An Engine Near Miss

If you don't have an engine monitor now, you'll want one after you read this


by Marc Cook

It’s said you learn more out in the halls than in the classroom, which we take as a clever way to suggest that real-world experiences count for a lot more than stale lectures and absorption by rote. One object lesson is worth a thousand texts.

Truisms are hardly universal, but this one applies to the pilot of a certain Piper Navajo and the way he learned about fuel contamination, detonation and preignition, and the true value of an all-cylinder engine monitor. We talked to this pilot shortly after his flight, and have reviewed the data exported from the J.P. Instruments’EDM-800 engine monitor–between the two, it’s possible to create a surprisingly detailed picture of the short flight. Let’s jump ahead a bit: There is no doubt in the pilot’s mind that had he not paid attention to the engine monitor, he’d have cratered a very expensive turbo Lycoming and, quite possibly, have created an inadvertent training opportunity for a single-engine approach.

Problem child
A bit of background. This model of Piper’s venerable Navajo is powered by a Lycoming TIO-540-J2BD, a 350-horsepower turbocharged six-cylinder engine running significant ground boost (which nominally puts out around 43 inches, but, thanks to the system’s density controller, can pump as much as 48 inches of manifold pressure on hot days) at a maximum of 2575 rpm without the soothing balm of an intercooler. Among piston-engine experts, the -J2BD is often regarded as a “problem child,” a powerplant at the stress limits of prevailing technology. Because the TIO-540-J2BD runs with such slim detonation margins, this is the engine used by General Aviation Modifications Inc. (GAMI) to certify its PRISM electronic ignition system for that very reason. In short, in the -J2BD you have an engine that on the best of days is highly stressed.

And in this particular Navajo there’s a further complication. The pilot was informed by line personnel that they had inadvertently given the Piper a dose of Jet A to one of the tanks feeding the left engine. (No other tanks were contaminated.) After suitable mea culpas, they drained the tank, refilled with 100LL and ground-ran the engine to everyone’s satisfaction. Like any good pilot, however, our hero was concerned about undetected contamination and vowed to keep a mighty close watch on the engine gauges during the solo, good-weather flight to check it out while his passengers waited on the ground. He was wise to do so: After the flight, it became clear that the wrong tank had been emptied; this Navajo took off running on a mix of 20-percent Jet A/80-percent 100LL.

Anatomy of an event
Engine start, taxi and runup were normal. According to the pilot, the takeoff began normally, with manifold pressures and fuel flows coming up as expected and comparable left to right. He noted a slight anomaly on the standard TIT gauges–the left engine wasn’t up as high as expected–but both engines seemed to be making good power. He busied himself with transitioning to the climb phase. So far, so good.

But had our pilot been really hawking the JPI, he would have noticed two important parameters diverge from the norm. First, the left engine’s TIT was indeed low–at the same instant the right engine’s TIT was 1440 degrees, the left’s was 1280. That’s a big enough split to notice even on analog gauges, which he did; however, the magnitude of the problem still eluded him. (How many times have we all done this, considering first that the gauge may be wrong before believing there could be a real problem?) The other, and indeed more important, telltale would have been the dramatic upward march of the left engine’s cylinder-head temperatures.

Thirty seconds from the beginning of the takeoff roll, the left engine’s number-three cylinder-head temperature (CHT) had climbed from 310 degrees to 384 (the right engine’s climbed from 297 to 347); 60 seconds later, #3 CHT was at 463 (the right engine’s was 368); and 90 seconds into the flight the number-three cylinder on the left engine reached a scorching 489 degrees (370 for the #3 on the right.) Shortly before the hottest cylinder reached its highest CHT, the pilot, noticing the thermal runaway, reduced power and leaned for a target fuel flow in the climb. (We wouldn’t have done both; more on that later.)

Analyzing the data
Despite the pilot’s leaning from full-rich, CHTs on the left engine started downa result of the power reduction as well as extra cooling air resulting from shallowing the climb. At this point, our hero knew something was definitely wrong, but wasn’t quite sure what.

So what happened; why did the CHTs go whacko? This is a clas sic detonation event, beautifully captured in the data and substantiated by the pilot’s knowledge of the misfueling. (For more on detonation, see sidebar.) And then there’s the other engine’s data: Overlay the data from the left and right engines and there is, at least in retrospect, no question something was very wrong with the Lycoming on the port wing. In this case, the CHTs for all of the left engine’s cylinders shot through the roof, gaining heat with dramatic ease.

But the big question is: If this had happened to you and you were unaware of the fuel contamination, what would you have done? And, more importantly, how quickly would you have reacted?

Step by Step
The first step is, of course, recognition. Keep the engine monitor in your scan and consider locking the display onto the cylinder that typically has the highest CHT. This is a key advantage of a combined analog/digital instrument; you get the big picture–all cylinders are alive via the EGT indications, and they’re all more or less in a line by checking the CHT markers–as well as the detailed information. Had our Navajo pilot been watching any of the left engine’s CHTs, he’d have noticed a dramatic rise and known to do something.

As ever, use the engine monitor’s various readings to corroborate the data. If one CHT is going up, what is that cylinder’s EGT doing? Here you may be surprised at what the readings portend. The classic detonation event not caused by contaminated or below-spec fuel (i.e., putting 80-octane in an engine needing 100LL) results in a dramatic, rapid rise in CHT, with a long-term upward trend in oil temperature, but virtually no change in EGT. It’s been accepted wisdom that EGT will drop during detonation, but recent testing has shown that not to be the case. So, don’t assume that a skyrocketing CHT–even in one cylinder–isn’t detonation just because the EGT hasn’t changed.

However, fuel contamination can change the fuel’s burn rate sufficiently to throw off the absolute EGT or TIT value, but the difference is often subtle enough that you wouldn’t notice it unless you were extremely conscious of your engine’s normal thermal effluvia. (The Navajo’s -J2BDs can be considered an extreme test of this theory.) What’s more, don’t get into the “Oh, it’s just a probe” mindset. Engine monitors have been flying for more than a decade and experience has shown that probes tend to fail open; that is with a zero reading. They don’t fail slightly high, they don’t ramp up and down on their owntrust the probes.

Setting the Alarm
One of the best ways to keep a watchful eye for detonation is to allow the monitor to do it for you. Current analyzers allow you to set limit alarms for most parameters. Resist the temptation to set the analyzer’s CHT warning anywhere near the redline maximum. Over the course of a few flights, determine that highest CHT obtained at the end of the climb segment and set the alarm 10 to 20 degrees above that. (Set it to lesser tolerance if you experience little season-to-season variance in ambient temps.) The idea is to have the limit just above the normal maximum, but low enough to call attention to a detonation-induced runaway in the earliest possible stages. Let’s say you have a detonation event. If your engine’s redline is, say, 450 degrees CHT and you set the alarm for 440, by the time you’ve acknowledged and made corrections, the CHT will likely be above redline and climbing.

Graphic monitors are subject to considerable scale compression in EGT mode to make all the bars fit the screen. As an aid, the JPI employs a normalize mode that levels the bars and increases the resolution from approximately 35 degrees per bar to 10. Use this mode in cruise to make EGT shifts more obvious. Normalize also works in the climb phase for turbocharged aircraft.

Back to our Navajo flight, then. Five minutes after the first detonation event, our pilot felt the engine was still in good health; oil temperature and pressure was fine, the CHTs had come down, all was right with the world. Relaxing a bit, the pilot advanced power for a climb a noticed another rapid increase in CHTs for the left engine. (The CHTs on the right increased as well, but not nearly as dramatically.) Seeing the CHT digits flash toward the expensive zone, our pilot once again pulled the power back.

Beyond Detonation
Good thing he did so quickly because the second incident wasn’t just detonation but a full-house case of preignition. (See sidebar.) How do we know? Look at the EGT data for this portion of the flight and you can see the #3 cylinder’s value drop from its normal position about mid-pack among the other EGTs to become the “coolest” of the six. At the same time, all cylinders’ CHTs are rising. Again, the distinction: detonation normally does not influence EGT while preignition most certainly does. During a detonation event, the bulk of the combustion event takes place in about the same amount of time as a normal event, but during preignition the combustion pressures peak hard and early, leaving most of the gas expansion to happen well before the exhaust valve opens. Most likely, the first detonation event damaged one of the spark plugs in #3, leading to preignition later when engine power–hence cylinder pressures and CHTs–was increased.

In both events, our Navajo pilot had the right basic idea: reduce power. Next–or perhaps first if you’re in the initial climb phase and can’t afford to give away horsepower–be sure the mixture is full rich and back it up with the boost pump. Excess fuel slows the combustion process, reduces chamber temperatures and pressures, and may well stop the detonation event immediately. Can you crater an engine in a matter of seconds? Hard to offer an airtight alibi that you can’t, ever, but this Navajo sustained one good detonation event and two more preignition events yet answered the call. Postflight borescope and compression tests gave it a clean bill of health, although the pilot quite astutely insisted that all of the left engine’s spark plugs be replaced regardless of apparent condition.

One other bit of automation might have assisted our Navajo pilot. In the JPI unit, you can select a maximum EGT differential alarm that alerts you when the highest and lowest absolute EGTs are far apart. Normally this is not a hugely valuable number, as absolute EGT is not an important consideration–only the relationship of the EGTs to each cylinder’s peak and other limits as set by, say, a TIT maximum. However, during the preignition event, when the #3 cylinder dropped from the pack, the differential value climbed from a nominal 60 degrees–that is the hottest EGT was 60 degrees higher than the coldest–to 167 degrees. If you watch the numbers carefully, you’ll know your engine’s typical EGT spread and can set the alarm to something just above. Any large-scale divergence from the norm is reason for further investigation.

Finally, take advantage of the data recording available to you by periodically downloading it and taking a critical look at it. Have the average CHTs been trending up or down? How does the oil temperature track the CHTs? Did one cylinder suddenly drop off line during a trip when you might not have noticed?

As with any part of intelligent flying, look for trends and act immediately to remedy any apparent out-of-bounds indication.

Also With This Article
“Detonation And Preignition”
“Decoding Engine Monitors”

-Marc Cook is a freelance editor and writer. He lives in Long Beach, California.


Please enter your comment!
Please enter your name here