The big Q one cannot avoid asking is this: cracks absolutely do happen, so what causes them?
They also happen much more in turbo engines, but most turbo piston engines are not really boosted; they are only “normalised”, so the cause can’t be just straight cylinder pressures, which are no higher than in NA. It also can’t be instantaneous pressure plus instantaneous temperature, because these are the same as NA in the early part of a climb. The explanation has to be a combination of pressure, temperature, and time, in some combination, and that precisely fits the turbo operating profile (SFC to FL180 in 18 mins, etc) and, sure enough, turbo engines develop loads of cracks.
The evidence from glider towing ops is strong, and hard to dispute due to the reasonable sample size.
If someone says shock cooling doesn’t exist, I would challenge them to buy an IO540 or IO-550 engined plane (they will have to buy it because nobody will rent them one 
) climb it at Vx or Vy to say 5000ft, CHTs close to 400F, and then close the throttle and glide back. Repeat a few times, inspect, and report the result.
Also, I bet you lots of people do have cracks but don’t find out, because
This is very similar to cutting open oil filters. A lot of them don’t get cut open – as I very well know. And almost nobody (as a %) does oil analysis.
The bottom line is that not much good data comes out in GA, partly due to things naturally not getting detected and partly due to people keeping quiet. In general the only people who post negative reports are ones who aren’t protecting something, or ones who plan to never sell their plane.
Peter wrote:
They also happen much more in turbo engines, but most turbo piston engines are not really boosted; they are only “normalised”, so the cause can’t be just straight cylinder pressures, which are no higher than in NA.
People who buy a turbo normalized plane use it differently from people who buy an NA plane.
TN planes are typically flown at higher altitudes (so lower environment pressures and lower OATs) and for longer share of the flight at higher %-Power.
In towing ops it’s also that they fly a high percentage of flight time at max power. That also has implications for CHT (both in terms of peak as well as in terms of average CHT).
And that is the whole point: There are obviously aircraft/engines that are more prone to cracks. But for all of these engines, rapid reduction of temperatures is only one of many factors that does differentiate them from others
If you are flying at say 20,000 ft and you pull the throttle to idle, you will start a decent at say 500 ft per minute which leads to a OAT rise, excluding any RAM air effects of 1 degrees per minute. How fast does an engine at idle cool under such conditions? I don’t have an answer but I’m sure many of you mathematicians and physicists will.:)
IME it would be of the order of 200F per minute.
That’s about 3x the engine mfg’s recommendation for max cooling rate.
I routinely see 70F/min on short final, when closing the throttle from what is almost a glide approach, but that is
I get the shock cooling warning from the EDM700 then.
So if idle and a descent rate of 500ft per minute equals engine cooling 3 x manufacturer’s recommendations you therefore need a lower descent rate (eg 150 ft per minute) slowing the aircraft by reducing MP as one descends. Pilot_Dar recommends 1 inch per 1000ft as do many others. That suggests that at circuit height you are between 17" and 20" MP lower than when you started. But how do you calculate your TOD and descent rate based on this, especially as I would assume that all engines/airframes behave differently with regards to engine cooling. Just watching temperatures is a bit hit and miss for IFR flight planning don’t you think?
gallois wrote:
g 150 ft per minute
So that is just 100 minutes for a descent from FL150 to SL. In a 180kt plane that covers 300NM/550km.
Would be “funny” to ask the controller 100km west of Paris if I could start my descent into Frankfurt ;-)
Malibuflyer wrote:
Unfortunately there is a certain tendency with most people, to regard anecdotes that support ones basic belief as facts while as the same time asking people who tell anecdotes that oppose their beliefs for “the real facts”.
I have found that it can be pretty expensive to get all of that experience yourself. I have found myself more at the center of some of this experience than other pilots, and in some cases, understand the cost :(.
As said, often, cracked cylinders are not discovered until a routine inspection. Commonly, the cracks will be found in the casting between a spark plug hole and a valve hole, though cracks can certainly form elsewhere on the cylinder too. For my experience, it is uncommon for a piece of cylinder to actually come off, though I have been told that’s happened (the whole cylinder head came off). The cylinder head is cast aluminum, so does not have the advantage of good structure as wrought metal parts would have, The cylinder head has very irregular material thickness, so the effects of temperature change (aluminum expansion/contraction) are different locally. And, the measurement of temperature (if the plane even has a CHT indicator) is only at one point on the cylinder, (a washer under the spark plug, or specific port), so the temperature information relayed by the probe is only for that local area of the cylinder, and the “averaged” CHT will be presented to the pilot after it has averaged following a change.
So, if the pilot is seeing the CHT move (which I have once, right seat in the T207, when an abusive pilot closed the throttle), that local area is experiencing a rapid temperature change, where other parts of the cylinder may have to catch up – so unequal temperature change, and stress on the local aluminum structure. Some engines have poor or missing baffles, so cooling may be unequal.
I have seen cracks in cylinder heads which can appear as a smoky black line, as exhaust gasses are escaping through the crack. Certainly, once a cylinder is removed and cleaned up (glass bead blasting), cracks become very visible. At the engine shop, we did send cylinders out for weld repair – it was somewhat successful, though not entirely worth the effort. As aftermarket cylinders became available, and OEM’s reduced their prices to compete, simply replacing the cylinder with new was more cost effective and dependable.
I am very leery of accepting at face value remarks like [that operator/display pilot] did XYZ with the plane for years, and never changed a cylinder. Well… as said, people don’t always provide complete maintenance information to those of us in the internet. I am aware (having worked for the engine shop for years), that some well off operators just changed whole engines more regularly that most of us. The operator of the Aztecs I used to fly just kept a spare engine preserved, and ready to install, in case it was ever needed. I remember it sitting not needed for years.
My piston engine experience is not the most, and not little either. Through experience, both as a pilot and maintainer, I have seen some vulnerabilities of certain engine models, particularly some models not common to rental airplanes (so much smaller number of pilots operating any particular one of them). Shock cooling is a vulnerability of many engine models, and I personally treat all air cooled piston engines as though shock cooling damage is a risk.
I guess the only way to definitely settle this dispute would be to buy 20 or so identically equipped planes, randomly assign them into two different groups of 10 and then have one of the groups flown like shock cooling exists and the other like it does not. Then compare the results every X hrs of operation.
MedEwok wrote:
I guess the only way to definitely settle this dispute would be to buy 20 or so identically equipped planes, randomly assign them into two different groups of 10 and then have one of the groups flown like shock cooling exists and the other like it does not. Then compare the results every X hrs of operation.
That will sadly kill this debate forever? but you still need to cross-validate asking each group to fly the other group style, that makes 20 other engines
MedEwok wrote:
randomly assign them into two different groups of 10 and then have one of the groups flown like shock cooling exists and the other like it does not.
The pilots are not allowed to know if they fly the planes with or with our shock cooling (to avoid that they make other things differently as well) ;-))))
