the broken studs. He noted some of them were darker than others (you can see this in the picture) indicating one failed first followed by the rest over a longer period of time. He thought likely months. If that is true then this didn't happen all at once.

[PT20J]

1 hour ago, gevertex said:

Another thing my mechanic mentioned was the color of the broken studs. He noted some of them were darker than others (you can see this in the picture) indicating one failed first followed by the rest over a longer period of time. He thought likely months. If that is true then this didn't happen all at once.

That’s what I would expect. Each one that fails puts Increased load on the rest and the next weakest fails and on it goes.

Since the cylinder was still attached it’s hard to see how the failed studs could affect EGT. I’m not saying the two aren’t related, but it’s hard to see how.

[Pinecone]

9 hours ago, gevertex said:

Another thing my mechanic mentioned was the color of the broken studs. He noted some of them were darker than others (you can see this in the picture) indicating one failed first followed by the rest over a longer period of time. He thought likely months. If that is true then this didn't happen all at once.

I was going to say, the 3rd picture shows a stud that was about half gone before it finally broken.  Possibly the same for one in the first picture, or it may be some contamination.

Where the metal is frosted looking, that is a fresh break

[gevertex]

On 5/24/2024 at 9:55 PM, PT20J said:

Anything that relieves the preload on the studs can cause this.

What’s more interesting is how this is related to the EGT drop. Or, perhaps, the two events have unrelated causes. It would be valuable to learn the details. 

Not entirely sure on this one. My guess is the lack of bolts changing spacing between something critical while it’s running. If the cylinder pulled away from the case a couple mm that could be a couple mm less the valves open. I don’t know enough about this engine to know if that would happen.

Also, we found some metal in the oil filter. Looks like aluminum. Oil analysis confirmed aluminum doubled up. 
 

[EricJ]

2 minutes ago, gevertex said:

Not entirely sure on this one. My guess is the lack of bolts changing spacing between something critical while it’s running. If the cylinder pulled away from the case a couple mm that could be a couple mm less the valves open. I don’t know enough about this engine to know if that would happen.

Also, we found some metal in the oil filter. Looks like aluminum. Oil analysis confirmed aluminum doubled up. 
 

Yes, the valves opening less and a little less compression will both contribute to less output power.

Are you worried about those big flakes?  That's not an unusual amount of flakes in the filter unless there were a lot more scattered around in the rest of the media.    You do expect to see some in there from normal wear.

[gevertex]

1 hour ago, EricJ said:

Yes, the valves opening less and a little less compression will both contribute to less output power.

Are you worried about those big flakes?  That's not an unusual amount of flakes in the filter unless there were a lot more scattered around in the rest of the media.    You do expect to see some in there from normal wear.

More the oil analysis showing aluminum doubling up. I also hadn’t seen flakes like that in prior oil changes. So it’s different at least. 

[PT20J]

Bolts are springs. They work by preloading them so that they stretch within the elastic limit of the material and the spring force thus created is what holds the parts together. The design requirement is that the preload must exceed the force on the stud under load. In this way, the stud doesn't feel the load and the joint stays tight. It is important that the materials being joined have less elasticity than the bolt. This is why paint under the nut or sealant between the joint is problematic. It is also why lock washers are not used under cylinder nuts. The first thread from the shank is the most heavily loaded, and that's usually where the stud fails.

There are two primary failure modes:

1. The nut was installed with insufficient preload (torque), or lost it's preload, which allows the stud to feel the load and eventually fail due to fatigue. This causes a pretty clean break.

2. The nut was overtightened which causes the stud to stretch beyond its elastic limit and neck down leading to a tension stress failure. 

From the pictures, it looks like 1 to me. 

If the cylinder had actually become loose enough to move a mm or two, the studs would have all stretched under the load and failed under tension and the cylinder would likely have departed the crankcase.

[PT20J]

1 hour ago, gevertex said:

More the oil analysis showing aluminum doubling up. I also hadn’t seen flakes like that in prior oil changes. So it’s different at least. 

Oil analysis and filter examination detect different issues, and that's why we do both. Chips in the filter are often too large to show up in oil analysis and suspended metals in the oil are too small to be trapped in the filter media. That metal in the filter is not excessive and may or may not be related. The significant increase in aluminum in the oil analysis means something. Aluminum usually comes from pistons or piston pin plugs, but in this case it might be do to some fretting of the cylinder against the crankcase. While the cylinder is off will be a good time to check the camshaft.

I would carefully preserve any evidence if I suspected that this failure was due to an installation error.