Over two years, in a fleet of about 300 piston airplanes under Savvy Aviation's management, six magnetos failed in the ugliest way a magneto can fail. That works out to roughly one per hundred airplanes per year, which is a lot more often than most owners would guess.
The pilots were a fair cross-section. Some were new, some were CFIs with thousands of hours. The failures happened everywhere from the traffic pattern to FL210. And in all six, with all those different people in all those different situations, not one of them tried the magnetos one at a time to work out which one had gone bad.
Every one of them declared an emergency, pulled the power back because the engine was shaking itself apart, and landed at the nearest airport. Nobody got hurt and nothing got bent. But every one of those flights could have ended at its planned destination, with a smooth-running engine, if the pilot had reached up and turned a key one notch.
The failure everybody pictures
Part 33.37 is short about it. Every certificated spark ignition aircraft engine must have a dual ignition system, with at least two spark plugs per cylinder, and two separate electrical circuits fed from two separate sources of energy. The FAA wrote that rule because ignition failures are ordinary events, not exotic ones.
The most common failure of all is a spark plug, and it is almost always harmless. With two plugs in the cylinder, one is enough to keep it making power. Unless you are watching an engine monitor in normalized mode, where the bars line up in the middle of the display and the sensitivity goes way up, you will probably fly the whole flight without noticing. The only sign is the EGT on that cylinder rising 50 degrees or so, because one flame front takes longer to cross the chamber than two, so combustion is still going on when the exhaust valve opens.
What kills plugs is usually crud in the electrodes. Oil, carbon, lead. Sometimes it clears itself, which Busch has watched happen in his own airplane more than once: an EGT pops up, he does an in-flight mag check, identifies the dead plug, writes it on a Post-it so he can change it after landing, and fifteen minutes later the problem has gone away on its own. The pounding of a few thousand combustion events knocked the deposit loose. Carbon and oil fouling will often shake free like that. Lead usually will not.
He has a photograph of a fine wire plug that came out of a client's engine with a glob of metallic lead sitting in it, big enough that a massive electrode plug would have shorted out cold. The fine wire kept firing, because the lead happened to land somewhere that did not bridge anything.
A magneto that dies outright is nearly as undramatic. Points fail, a coil goes open, a condenser shorts, and half the spark plugs in the engine stop working. Every cylinder keeps producing power on single ignition, all the EGTs rise together, and you fly on to where you were going and get it fixed there.
That is the failure pilots have in their heads when they think about magnetos, and it is the reason the other kind catches them so badly.
The gear made of plastic
Inside the magneto, a large distributor gear carries a rotating finger that hands the spark to the right plug at the right moment. It is made of plastic, as a lot of a magneto's insides are, because it has to stay non-conductive with tens of thousands of volts going past it.
Sometimes it sheds teeth. When it does, the gear either stops turning or, more often, starts turning erratically and slips out of sync with the engine. Now the magneto is firing spark plugs in the wrong cylinders at the wrong times, and the engine goes somewhere no checklist covers.
Change of underwear rough.
That is the phrase Busch uses, and pilots who have had it happen do not argue with it. It is not the familiar lopsided shudder of one weak cylinder. It is the whole engine going wild, because the sparks are lighting the mixture whenever they feel like it.
The half hour over Cincinnati
Here is where the six failures stop being a maintenance statistic.
Every one of those pilots had a perfectly healthy second magneto sitting right there. Selecting it, on its own, would have brought the engine back to smooth running immediately. That is the entire reason the FAA requires two.
The one that stayed with Busch was a turbocharged Cirrus at FL210 near Cincinnati, flown by an experienced pilot. It took him nearly half an hour to descend, power off, and land at Lunken. Half an hour of the engine tearing itself about, with nothing else to do but think, and the idea of trying one magneto at a time never came to him.
None of this is stupidity. When an engine starts shaking that hard, reaching up and switching off half its ignition is close to the last thing instinct suggests. Pilots go to full rich, they turn on the boost pump, they do everything the training gave them. Turning off a mag is not on that list, partly because nobody told them a magneto can fail in a way where that is the answer.
It is counterintuitive to push the nose down when the airplane stalls. We train it until it is reflexive anyway.
The procedure is not complicated. If the engine goes violently rough, try each magneto on its own. It takes a couple of seconds each. If one of them makes the engine run smoothly, leave it there and go land somewhere sensible. If neither helps, you have learned something useful too and you are no worse off.
One and a half magnetos
Some Lycoming engines do not have two magnetos in the ordinary sense. They have a Bendix D-3000 dual magneto: two magnetos packed into one housing, on one drive shaft, bolted to one pad on the accessory case. If your Lycoming's model number ends in the letter D, an O-360-A1F6D or a TIO-540-F2BD or an O-320-H2AD, that is what you have.
It was a reasonable-sounding idea. Accessory cases are crowded, and this saved a pad and some gearing. Busch wrote a column about it years ago titled "Are one and a half mags enough?", and some CFIs he knows will not fly a single-engine airplane fitted with one.
The problem is a handful of single points where one thing failing takes out both magnetos at once.
The mounting clamps are the classic. Two clamps hold the magneto to the accessory case and let it rotate for timing before the nuts are tightened. They have a history of loosening, and the dual mag is a heavy thing to hold still. When it creeps in the mount, both magnetos go off timing together. Lycoming is on its third clamp design trying to solve this, the newest having considerably more contact area with the mag flange. There is a service bulletin and no AD, which tells you roughly how seriously to take it: seriously.
The impulse coupling is another. It sits on the drive shaft to help the engine start, it has a fair number of moving parts, and its failure rate is not trivial. One shaft, one coupling, both magnetos.
Then there is a screw. The dual mag uses a single cam to drive two sets of breaker points, one per magneto, and like most Bendix magnetos the cam is held by a screw that you loosen to set the internal timing. Busch only learned this one recently, from an owner who had to abort a takeoff when the engine failed. The screw had backed off and the cam had slipped, and with one cam serving both magnetos there was nothing left to fall back on.
The dual mag satisfies the letter of the two-source requirement. Whether it satisfies the intent is a question Busch is fairly blunt about. Most of these engines can be converted to conventional magnetos, but it means a different accessory case and different gearing, so it is an overhaul-time decision rather than a Tuesday one. The engine comes back with a model number that no longer ends in D.
The other single points
Two conventional magnetos have their own ways of failing together.
The ignition switch is one. It is a key-operated rotary switch that grounds one magneto, or the other, or both, and works the starter as well, and there are ADs against some brands. If it fails in a way that grounds both magnetos at once, the entire ignition system is gone. Busch has talked to owners who chased that fault by replacing both magnetos with overhauled units and then the whole harness, and only found the switch at the very end because it was the last thing anybody suspected. Twins tend to use individual toggle switches instead, which are far more reliable, and experimental owners can simply fit them.
The other one is stranger. Turbocharged airplanes, particularly those with the physically small Slick magnetos, pressurise the magneto cases with air bled from the induction system to stop them misfiring at altitude. That air goes through a small plastic filter first, to keep induction system crud and moisture out of the magneto.
Every pressurised installation Busch has ever seen uses one filter for both magnetos, with the output teed into two hoses.
One of Savvy's clients was up in the flight levels in a Cessna 414 when an engine started misfiring badly. A plastic nipple on that filter had fatigue-fractured, both magnetos lost pressurisation at the same instant, and both went into high altitude misfire. He declared, descended, and the magnetos started behaving again once the air got thicker. When they came off for inspection the insides were a mess, with plastic parts melted by all the internal arcing, and both needed major repair.
The physics of it is worth understanding. In a turbocharged engine the cylinder stays at sea level pressure or better all the way up, while the air inside an unpressurised magneto gets thinner with altitude. Dense air insulates well and thin air does not. Climb high enough and the easiest path for the spark is no longer across the plug gap out in the cylinder; it is a shortcut inside the distributor block, where the air has gone thin. There are two cures. Pressurise the case so the magneto sees the same pressure the cylinder does, or use a magneto physically big enough that nothing is close enough to arc across, which is what the large Bendix S-1200s do.
While we are ranking them: Slick magnetos have historically been the worst of the three, the smaller Bendix S-200s are better, and the S-1200s are far and away the best. The catch is size. Plenty of engines have nowhere to put one.
Why nobody opens a magneto
At an annual or a 100 hour, a magneto gets a look and very little else. The mag-to-engine timing is checked and corrected, sometimes the points get inspected, and that is about the extent of it. Nobody goes inside, because going inside means pulling the magneto off the engine and taking it apart.
Both manufacturers, Bendix through Continental and Slick through Champion, recommend exactly that every 500 hours. It is a recommendation, so Part 91 operators are free to ignore it, and plenty do.
Busch is philosophically a maintenance-on-condition man and says this recommendation goes against his grain. He follows it anyway, for two reasons.
The first is that there is no way to judge a magneto's condition from the outside. You cannot borescope one. There is no filter to cut open, no oil to send to a lab, no trend to watch. It is a sealed box full of plastic parts, felt wicks that need lubricating, and a carbon brush that carries the high voltage from the coil to the distributor gear and wears down as it does so. The only way to know is to look.
The second reason is the six pilots. Busch runs his vacuum pumps to failure, because there are two of them and one will fly the airplane. By the same logic he ought to run magnetos to failure too.
Redundancy does not help you if you do not know how to take advantage of it.
A 500 hour inspection means the magneto comes apart. The plastic parts get inspected and the distributor gear gets replaced if it looks the least bit doubtful. Consumables like the carbon brush get renewed. The internals that need it get lubricated with the specific greases magnetos want. The points get checked and the gap reset. And the E-gap, the magneto's internal timing, which determines how much voltage it actually produces, gets set properly. On a Bendix that means loosening that same cam screw and rotating the cam.
One thing worth getting right when you book it: ask for a 500 hour IRAN, not an overhaul. IRAN stands for inspect and repair as necessary, and the last two words are the point. Legally a magneto only has to be overhauled or replaced when the engine is overhauled. In an IRAN the technician judges each part on its condition and replaces what needs replacing. In an overhaul his hands are tied and a long list of parts gets replaced whether or not anything is wrong with them. An IRAN runs somewhere around five or six hundred dollars and takes about a week; a factory rebuilt Bendix S-1200 is closer to two thousand.
The condenser is the one part that ages on the calendar rather than the tach. They are electrolytic capacitors, they dry out, and about ten years is what you get. A shorted one kills the magneto outright. An open one is worse in a slow way: the points start arcing, they erode, the timing drifts off, and eventually the magneto quits anyway. A good shop checks the date code and swaps it if it is old, which is cheap.
The check that actually tests something
The mag check you do at the hold short line is a weak test, and it is worth being honest about why. At 1700 or 2000 RPM with a rich mixture, it takes almost nothing to light the charge. A marginal ignition system will pass that check comfortably. It only catches the failures that were already obvious.
The demanding version is done in cruise. Set up around 65% power, lean as aggressively as the engine will take, preferably lean of peak, and then check the magnetos. High power and a lean mixture both make the charge harder to ignite, so the ignition system has to be genuinely healthy to get through it. Savvy calls it an ignition system stress test, and publishes the whole procedure as a flight test profile alongside the GAMI lean test.
What you are looking for is simple enough. Switch to one magneto and every EGT should rise, none should fall, and after the rise they should sit fairly stable. The engine will run slightly rougher on one magneto than on two, which is normal. It should not run anywhere near change-of-underwear rough.
If you want the engine monitor data to be worth reading afterwards, go slowly. Stay on the single magneto for at least ten sample intervals: ten seconds if your monitor logs once a second, a full minute if it logs every six. Then both for the same, then the other magneto, then both again. Marginal plugs and drifting mag timing are obvious in that data and invisible from the cockpit.
One thing to know before you try it. If you switch to a magneto and the engine quits, which in Savvy's experience has never actually happened but is theoretically possible with a dead mag you did not know about, do not flip straight back to BOTH. The engine has been pumping unburned mixture into the exhaust system the whole time it was dead. Pull the mixture to idle cutoff first, then select both magnetos, then bring the mixture back in and let it relight. Otherwise you get a bang loud enough to ruin your afternoon and quite capable of destroying a muffler's flame cone.
When the timing goes wrong instead
The failures worth actually fearing are the ones that move the timing rather than stopping the spark, because those hurt the engine rather than merely inconveniencing you.
A magneto firing five degrees early sends cylinder head temperatures through the roof. Ten degrees early can melt a hole in a piston, wreck spark plugs, and separate a cylinder head from its barrel. Advanced timing lights the charge too soon, peak pressure arrives while the piston is still coming up, and the cylinder gets detonation and heat instead of work.
Which is why the answer to "how long can I fly it like this" is always the same, and it has nothing to do with magnetos. Watch the cylinder head temperatures and set an alarm. Busch uses roughly 400°F for Continental cylinders and 420 for Lycomings, and treats going above it as something to fix right now rather than something to think about.
He is fond of a recent example that was not a magneto at all. A pilot doing touch and goes forgot to richen the mixture for the go. One cylinder went into detonation and sat at 550°F for about five minutes. The borescope showed nothing obviously wrong afterwards, so the airplane was returned to service. The engine monitor data from the next flight made it clear the cylinder was badly hurt, and off it came.
The short version
Spark plugs fail constantly and it barely matters. Magnetos that die outright cost you nothing but a repair at the other end. The one to know about is the distributor gear that sheds teeth, starts firing spark plugs at random, and makes the engine feel like it is coming off the mounts.
That failure has an answer, and the answer is the thing nobody does. Try one magneto, then the other. Two seconds each. Six pilots in a row have proved that the reflex does not come for free.
Get the 500 hour inspections done, and ask for an IRAN rather than an overhaul. Do your mag checks in cruise with the engine leaned, where they mean something. If you fly behind a Lycoming with a D on the end of its model number, know that your two magnetos share more parts than you would like.
Magnetos are dinosaurs. Cars stopped using them decades ago.
Busch is not sentimental about any of this and would rather see electronic ignition on everything. The FAA currently allows one magneto on a certificated engine to be replaced with an electronic system, and not both, so most of us are flying with at least one of these things for a while yet.
This article is based on the Savvy Aviation webinar How Mags Fail by Mike Busch.