Exhaust valve failure!

[HeyChuck]

The exhaust valve in a cylinder of my Continental TSIO-360SB failed during flight (at 1554 hours SMOH). The engine ran rough and cylinder went cold; landed ASAP. The attached picture of the broken valve was taken after we removed the cylinder. This cylinder had compression 80/80 during the annual inspection 24 hours earlier, with no signs of burning during bore-scope inspections. No obvious CHT or EGT anomalies on my JPI EDM-700 were noticed before the failure, but the data were not being recorded. Has anyone seen a valve fail like this?

[GeeBee]

No cylinder is 80/80. If it is, something is not right.

[kortopates]

You won't be able to notice EGT anomalies in flight, they are much too subtle and only present when the valve is sticking in rotation which is a common cause of the valve developing a hot spot. If you got borescope pictures of your exhaust valves during your last annual 24 hrs ago it would be a good idea to look again for a lack of symmetry in color as well as for any green discoloration at the edge. Going from no signs at all to valve failure in 24hrs is very fast and 80/80 isn't believable.  Here is a poster of pictures of what too look for: 14FN0000-ASI-Valve-Safey-Poster_Final.pdf (aopa.org)

[N201MKTurbo]

I wonder if you had a lifter failure? It looks like the valve was forced lower in the seat than it would normally go. Normally, the lifter reduces the valve lash to zero. In that situation, the valve is just seated when the tappet is on the back side of the cam lobe. This will keep the valve from sinking too deep in its seat. If the lifter collapsed, it would give you about 0.080” of valve lash. This would allow the valve to hammer into the seat instead of smoothly riding  down the cam lobe.

Just a theory. 

[N201MKTurbo]

7 minutes ago, GeeBee said:

No cylinder is 80/80. If it is, something is not right.

I have cylinders that are 80/80

[dzeleski]

Hard to tell from that angle but the whole valve looks tuliped quite badly. Usually that only happens from overheating a valve.

Were all the plugs checked? Any missing pieces out of them? The rings and ring lands all in good shape as well?

[TheAv8r]

It reminds me of an old Dean Martin song: When the valve breaks and dies and is not a pizza pie, that's an overrrhaauuulll... 

[kortopates]

1 minute ago, N201MKTurbo said:

I wonder if you had a lifter failure? It looks like the valve was forced lower in the seat than it would normally go. Normally, the lifter reduces the valve lash to zero. In that situation, the valve is just seated when the tappet is on the back side of the cam lobe. This will keep the valve from sinking too deep in its seat. If the lifter collapsed, it would give you about 0.080” of valve lash. This would allow the valve to hammer into the seat instead of smoothly riding  down the cam lobe.

Just a theory. 

Good idea, but I think that is simply due to a normal amount of valve guide wobble at 1500+ hrs and then a portion of the valve head breaking off. The valves guide/stem fit can produce quite a bit of wobble and not actually lead to valve burning or failure, just higher oil consumption. But also just theory.

[GeeBee]

6 minutes ago, N201MKTurbo said:

I have cylinders that are 80/80

As I said, something is not right. Think about it.

[N201MKTurbo]

The fact is, unless you find a smoking gun, you will never know what caused the problem. Throw in a new valve and guide and get on with life.

[N201MKTurbo]

8 minutes ago, GeeBee said:

As I said, something is not right. Think about it.

Why can’t a cylinder have no leakage?

I had two last time. I have an engine inspection coming up next month. If they are still that good I’ll take a picture for you.

[HeyChuck]

26 minutes ago, GeeBee said:

No cylinder is 80/80. If it is, something is not right.

I was not present when the compression was tested during the annual, but the IA stated "Checked compressions: orifice psi 46, #1-55 #2-80, #3-80, #4-70, #5-79, #6-80."

Cylinders #1 and #4 were subsequently replaced. The valves on cylinders #2 and #6 were replaced about 200 hours previously.

[GeeBee]

15 minutes ago, N201MKTurbo said:

Why can’t a cylinder have no leakage?

I had two last time. I have an engine inspection coming up next month. If they are still that good I’ll take a picture for you.

It means you have no ring end gap. You may read it as 80/80 but it is not.

[HeyChuck]

13 minutes ago, N201MKTurbo said:

The fact is, unless you find a smoking gun, you will never know what caused the problem. Throw in a new valve and guide and get on with life.

Getting on with it.Because I was far from my home base, it was more expeditious to replace the entire cylinder assembly & piston (even though there was no other damage), then fly home (a week later). Fortunately, the IA who inspected the turbocharger found no damage, so the pieces presumably exited through the exhaust pipe.

[HeyChuck]

1 hour ago, dzeleski said:

Hard to tell from that angle but the whole valve looks tuliped quite badly. Usually that only happens from overheating a valve.

Were all the plugs checked? Any missing pieces out of them? The rings and ring lands all in good shape as well?

Yes, the plugs, rings, and piston were all fine. The busted valve is still in the removed cylinder, I may take it out for closer inspection.

[HeyChuck]

1 hour ago, N201MKTurbo said:

I wonder if you had a lifter failure? It looks like the valve was forced lower in the seat than it would normally go. Normally, the lifter reduces the valve lash to zero. In that situation, the valve is just seated when the tappet is on the back side of the cam lobe. This will keep the valve from sinking too deep in its seat. If the lifter collapsed, it would give you about 0.080” of valve lash. This would allow the valve to hammer into the seat instead of smoothly riding  down the cam lobe.

Just a theory. 

That is an interesting possibility. The mechanic visually inspected the cam and lifter while the cylinder was off, but didn't remove the lifter to test integrity. I don't think any anomalous lack of pressure was noticed when the lifters were bled down during cylinder replacement, but maybe nobody was paying much attention to that.

[Schllc]

56 minutes ago, HeyChuck said:

I was not present when the compression was tested during the annual, but the IA stated "Checked compressions: orifice psi 46, #1-55 #2-80, #3-80, #4-70, #5-79, #6-80."

Cylinders #1 and #4 were subsequently replaced. The valves on cylinders #2 and #6 were replaced about 200 hours previously.

why would you change a cylinder at 70# compression?   

Its not a criticism btw, I am genuinely curious.

I'm not sure I would have changed at 55# either without some other info, but I definitely wouldn't have pulled a jug over 70#. 

[GeeBee]

14 minutes ago, HeyChuck said:

That is an interesting possibility. The mechanic visually inspected the cam and lifter while the cylinder was off, but didn't remove the lifter to test integrity. I don't think any anomalous lack of pressure was noticed when the lifters were bled down during cylinder replacement, but maybe nobody was paying much attention to that.

On a Continental, I always change the lifters when I swap a cylinder. They are too cheap not to change out.

[EricJ]

2 hours ago, N201MKTurbo said:

I have cylinders that are 80/80

I've seen that once in a while, too.   Compression tests are not precision operations.

I think it's easier to see that on a Lycoming than a Continental.

[HeyChuck]

38 minutes ago, Schllc said:

why would you change a cylinder at 70# compression?   

Its not a criticism btw, I am genuinely curious.

I'm not sure I would have changed at 55# either without some other info, but I definitely wouldn't have pulled a jug over 70#. 

The cylinder compressions were not the concern. The cylinder with 55/80 compression had a leaking exhaust valve but was not burned. More importantly, there was some modest delamination of the Nickel plating on the cylinder wall, which could have progressed and caused further problems. The inspecting IA thought he also saw some anomalous wear on the wall of the cylinder (with 70/80 compression) during borescope inspection, so that cylinder was also replaced. That was likely unnecessary because the minor pit appeared insignificant upon closer inspection.

[HeyChuck]

44 minutes ago, GeeBee said:

On a Continental, I always change the lifters when I swap a cylinder. They are too cheap not to change out.

I will do that next time. Thanks.

[N201MKTurbo]

9 minutes ago, HeyChuck said:

The cylinder compressions were not the concern. The cylinder with 55/80 compression had a leaking exhaust valve but was not burned. More importantly, there was some modest delamination of the Nickel plating on the cylinder wall, which could have progressed and caused further problems. The inspecting IA thought he also saw some anomalous wear on the wall of the cylinder (with 70/80 compression) during borescope inspection, so that cylinder was also replaced. That was likely unnecessary because the minor pit appeared insignificant upon closer inspection.

One of my new cylinders (Continental Lycoming) had a leaky exhaust valve. It only had about 10 hours on it at the time. I pulled the keepers and springs off it using the rope trick and just spun it with my fingers against its seat, until my fingers got sore. About 20 minutes. It sealed up nice and tight after that and has been fine ever since.

[HeyChuck]

3 hours ago, kortopates said:

Good idea, but I think that is simply due to a normal amount of valve guide wobble at 1500+ hrs and then a portion of the valve head breaking off. The valves guide/stem fit can produce quite a bit of wobble and not actually lead to valve burning or failure, just higher oil consumption. But also just theory.

Oil consumption was normal before the failure. Maybe the valve started sticking some in the guide and was burnt with blowby?

[PT20J]

The valve doesn't look discolored, but it appears to have cracks and a chunk is missing. Maybe a defect in the metal that finally let loose?

[HeyChuck]

7 minutes ago, PT20J said:

The valve doesn't look discolored, but it appears to have cracks and a chunk is missing. Maybe a defect in the metal that finally let loose?

A couple other mechanics I have shown this to had a similar thought. No signs of the classic asymmetric burn pattern that sometimes foretell a failure.