What causes exhaust valve failure

[jaylw314]

7 hours ago, DXB said:

I'm pretty skeptical that an exhaust valve can reach the point of failure without showing clear visual evidence of a hot spot at a rim well in advance. That cannot be said of the SAVVY FEVA signature on the EGT trace, which is often a very late event when failure is imminent [also I believe they check for it for free when you upload data to their site, though they don't publicize this.]

I suppose the hours you fly between annuals makes a big difference - if it's a lot, getting your own borescope and checking every 50-75 hours is a quick, easy, cheap do-it-yourself option that most any owner can avail.

It did strike me as a lot of marketing talk, but then again, I have no basis for comparison...

[kortopates]

Now we have FEVA 2.1 that considers much much more than just the oscillating EGT signifying a valve not properly rotating. Unfortunately not all burnt valve as began with the rotating issues hence FEVA 2.1 which is a totally different approach based on a machine learning model that provides a measure of statistical risk.

 

https://www.savvyaviation.com/savvyanalysis/feva/

 

On LOP, sorry but not every engine has the mixture distribution and or strong enough ignition system to support LOP ops. Nothing wrong with leaning to roughness and then slightly enrichening. You may think your LOP but without a monitor many would learn they're solidly at peak. So best to limit this practice at % power setting less than 65% so there is zero concern for where you set the the mixture. But above that it's wise to set the mixture either rich enough or lean off from peak. nothing complicated about it with a modern engine monitor and a pilot that knows how to use it.

 

 

Sent from my iPhone using Tapatalk

[PT20J]

6 minutes ago, kortopates said:

Now we have FEVA 2.1 that considers much much more than just the oscillating EGT signifying a valve not properly rotating. Unfortunately not all burnt valve as began with the rotating issues hence FEVA 2.1 which is a totally different approach based on a machine learning model that provides a measure of statistical risk.

 

https://www.savvyaviation.com/savvyanalysis/feva/

 

On LOP, sorry but not every engine has the mixture distribution and or strong enough ignition system to support LOP ops. Nothing wrong with leaning to roughness and then slightly enrichening. You may think your LOP but without a monitor many would learn they're solidly at peak. So best to limit this practice at % power setting less than 65% so there is zero concern for where you set the the mixture. But above that it's wise to set the mixture either rich enough or lean off from peak. nothing complicated about it with a modern engine monitor and a pilot that knows how to use it.

 

 

Sent from my iPhone using Tapatalk

Paul, from posts I see in BeechTalk, failed rotators seem to be a thing. Wondering if your data shows that it is more prevalent in Continentals vs Lycomings. 

[A64Pilot]

If you suspect a rotator, they are easy to test, remove the rocker etc and mark the collar of the rotator, using a wooden dowel smack the valve, it should turn evenly and  in one direction every time you smack it. From memory I think clockwise

Smack it too hard and the keepers could fall out, so don’t wail away on it.

‘Sorry if everyone knew this already

[DXB]

On 6/18/2021 at 2:44 PM, kortopates said:

Now we have FEVA 2.1 that considers much much more than just the oscillating EGT signifying a valve not properly rotating. Unfortunately not all burnt valve as began with the rotating issues hence FEVA 2.1 which is a totally different approach based on a machine learning model that provides a measure of statistical risk.

 

https://www.savvyaviation.com/savvyanalysis/feva/

 

On LOP, sorry but not every engine has the mixture distribution and or strong enough ignition system to support LOP ops. Nothing wrong with leaning to roughness and then slightly enrichening. You may think your LOP but without a monitor many would learn they're solidly at peak. So best to limit this practice at % power setting less than 65% so there is zero concern for where you set the the mixture. But above that it's wise to set the mixture either rich enough or lean off from peak. nothing complicated about it with a modern engine monitor and a pilot that knows how to use it.

 

 

Sent from my iPhone using Tapatalk

Machine learning is an interesting approach to extract more reliable prediction of exhaust valve failure. I imagine this FEVA 2.1 algorithm is patent pending - is there any plan to publish the detailed methodology used to create the predictive model?  I suspect it was developed by analyzing EGT traces from lots of planes with known normal valves and from planes with valves at various levels of burning by visual inspection. Inevitably the model will be overfit to this particular "case-control cohort" of planes and perform worse in the real world.  In the graph from the website, I'm curious  what validation cohorts were used and how the numbers on pre test probability (average failure risk ~7%) vs. post test probability (3-5% for 5 cylinders, ~12% for one cylinder) were derived. Even accepting that to be the real world performance of the test, a pre to post test increase of 7% risk to 12% risk doesn't make me want to  borescope that cylinder any sooner than I normally would.  Nor does a decrease of 7% risk to 3-5% risk make me feel comfortable increasing my interval of borescope exam for the other 5 cylinders. 

Based on the limited information available to me, this algorithm doesn't seem to impact my practice of routine borescope exam every 75 hours or so, which I suspect is a sufficient screen to catch a burning exhaust valve in time.  Regardless, I'm curious to see how machine learning analysis of aircraft piston engine monitor data evolves - I'm guessing it will have some real utility in the end.  

Edited June 21, 2021 by DXB

[exM20K]

38 minutes ago, DXB said:

Based on the limited information available to me, this algorithm doesn't seem to impact my practice of routine borescope exam every 75 hours or so, which I suspect is a sufficient screen to catch a burning exhaust valve in time.  Regardless, I'm curious to see how machine learning analysis of aircraft piston engine monitor data evolves - I'm guessing it will have some real utility in the end.  

Ditto.  I spent 20 years building trading systems, and the road to ruin is paved with in-sample data to which the model is perfectly fit.  Out-of-sample (real-life trading) will often show dramatically different result.

Mine came back with "All Hands On Deck" for four of six cylinders.  'Scoped them, and everything is jiggy.  Whew.

I did send them my photos and results so that they can perhaps learn the cause of my false positive, but with small samples, the likelihood of coming up with an over-fit model is very high.  The rotator issue is a single malfunction mode which expresses itself in EGT traces.  The multi-factor model is a whole different animal.

I'm a big fan of what Savvy and @kortopatesdo with respect to analyzing engine monitor data traces, and I'm with @DXB: Cylinders get borescoped every 100 hours and at the annual.  

-dan

[carusoam]

Looks like our engines are no longer the low tech tractor engines they once were…

When supported by a decent AI.

More of what stops valve failures than what causes them…

-a-

[DXB]

7 hours ago, exM20K said:

 

I'm a big fan of what Savvy and @kortopatesdo with respect to analyzing engine monitor data traces, and I'm with @DXB: Cylinders get borescoped every 100 hours and at the annual.  

-dan

Agree - I’ve learned a ton from Mike Busch’s writings and from @kortopateson this site.  And Savvy provides a number of valuable services - just not convinced yet these type of AI data analytics from our engine monitors are ready for prime time.

[kortopates]

On 6/18/2021 at 11:55 AM, PT20J said:

Paul, from posts I see in BeechTalk, failed rotators seem to be a thing. Wondering if your data shows that it is more prevalent in Continentals vs Lycomings. 

Failed rotators are real concern yet easy enough to test. However, when ever we have a client lapping a valve in situ, we always recommend replacing both springs and the rotator at the same time to give it every chance of good rotation after the work. The parts are cheap enough its not woth taking the risk not to replace in our opinion.

I don't think we have sufficient data to statistically say one of the other has more failed rotators. Nor do we always know that the issue.

I have one client with an IO-550 that is an artist at laping valves in situ. He's not an A&P but a past Porshe mechanic and taught at the college and does this for lots of people. At any sign of exhaust valve leakage at all, he goes in and laps the valve on his engine. He does a borescope and compression test at every oil change (3-4 months) and he has over 150% TBO hrs  on his 6 cylinders now and still goin strong. 

[Guest]

2 hours ago, kortopates said:

Failed rotators are real concern yet easy enough to test. However, when ever we have a client lapping a valve in situ, we always recommend replacing both springs and the rotator at the same time to give it every chance of good rotation after the work. The parts are cheap enough its not woth taking the risk not to replace in our opinion.

I don't think we have sufficient data to statistically say one of the other has more failed rotators. Nor do we always know that the issue.

I have one client with an IO-550 that is an artist at laping valves in situ. He's not an A&P but a past Porshe mechanic and taught at the college and does this for lots of people. At any sign of exhaust valve leakage at all, he goes in and laps the valve on his engine. He does a borescope and compression test at every oil change (3-4 months) and he has over 150% TBO hrs  on his 6 cylinders now and still goin strong. 

Care to have him share his process?  Lots of cylinders get pulled for a bad valve.

Clarence

[kortopates]

On 6/22/2021 at 1:29 PM, M20Doc said:

Care to have him share his process?  Lots of cylinders get pulled for a bad valve.

Clarence

Sure, here is an article from one of our guys on valve lapping in situ:

https://www.tennesseeaircraft.net/2011/11/13/exhaust-valve-lapping/

It has some limited pictures, yet explains the process in detail. But the article is very conservative and just shows lapping the valve by hand - which os safest. But most of us need mechanical assistance to fully clean up the valve to seat junction, so here is an additional article that has more pictures and shows how one can safely and carefully attached the valve stem to an electric drill without attaching the drill chuck to the valve stem directly which would definetly cause damage. You'll notice he uses vinyl tubing, while most of us use rubber tubing - rubber should be a bit beefier - but the concept is the same. He also gives a good tip on cleaning the valve & seat after. Scroll down to article 131 on this page:  https://airplaneownermaintenance.com/tag/borescoping/

Edited June 23, 2021 by kortopates

[carusoam]

That has to be a good article…

It references the rope trick and the wobble test….  
 

Two things we hear about around here often…

Best regards,

-a-

[nitpick]

On 6/9/2021 at 9:51 PM, EricJ said:

* Most aircraft engines don't make 100% power for extended periods

* for the most part they're still ancient designs

* My O-360 spends nearly all of its life at 75% power.  It's made it to TBO several times.  2000 hours at an average speed of 120 miles per hour is the equivalent of 240,000 miles

* All reciprocating engines are "ancient designs."  There's no difference in car, boat, airplane or stationary engines in this regard.  They all have crankshafts running in plain bearings with reciprocating pistons and a valve train.  There's no fundamental difference between a car engine and an aircraft engine in terms of design

That said, there's a ton of technology in an aircraft engine.  There's a forged crankshaft that was cast in a vacuum.  You don't see that in automotive applications.  There are sodium-filled exhaust valves.  Again, you don't see that technology in an automotive application.  There are fuel injection systems, dual sets of spark plugs, crankshaft vibration (Sarazin/Chilton) dampers.  Won't find those on an automotive crankshaft

And, by virtue of being high-displacement, low-RPM engines they are remarkably fuel efficient -- especially when compared to automotive engines.  The direct drive versions have no losses associated with reduction gearing, like in an auto engine.  The lower swept area per HP output means less loss to friction, etc.

[tmo]

On 6/18/2021 at 8:44 PM, kortopates said:

You may think your LOP but without a monitor many would learn they're solidly at peak. So best to limit this practice at % power setting less than 65% so there is zero concern for where you set the the mixture.

Is the 65% applicable for turbocharged engines as well?

[EricJ]

34 minutes ago, nitpick said:

* All reciprocating engines are "ancient designs."  There's no difference in car, boat, airplane or stationary engines in this regard.  They all have crankshafts running in plain bearings with reciprocating pistons and a valve train.  There's no fundamental difference between a car engine and an aircraft engine in terms of design

If by "design" you mean "fundamental basic architecture", I might agree, partially.   I think "design" includes the details of clearances, actuation, metallurgy, fuel metering and delivery, sensing and monitoring, ignition, control, timing, etc., etc., in which case I'll stick with my characterization of GA reciprocating aircraft engines as being "ancient".   

[kortopates]

27 minutes ago, tmo said:

Is the 65% applicable for turbocharged engines as well?

Yes

[nitpick]

1 hour ago, EricJ said:

 I think "design" includes the details of clearances, actuation, metallurgy, fuel metering and delivery, sensing and monitoring, ignition, control, timing, etc., etc

Almost all of those address weaknesses of the reciprocating internal combustion engine in the automotive application -- issues such as drivability and emissions.  

Weaknesses that simply do not exist in steady-state, high-power applications such as aircraft application.  They are fine in the automotive application but add only cost and complexity and no value in an aircraft application

If they had relevance and value in the aircraft application, the Porsche PFM engine would not have been the heavy, expensive, unreliable and inefficient failure that it was

[EricJ]

1 hour ago, nitpick said:

Almost all of those address weaknesses of the reciprocating internal combustion engine in the automotive application -- issues such as drivability and emissions.  

Weaknesses that simply do not exist in steady-state, high-power applications such as aircraft application.  They are fine in the automotive application but add only cost and complexity and no value in an aircraft application

If they had relevance and value in the aircraft application, the Porsche PFM engine would not have been the heavy, expensive, unreliable and inefficient failure that it was

I disagree.   Apparently Lycoming does, too, as those are the general directions they take when they do try to modernize a gasoline "aircraft" engine.

https://www.lycoming.com/engines/ie2

[carusoam]

Welcome aboard Nit!

Great first post…

We could always use more engine technology people around here.

Go MS!

Best regards,

-a-

[N201MKTurbo]

18 hours ago, EricJ said:

I disagree.   Apparently Lycoming does, too, as those are the general directions they take when they do try to modernize a gasoline "aircraft" engine.

https://www.lycoming.com/engines/ie2

Will it make more power than our current engines?        Probably not.

Will it use less fuel than our current engines?                  Probably not.

Will it be more expensive to maintain?                               Yes.

Will it cost more?                                                                  Yes.

Will it be more reliable?                                                        Nobody knows.

Will it start easier?                                                                 Yes.

[nitpick]

19 hours ago, EricJ said:

I disagree.   Apparently Lycoming does, too, as those are the general directions they take when they do try to modernize a gasoline "aircraft" engine.

https://www.lycoming.com/engines/ie2

Both Lycoming and Continental developed electronic engine management systems / electronic ignition for their engines in the late 1970s.  Neither could find an OEM willing to put up with the added cost and complexity.  Unison's LASAR system of the mid-1990s was a great system as well that could not find a market

Continental developed liquid-cooling technology for its aircraft engines in the 1980s, at a cost of millions of dollars and could not find a market.  Lycoming and John Deere developed the SCORE rotary aircraft engine in the same period, at a cost of nearly $30 million in today's dollars -- and could not find a market

The ie2 technology is really great and I am very happy to see it on the Tecnam P2012 -- that aircraft may single-handedly bring back the piston-powered commuter airliner market.  But I doubt it will be adopted by the General Aviation OEMs selling personal airplanes to the market (i.e. Piper, Cirrus, Cessna, etc.) because of ROI

[nitpick]

7 hours ago, carusoam said:

Welcome aboard Nit!

Great first post…

We could always use more engine technology people around here.

Go MS!

Best regards,

-a-

Thanks, I've been writing about general aviation engines/technology for a few decades now -- you can check out my early work in back issues of the Aviation Consumer from the 1990s

[Shadrach]

On 6/13/2021 at 5:23 PM, DXB said:

I'm pretty skeptical that an exhaust valve can reach the point of failure without showing clear visual evidence of a hot spot at a rim well in advance. That cannot be said of the SAVVY FEVA signature on the EGT trace, which is often a very late event when failure is imminent [also I believe they check for it for free when you upload data to their site, though they don't publicize this.]

I suppose the hours you fly between annuals makes a big difference - if it's a lot, getting your own borescope and checking every 50-75 hours is a quick, easy, cheap do-it-yourself option that most any owner can avail.

My Dad and Brother made a 3 cylinder approach into Winner, SD after #4 swallowed an exhaust valve. This was in the late 90s; we had yet to install an engine monitor. The plane was less than 20 hours out from annual. AFAIK, no cylinders were borescoped at annual because there was no indication of a problem. Compressions all in the mid-70s and no leaks. The valve snapped at the stem. The engine was 20 years past calendar TBO and getting close to hourly TBO, mid 1700s.

That event still looms large in my mind for two reasons:

#1 To the humans in the cockpit, the failure seemed binary. Everything seemed to be operating just as it should right up into the split second that it turned into a $hit $how.

and

#2 when something like this happens you’re a hostage to whomever’s on the field where you land. There was one A&P on the field. He sourced and installed an overhauled cylinder.  My father departed solo for a test flight and lost a cylinder on downwind after it ingested debris from the previous failure. Made another three cylinder approach...The mechanic refused to except even partial responsibility for neglecting to thoroughly clean the intake. Dad‘s hands were tied and he paid the same mechanic full price to install a second overhauled cylinder. That was a $4000ish hit in the mid 90s and we were left with an engine that we didn’t trust (overhaul and engine monitor install scheduled immediately upon arriving at home base).

I agree that monitoring is the way to go. Engine monitors, borescope whenever the plugs are rotated...etc..etc. Even with good monitoring and preventative maintenance there are still times when things go pear shaped without warning. That being said, I think it happens a lot less to pilots that utilize modern technology to run and monitor their engines optimally.

 

Edited August 6, 2021 by Shadrach

[nitpick]

17 minutes ago, Shadrach said:

I agree that monitoring is the way to go. Engine monitors, borescope whenever the plugs are rotated...etc..etc. Even with good monitoring and preventative maintenance there are still times when things go pear shaped without warning. That being said, I think it happens a lot less to pilots that utilize modern technology to run and monitor their engines optimally.

 

The development of the cheap digital borescope has been a godsend to the industry.  There is absolutely NO reason not to borescope every cylinder and develop a digital photographic record at every 50-hour plug rotation/cleaning.

A simple picture of each piston crown and every exhaust valve face will provide "early warning" of 90% (in my opinion) of cylinder failures, comfortably before the failure would occur

The other godsend is the multi-cylinder engine monitor which, together with a computer and a reasonable visualization program, can provide you with a picture of your last flight and compare that picture to previous flights -- making it easy to spot anomalous EGTs and/or CHTs on an individual cylinder basis, post-flight

[KB4]

Can’t help with causes, but good write up on WHAT TO LOOK FOR an annual is here MooneyFlyer August issue.  
 

https://themooneyflyer.com/issues/2021-AugTMF.pdf