Rotax Piston Meltdown

[rv6ejguy]

Originally Posted by Jolly

I have gathered at least a dozen pictures from detonated Rotax engines from classes I have taken. I would post them, but not sure how to post a picture here. The three top common causes are air leak, someone leaning or playing with carb jetting and bad or low octane fuel. Having a seriously over pitched prop could add to your trouble if the right conditions existed. Detonation can happen fast enough that you won't know it until it's too late which is why you see the aftermath with the detonated damage.

I'd be very interested to see photos of detonation damage to Rotax pistons. Email them to me if you can. They might help us understand more about the ways the piston fails.

[waterboy2110]

Originally Posted by rv6ejguy

Yes, all good points and I think that should be investigated in this case because it's so odd. I've found things like this before too- like paper towel, rags stuck inside water/ oil passages somewhere, core or oil plugs missing etc.

Great comments folks.

The intake O rings and water jacket O Rings were in place and there were no signs of being pinched or damaged. The carb manifold / drip pan assemble was as described by the Vans KAI. As Scott pointed out it would be nearly impossible to get the O Ring on the wrong side of the drip tray (someone did manage to as Scott pointed out). I checked the other day to see how hard it would be to get it wrong when I took the latest pictures. The rubber carb flange groves are much smaller than the O Ring in the manifold and the carb flange appears to have a molded O Ring in it. This part would be discarded at rubber change out I believe. I wanted t have a good look at this since in the illustrated parts book it actually shows the intake O Ring against the carb flange when using Rotax drip pans. It's the same part number in the drawing for either setup so I assume it's a mistake. The way Vans set this interface up is correct IMO and as Scott has said hundreds of these are flying this way.



The compression on this engine was always 80/80 on all cylinders as was last checked at 25 hours TTSN.

I believe I mentioned this before but I'm using Prestone Dex Cool 50/50 which I picked somewhat randomly from the approved list. I believe it says it's good to 129C.

I had a great conversation with Bill Sherlock who I was introduced to by a fellow VAF member a couple days ago. He's a seasoned Rotax Mechanic who works locally in my area. One of the things he's mentioned is that the support for Rotax in the US is very different than overseas (he sounded British). My partner Andre, also a long time Rotax mechanic spent much of his time working on these engines in the United Emirates (very hot and no AvGas) and he's said all along that he is baffled by the lack of interest in this failure. He spoke highly of the Rotax factory folks and said they told him the customer always comes first. He's called the factory for me and they confirmed this would be a warranty issue but it would need to go through the right channels. I haven't exhausted all my options WRT getting around the US distributors yet.

I too would think they would be interested in the failure. I don't know if they know about it or how I would go about confirming that. Andre feels I should reach out to Eric Tucker - the guy in the Bahamas we register our engines with. That's my next step with the factory.

[AndrewR]

Originally Posted by jnmeade

How well can we compare information from a air-cooled engine to one that has air-cooled cylinders and a water cooled head

My point was that we would be concerned if a Lycoming hit 450F on the ground and 480F after takeoff, why are we not concerned about a Rotax similarly close to it's limits?

Is it just smaller numbers due to temps in C and liquid cooling just aren't as scary?

[AndrewR]

Originally Posted by rv6ejguy

I think you'll find the detonation failures will all look pretty much like the photos I've already posted with simple fracture of the 2nd and 3rd ring lands. You won't find heat damage to the crown unless pre-ignition follows the detonation like in the second photo I published.

This looks like textbook detonation damage. I haven't found any other references saying you won't find heat damage from a detonation event - everything seems to say the opposite.

You said your photos were of events that lasted only seconds. What would they look like if the engine continued to run WOT for another minute or 2?

Originally Posted by rv6ejguy

Heavy detonation usually ends with the failure of the ring lands because of the massive compression loss which lowers the cylinder pressure below the point where detonation can occur.

Sure, the detonation ends but what does the loss of compression mean? It means that gas is escaping past the rings. At WOT, a lot of gas.

Originally Posted by rv6ejguy;

the piston then loses gas seal, hot gas flow past the piston then quickly melts the aluminum.

After the ring land breaks, you have a loss of compression and lot of blowby. That chunk of ring land suddenly has a lot more surface area, hot gases blowing past, some nice sharp corners to absorb heat and no path to sink heat back into the rest of the piston. So it will melt. The rings are unsupported at that point so being brittle, they are likely to break. The next part of the piston to melt due to the blowby is probably the crown/top ring land because it also has a lot of surface area, only a small cross section to sink heat into the rest of the piston, and it's in the hottest area anyway.

My opinion: that's what your detonation photos would show if you maintained full power for long enough after the ring land broke. The end result would look a lot like this failure.

[rv6ejguy]

Originally Posted by AndrewR

This looks like textbook detonation damage. I haven't found any other references saying you won't find heat damage from a detonation event - everything seems to say the opposite.

You said your photos were of events that lasted only seconds. What would they look like if the engine continued to run WOT for another minute or 2?



Sure, the detonation ends but what does the loss of compression mean? It means that gas is escaping past the rings. At WOT, a lot of gas.



After the ring land breaks, you have a loss of compression and lot of blowby. That chunk of ring land suddenly has a lot more surface area, hot gases blowing past, some nice sharp corners to absorb heat and no path to sink heat back into the rest of the piston. So it will melt. The rings are unsupported at that point so being brittle, they are likely to break. The next part of the piston to melt due to the blowby is probably the crown/top ring land because it also has a lot of surface area, only a small cross section to sink heat into the rest of the piston, and it's in the hottest area anyway.

My opinion: that's what your detonation photos would show if you maintained full power for long enough after the ring land broke. The end result would look a lot like this failure.

The thing you're missing is that when the ring lands break the engine will effectively go down to about half power on that cylinder and it's immediately obvious due to engine roughness. Break lands on 2 or more cylinders and it turns into a severely vibrating beast. The instinctive reaction in a plane or car is to reduce power immediately, not to leave it WOT.

This is why I mentioned another race failure where the driver noted the miss but continued to try to "drive through the miss" hoping it would clear out. What he did was completely destruct all 4 pistons with severe pre-ignition as well as melt the aluminum around the exhaust seats. This was the most destroyed engine I've ever seen that still had the rods attached to the crank.

Yes, you could be right, with a cast piston if you break the ring lands and leave it at high or full power for a few minutes, you might cause further piston damage. From photo 12, we can't see a broken ring land which is ALWAYS step 1 in a detonation failure on a cast piston. The land usually snaps right off from the piston body. If this had happened, there would be no more heat sink ability of that land to transfer heat to the piston and the lands would have quickly, completely melted away.

The difference between real world experience in a field and reading some articles thinking you understand all facets of this topic should be obvious.

[Jolly]

Rotax E-Learning on detonation prevention

http://www.rotax-owner.com/en/videos...-detonation119

[AndrewR]

Originally Posted by rv6ejguy

The difference between real world experience in a field and reading some articles thinking you understand all facets of this topic should be obvious.

Part of my job for 25 years has been problem and failure investigation, often with multiple experts who can't agree on a cause or are all pointing the finger at each other. I have to research the problem and work with the experts to find a cause and solution.

In this thread we have:

• Rotax via a SL saying that operation in this regime can cause detonation and engine damage
• Vans saying they have demonstrated that it does not cause engine damage
• APS saying this is obviously not detonation because engines can operate for hour with detonation without damage
• Yourself saying it is not detonation because detonation can break the ring lands within seconds, but does not cause heat damage.

So I am not an engine expert, but in other ways I am right at home. I know despite their expertise, experts are rarely right 100% of the time. A large part of the puzzle is researching opinions and evaluation how consistent they are with other information that is available.

[AndrewR]

Originally Posted by rv6ejguy

The thing you're missing is that when the ring lands break the engine will effectively go down to about half power on that cylinder and it's immediately obvious due to engine roughness. Break lands on 2 or more cylinders and it turns into a severely vibrating beast. The instinctive reaction in a plane or car is to reduce power immediately, not to leave it WOT.

If you're at 200' after takeoff your reaction might be different.

We have the EFIS data. We can see that the manifold pressure is 30" from about point 800 through 910 (I think these are seconds).

RPM drops from a peak almost immediately, but severely drops at around point 850, varying around 4100-4500. At this point altitude was somewhere around 200-400 feet. Manifold pressure remained at 30" for perhaps another minute.

My opinion, if you lose 10% rpm at 300' you don't reduce power, you take whatever the engine is still giving you to get safely on the ground. If you reduce power, you might find you lose the rest.

[RV10inOz]

Rotax via a SL saying that operation in this regime can cause detonation and engine damage
Vans saying they have demonstrated that it does not cause engine damage
APS saying this is obviously not detonation because engines can operate for hour with detonation without damage
Yourself saying it is not detonation because detonation can break the ring lands within seconds, but does not cause heat damage.

Andrew,

Detonation is not a definitive thing, I have participated in to a small extent the unleaded avgas certification of G100UL. I have plenty of material which I cannot release publicly but there are some things that George has publicly presented at OSH etc. The take away from this is detonation is a bit like saying weather. The weather could be cool, mild warm, warmer, hot, quite hot, really hot unbearably hot......you get the idea.

Not every cycle will detonate, given there is 20 (or 40+) events per second, and then detonation intensity varies from each event. Now if you really go out of your way, use a very low octane for the CR of the engine, maybe advance the sparks a bit, heat up the oil to 240+, get the IAT up around 120, and cut off the cooling flow to the heads/cylinders and do that with an engine that has a propensity for it, then you might get destructive detonation. The thing is often the plug ceramic fails before anything else as it is fragile, and then mixed in with the badness already in place you add the occasional (not every cycle) a reignited one or two, and bingo, now we have broken ring lands or holes in pistons, whichever fails first.

Here is a cylinder, (I don't have the piston pic) that was sent to a big name shop in Tulsa, for comment. They thought it was quite new and in good shape. It had the following history;
1700 hours in a TNIO520 or 550(can't remember)
20 hours of light detonation testing
3 hours of medium detonation testing
30 minutes of heavy detonation testing.

All done with abusive parameters.

https://scontent.fbne1-1.fna.fbcdn.n...fc&oe=5ACA7704

Now to add another variable to this mix, think about the kind of detonation events, as they are all different. George describes it as like being in a hail storm. Lets assume you are going to get 'X' tons per acre of hail, and you have to leave your car outside in the storm. Do you want the hail to be pea sized or apple sized? It is the mixture of the events and the peak severity of them that matters. None are the same as the next, and of course not every cylinder comes to the same party. And some are born stronger than others.

Hope that helps.

[AndrewR]

Originally Posted by RV10inOz

Detonation is not a definitive thing... The take away from this is detonation is a bit like saying weather. The weather could be cool, mild warm, warmer, hot, quite hot, really hot unbearably hot......you get the idea.

Absolutely. Which is why statements like detonation always breaks the second ring land, and does not cause heat damage seem a bit too absolute to me. Or that engines can run for hours with detonation without damage.

From the Allen W Cline article:
An engine that is making 0.5 HP/in3 or less can sustain moderate levels of detonation without any damage; but an engine that is making 1.5 HP/in3, if it detonates, it will probably be damaged fairly quickly

This makes me a little wary about extrapolating testing done on Lycoming to Rotax.

Another tidbit from someone I know who consults on engine management systems to the major auto makers: He said that this is all understood well enough that they don't use test cells so much anymore, a huge amount of it can be done with computer modelling. The manufacturers know what they are doing. I would treat information from them e.g. the Rotax service letter as reliable.