The maddening aspect of drawing conclusions from GA accident reports is their lack of reliable detail. The NTSB and FAA are satisfied to close out a wreck probe with engine failed in flight; post-accident examination revealed no mechanical discrepancies.
Not that we blame the poor pukes doing the field investigations. Theyre understaffed, have few technical resources and too much tin to kick around during even a slow month. Moreover, few investigators are trained in the finer points of piston-engine forensics so checking for valve and crank continuity and gas in the lines might be as much as we can expect.
It may be better that the cause remains shrouded in mystery than to launch off on a theoretical wild goose chase with no data to support the conclusions. Yet that appears to be exactly what the Australian Transport Safety Bureau did in its investigation of a May 2000 crash of a Piper Chieftain into Spencer Gulf north of Adelaide. Normally, we wouldnt have the slightest interest in a foreign accident investigation gone awry but the ATSBs report is attracting interest in the U.S. because of its conclusions: the bureau blamed the crash on improper leaning and, by faint implication, it questioned the practice of operating engines lean of peak EGT.
In the U.S., many owners have educated themselves on the subject of leaning in general and lean-of-peak specifically and have come to a conditionally comfortable conclusion that LOP operation saves gas, money and engine wear.
The effect of the ATSB report-government sanctioned, no less-is to pull the rug out from under this hard-earned wisdom. Its an exaggeration to say the ATSB report will tumble us back into the dark ages of leaning knowledge, but it hardly advances the cause, either. (Find the full report at www.atsb.gov.au/. Date: May 31, 2000.)
The ATSBs report details VH-MZK, a Whyalla Airlines Piper Chieftain flying with a pilot and seven passengers from Adelaide to Whyalla, a distance of 125 miles, all of it flown at night, including two overwater legs. About 24 miles from Whyalla and after he commenced his descent, the pilot declared a mayday, reporting that both of the Chieftains engines had failed and that he was ditching in Spencer Gulf.
The following morning, SAR forces found floating debris but no survivors. Eventually, the aircraft was recovered with six of seven bodies. Although a ditching was apparently attempted, damage to the aircraft suggested a hard impact and the investigators couldnt tell if passengers found in the sea had egressed from or been ejected from the aircraft. Curiously, although it was a for-hire flight over water, the aircraft was equipped with neither a raft nor personal flotation devices; they were not required.
At ATSB direction, the wreckage-including both engines-was recovered and examined. In December 2001, the bureau completed what is, by anyones standards, a massively detailed and apparently thorough report of the crash, totaling 147 pages, complete with color photos, charts and footnotes. In the U.S., the equivalent crash would generate a fat factual file but a few thousand word conclusion, if that. Unfortunately, much of the ATSBs report volume is taken up by information that has nothing to do with the crash. Nonetheless, the report eventually states the reasons for the Chieftains dual engine failure:
First, the left engine failed when the crankshaft fractured due to fatigue cracking caused by a failing connecting rod bearing. Contributing to that failure were high bearing loads caused by lead oxybromide deposit-induced pre-ignition.
The bearing failure was precipitated, says the ATSB, by the inclusion of an anti-galling compound between the bearing shell and the cap, causing the bearing to slide in its cap. The lead deposits were worsened by Whyallas leaning practices, according to the ATSB.
The right engine failure, said the ATSB, was dependent in that the failure of the left caused the pilot to set higher power on the right where the very same lead deposits caused by Whyallas leaning practices set the engine into pre-ignition, which burned through a piston, leading to the fatal ditching.
The lead oxybromide theory neatly ties the two engine failures together but only the careful reader will note that its a theory utterly unsupported by convincing engineering data and one built on a flawed understanding of piston engine leaning.
The ATSBs report was released as General Aviation Modifications, Inc., the company that developed GAMIjectors, was knee-deep in test-cell ignition research using the same Lycoming TIO-540-J2B found in the Whyalla Chieftain, save for the fact that it was the dual-mag J2BD variant, rather than the twin-mag J2B.
GAMIs has dual interest in the ATSB report: the company has embarked upon a near-evangelical quest to educate pilots about leaning and any aspersions cast upon lean-of-peak EGT operation are both bad for business and, as GAMI sees it, bad for engines.
With its highly instrumented test cell in Ada, Oklahoma, GAMI donated cell time on the TIO-540-J2BD to test the ATSBs theories after the fact, since in its report, the Australian agency showed limited understanding of what was actually transpiring in the engines cylinders and, evidently, had no means to gain that understanding.
In its exhaustive analysis of the Whyalla crash, the ATSB concluded that the left engine failed first, beginning with a fatigue crack in the crankshaft initiated by the faulty rod bearing, which the ATSB says propagated over the course of 50 flights. The agency used microphotographs to substantiate this theory but the evidence is puzzling, at the least.
More microphotography illustrated that the number 6 connecting rod bearing showed evidence of sliding in its cap, which the ATSB says was caused by the use of anti-galling compound between the shell and cap, a common practice in engine building. According to the ATSB, the rod bearing failure was further encouraged by abnormally high bearing loads brought on by pre-ignition caused by lead bromide deposits.
That dovetails nicely-in theory, at least-with the ATSBs explanation of why the right engine failed. When the pilot advanced the right engines power after the left was feathered, pre-ignition caused by the lead oxybromide deposits burned a hole in the number 6 piston within minutes, causing it to fail and leading to the ditching.
The ATSB made much of the airlines leaning practices and, despite any data or engineering research to back it up, blamed the crash, in part, on leaning, a finding which led one Australian newspaper to declare breathlessly that the ATSB had made a world-first discovery linking lean fuel…causing premature ignition.
War of the Faxes/E-mails
GAMIs review of the ATSB report touched off a spirited round of faxes and e-mails between Ada and the safety bureau, some of the discussion centering on the lead deposits, how they got there and what they meant. In its investigation, the ATSB learned that in order to save fuel, Whyalla pilots were trained to lean the Chieftains engines during climb and to lean to best economy during cruise. The POH gives that as peak TIT or 1650-degrees, which ever occurs first. Whyalla pilots may actually have been using 50-degrees rich of peak.
Further, according to the accident report, the airlines manager had experimented with lean-of-peak EGT operation and the report surmises that the Chieftains engines were, from time to time, actually operated lean of peak.
Based entirely on observations that pistons from other engines that had failed in Australia exhibited lead oxybromide deposits and that similar leaning practices were used, the ATSB leapt to the conclusion that best economy leaning may promote the build-up of deposits, causing pre-ignition and ultimately engine failure.
Its the ATSBs contention that there may be an optimumcharge temperature for the formation of lead bromide. Yet the report gives no convincing data–in fact, no data at all-on what this temperature is.
In a stunning example of bending thin data to suit the theory, the ATSB seems to conclude that since Whyallas engines were operated at best economy, that must be the temperature for deposits to form. On the basis of this unsupported could be theory, it gives pre-ignition caused by leaning-induced deposits as a contributory cause, despite years of operational, engineering and recent test cell data suggesting that best economy leaning produces fewer not more deposits, although it may produce high CHTs which are bad for a different reason.
Although lean-of-peak operation is not specifically implicated and is mentioned without comment could be ignored except for what happened next. The ATSB convinced the Civil Aviation Safety Authority-the Aussie FAA-to issue a safety bulletin recommending against aggressive leaning, which CASA defined as lean-of-peak operations. As noted in the sidebar, the CASA bulletin has significant and egregious errors.
GAMIs chief engineer, George Braly, took exception to the ATSB findings and in a series of critical faxes sent to the ATSB, he argues that, if misinterpreted, the Whyalla accident report will encourage dangerously erroneous notions about the operation of high-power piston engines.
Theres nothing new about private sector companies squabbling with government agencies over accident reports. Boeing makes a sport out of doing the same with the NTSB. Whats different about this example is that GAMI has the wherewithal to actually test ATSBs theories in the test cell and can-and has-produced reams of data that run counter to the report findings.
Fewer Deposits, Not More
One confusing term is aggressive or over leaning. Depending on whos using it, the term is sometimes taken to mean operation at peak EGT, lean-of-peak EGT or something on the rich side of peak, but not full rich.
Lets insert a baseline here: A fuel/air mixture that results in complete burning of fuel is said to be chemically stoichiometric and this will normally occur slightly on the lean side of peak EGT. If manifold pressure is constant, leaning further will slow and cool combustion because theres more air than fuel. A chemically stoichiometric mixture will yield minimal deposits, a leaner mixture even less.
GAMI says that many mechanics erroneously use the term too lean to describe what is in reality nothing but a richer-than-stoichiometric mixture, a mistake the ATSB also made or at least didnt elucidate clearly in its report.
In any case, ignoring temperatures for the moment, any mixture leaner than full rich will produce fewer deposits than a full rich mixture will. Pilots who fly engines susceptible to lead fouling-Lycomings O-540, for instance-soon learn that attention to leaning is the only way to keep the plugs free of lead. In some engines, over- rich mixtures produce heavy carbon deposits, something thats well documented.
Oddly, the ATSB argued that leaner mixtures produce more deposits and that this lead to pre-ignition and destruction of both engines in the Whyalla Chieftain, despite the fact that lean mixtures demonstrably burn cleaner.
GAMIs Braly argues that the deposits found on the Whyalla aircraft pistons were perfectly normal and what youd expect to see in an engine run leaner than full rich, if not stoichiometrically lean. But lean ops cant ignore temperature considerations and Bralys analysis reveals the problem therein. The ATSB fairly well established that Whyalla pilots routinely leaned during high power climbs. If thats accurate, its a bad way to run the TIO-540-J2B series, even if it wasnt causal in the Whyalla accident.
Setting up the leaned-climb scenario in GAMIs test cell revealed that peak cylinder pressures in a leaned climb run at some 900 PSI; essentially the same as takeoff power but without benefit of excess fuel flow to slow combustion and cool the cylinders.
At slower climb speeds, its easy to imagine that the leaned Whyalla engines had CHTs of 450 degrees or more, hot enough to provoke light detonation. If the crank/bearing failure occurred due to high cylinder pressures-a reach itself, says Braly-it probably came from the climb leaning, not from pre-ignition from lead deposits which would have been burned away and scavenged out the exhaust pipe.
As for the right engine, a hole burned in the number 6 piston was a clear indication of high cylinder temperatures and pressures. But why? The ATSB again blames lean-induced lead bromide deposits for pre-ignition, this despite the fact that except for the heat damage, the pistons appear to be perfectly normal.
Braly believes the more likely scenario-which he reproduced in the test cell-is that following the left engine failure, the pilot merely advanced the power without richening the mixture, instantly driving the engine into pre-ignition not because of engine deposits but because of heavy detonation caused by spiking temperatures.
In the test cell, Braly says this occurred in a matter of minutes and would have destroyed the right engine easily within the time frame allowed. If pre-ignition did occur, the more likely cause wasnt piston deposits but a glowing helicoil tang in a spark plug hole, something Braly says GAMI has also seen in the test cell at least twice.
Pound of Salt
In our view, the ATSBs report, however well intentioned, should be read with skepticism bordering on flat out rejection. Based on the data ATSB developed, it may be plausible that detonation caused by leaning during high-power climbs played a role.
Oddly, on the one hand, the ATSB seems to look askance at leaning during the climb but on the other, it claims that Whyalla operated the engines according to manufacturer/FAA recommendations. (The POH is silent on leaning the Chieftains TIO-540-J2B during climb.)
In any case, the idea that leaning practices increase piston deposits runs counter to empirical experience and isnt supported by the evidence at hand. Unfortunately, its inclusion as a contributing cause in an accident report only serves to muddy what is already a limited understanding of piston leaning practices.
Although the ATSB cant be faulted for not being clairvoyant, worth noting is that in February of 2002-two months after ATSBs report appeared and following field reports of several fractured crankshafts-Lycoming announced SB-550, recalling for inspection some 400 new crankshafts used in 300 HP-plus 540 series engines manufactured after September 30th, 1999.
According to the ATSB report, VH-MZKs left engine-the one with the broken crank-was a factory overhaul installed in February of 2000. The report doesnt say if the overhaul included a new crankshaft. But we suspect the ATSB may want to re-open its investigation to find out.
Also With This Article
Click here to view “Leaning: A Terrible, Stinking Muddle.”