Holiday opinions on EDM data

[DVA]

Hi ...

And happy Thanksgiving to everyone!

For those who have seen some of my posts you know I am a big fan of engine data and watching for things the might be a potential future issue. Below is a trace from a flight on an M20J using an JPI EDM830 and uploaded to Savvy’s site. The flight is about an hour long and as you can see the engine was set up in cruise and left that way; steady RPM, MAP and FF and normal CHTs. Notice the oddity on the EGT for Cyl#2 - this pattern has been seen off and on during other flights - however on some flights, #2 is steady and grouped well with the other EGTs. Occasionally, #2 will exhibit a random tendency (with no pattern) to divert from the group, curing itself a few minutes later.  At all times the rest of the engine parameters stay rock steady; nothing is felt or sensed by the pilot on this flight.

What do you think are differential diagnoses for this? Please share details on why you think what you think.

Notes on the above graph (as seen on Savvy Analysis’ site) the top traces are the four EGTs as a group plus FF. The bottom traces are the isolated #2 EGT (for clarity) and MAP. The temperature delta in the #2 swing peaks at ~50dF. CHTs are all in the 380’s and the #2 CHT does not fluctuate with EGT fluctuations.

Is that not a better conversation than what is likely to happen at the Thanksgiving dinner this year?  Ummm hmmm.

[Andy95W]

First thing I would do is swap the #2 probe with another.  And probably clean the fuel injector while you're in there.

[One Whiskey Hotel]

It appears to be periodic in nature... I would examine for some sort of electrical interference on the number 2 EGT, but I believe the more likely cause is a failing probe. Swapping the probe with its neighbor and examining the resulting download would confirm. 

[carusoam]

Nice graphs, DVA...

+1 For Andy's suggestion...

The EGT fluctuations would require FF variations or air flow variations in that cylinder.  Air bubbles leaking into the fuel line/FI might cause something like that...

Expect a high probability of EGT sensor issue.

Starting with the easier, check the probe and its connections.  Then check to see what is in the fuel injector and it's line.

The repetive nature is really interesting.  That is worth checking into...

PP ideas, not a mechanic...

Best regards,

-a-

[ArtVandelay]

IIRC, mike Busch says this is an indication of a failing exhaust valve, I would borescope that cylinder ASAP. The exhaust valves rotate slightly on each revolution causing the periodic movement.

[DVA]

8 minutes ago, teejayevans said:

IIRC, mike Busch says this is an indication of a failing exhaust valve, I would borescope that cylinder ASAP. The exhaust valves rotate slightly on each revolution causing the periodic movement.

Nice call teejayevans!

I was hoping that more people would comment on this, but I guess the holiday got in the way.

I wanted to post this because I don’t think we as a group spend enough time looking for predictive signs of future problems in our aircraft. This is a pattern of a failing exhaust value - caught before it became a safety issue. The only reason this was caught was because we looked at the EDM data regularly. If you’re not looking at your EDM data you’re missing out on a boat-load of important things. I will post the borescope pics if I can get my hands on them.

To further the discussion, why was this never a EGT probe or wiring issue, or an injector based on the tracing?

[Andy95W]

51 minutes ago, teejayevans said:

IIRC, mike Busch says this is an indication of a failing exhaust valve, I would borescope that cylinder ASAP. The exhaust valves rotate slightly on each revolution causing the periodic movement.

Nice call, TJ.

I was expecting to see a more rhythmic and distinctly cyclical indication, like Mike Busch presented in his seminar at Oshkosh:

 

Interesting to see the EGT graph can take different forms for the same cause.

[DVA]

Just now, N1395W said:

I was expecting to see a more rhythmic and distinctly cyclical indication, like Mike Busch presented in his seminar at Oshkosh:

Interesting to see the EGT graph can take different forms for the same cause.

I think Mike and kids at Savvy would agree that it all depends on how fast (or slow or not) the valve rotates and how leaky the value is. Lots of variables affect that such as trueness of the value stem, carbon buildup, friction, luck, etc., Vibration has a mind of its own so don’t count on consistency with this type of diagnosis. 

And in the case of this engine, the failing exhaust value pattern is not consistent with each flight. On some flights there is oscillation, on others none, but if you look back in time on the EDM traces, there is a definite pattern to this cylinder’s exhaust profile.

Also on this engine we see the exhaust temp on this cylinder running cooler and cooler on average over time - about 75dF. Then are flights where the temp moves around substantially for a few minutes then clears up, odd stuff like that was all pointing to a potential problem. Note you would not likely see these variations just by looking at the EDM in flight, you have to look at the traces to see this stuff early. Savvy Analysis makes it easy, but you can use the EDM manufactures’ visualization tools as well. Regardless, you should use something.

Here are pics of the suspect value taken a month before the graph date:

 

This next pic is a bit blurry, but you can see that the seat is pretty crummy:

I’ll try to get up to date pics to see how much worse it has become.

Here is the Good News... Catching this type of stuff early saves thousands of dollars in repair costs and maybe your ass as well.

[Marauder]

I think Mike and kids at Savvy would agree that it all depends on how fast (or slow or not) the valve rotates and how leaky the value is. Lots of variables affect that such as trueness of the value stem, carbon buildup, friction, luck, etc., Vibration has a mind of its own so don’t count on consistency with this type of diagnosis. 
And in the case of this engine, the failing exhaust value pattern is not consistent with each flight. On some flights there is oscillation, on others none, but if you look back in time on the EDM traces, there is a definite pattern to this cylinder’s exhaust profile.

Also on this engine we see the exhaust temp on this cylinder running cooler and cooler on average over time - about 75dF. Then are flights where the temp moves around substantially for a few minutes then clears up, odd stuff like that was all pointing to a potential problem. Note you would not likely see these variations just by looking at the EDM in flight, you have to look at the traces to see this stuff early. Savvy Analysis makes it easy, but you can use the EDM manufactures’ visualization tools as well. Regardless, you should use something.
Here are pics of the suspect value taken a month before the graph date:
 
This next pic is a bit blurry, but you can see that the seat is pretty crummy:

I’ll try to get up to date pics to see how much worse it has become.
Here is the Good News... Catching this type of stuff early saves thousands of dollars in repair costs and maybe your ass as well.

So, what is the fix? Pulling the cylinder and replacing the valve?


Sent from my iPad using Tapatalk

[DVA]

6 hours ago, Marauder said:

So, what is the fix? Pulling the cylinder and replacing the valve?

 

Yes.

[cnoe]

It's hard to say from one pic but it looks to me like that lighter crescent-shaped area is the beginnings of valve failure.

And regarding engine data review, I'm with you brother. I keep my 830 set on 2-second sampling and download after every flight. At least once a week I give the graphs a good going-over.

Good job on the early catch!


Sent from my iPad using Tapatalk

[Marauder]

3 hours ago, DVA said:

Yes.

 

I am really curious what this valve looks like when you pull it. Your pictures look really close to what my valves look like. (I just recently took these pictures)

This is the SavvyAnalysis uploaded flight right after those pictures were taken.

Sent from my iPad using Tapatalk

 

[carusoam]

The prefect circular image is the good...

The crescent moon on one edge is clearly not the pizza we are familiar with...

Chris, I'm not sure what you are seeing, it is circular like the pizza, but has variations of being very clean..?

Is this an artifact of running cleanly LOP?

 

Catching a failing valve before it becomes a bent and or broken valve is a low cost alternative...

Best regards,

-a-

[Marauder]

31 minutes ago, carusoam said:

The prefect circular image is the good...

The crescent moon on one edge is clearly not the pizza we are familiar with...

Chris, I'm not sure what you are seeing, it is circular like the pizza, but has variations of being very clean..?

Is this an artifact of running cleanly LOP?

 

Catching a failing valve before it becomes a bent and or broken valve is a low cost alternative...

Best regards,

-a-

I see the crescent moon artifact on one of the three pictures. He indicated they were pictures of the same valve done one month earlier than the graph. I see the crescent moon on one of the photos and not the rest. Are they different valve photos?

Also, the pictures I posted were done as I was installing the new Tempest plugs and Kelly harness. They look very similar to his non-crescent moon photos. Just trying to understand what other physical indicators are there besides the crescent moon artifact. The crude he mentions is also seen on my seats as well.

[NotarPilot]

What kind of borescope are you guys (DVA and Maurader) using?

[Marauder]

What kind of borescope are you guys (DVA and Maurader) using?

Im using this one: http://www.oasisscientific.com/vividia-va-400-rigid-usb-articulating-borescope-videoscope-inspection-camera.html


Sent from my iPad using Tapatalk

[HRM]

Speaking of MB, have you considered his service where he analyzes all of the data from your monitor?

[DVA]

38 minutes ago, Marauder said:

Im using this one: http://www.oasisscientific.com/vividia-va-400-rigid-usb-articulating-borescope-videoscope-inspection-camera.html
 

 

Did you get (or see) yours at KOSH this year?

I have the same one.

[Marauder]

Did you get (or see) yours at KOSH this year?
I have the same one.

No, I picked it up afterwards.


Sent from my iPad using Tapatalk

[Marauder]

Speaking of MB, have you considered his service where he analyzes all of the data from your monitor?

I have been using it. When I had problems with my spark plugs, I sent the data in and confirmed what was going on. Later I did a GAMI & Mag check with them. And then again after I got the new plugs and harness installed. Nice have another set of eyes on a problem.


Sent from my iPad using Tapatalk

[DVA]

Here’s a great value shot of the progress to damage. Note the margins where the heat profile changes up to the actual failure. Remember, that missing piece of metal had to go somewhere. It might have rattled around inside the cylinder for a bit destroying other stuff, or maybe it was pushed right out the exhaust and overboard before it did any damage. If this comes off in a turbocharged plant, you can virtually guarantee an immediate turbo failure.

 

From an AOPA article a few years ago, here is a valve chart that could come in handy.