Systems Reliability - Need Component Failure Probability Estimates

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,589
Location
Saline Michigan
I and another EAA chapter member are working on architecture decisions in our respective airplanes. In the process, I have assembled Failure Modes and Effects Analyses for several common options. Unfortunately, we have too small an experience base for confidence in estimating failure probabilities on the many components. We are looking for a broader experience set in estimating a number of things:
  • Likelihood of Failure
  • How likely individual items are to fail at engine start as opposed to in-flight?
  • Will we be able to detect the failures using common post-start and pre-takeoff checklist processes?
  • How likely are detection measures in finding impending failures?
If an item is available from B&C, please assume they are the source. My colleague in this exercise is using a Lycoming with factory supplied Surefly Ignition and Bendix Fuel Injection. My homebuilt is planned with SDS fuel injection and ignition. Estimates specific to the hardware are great.

Please help by giving your estimates of in-flight failure probabilities. Even if you can tell me how many hours you have and what failures you have experienced, I can collect that into an estimate of failure probabilities. I will take you thoughts on detection and failure modes on the following components:
  • Electrical components:
    • Batteries;
    • Carling DPDT switches;
    • Continuous Duty Contactors;
    • Continuous Duty Relays;
    • Alternators;
    • Regulators;
    • Diode packs and Diodes;
  • Ignition Equipment;
    • Conventional Magnetos;
    • E-Mags;
    • SDS Ignition and Coils;
    • Sure Fly Ignition;
  • Fuel Handling;
    • Mechanical Fuel Pumps;
    • Electric Fuel Pumps, Traditional;
    • Electric Fuel Transfer Pumps, Facet;
    • Electric Fuel Injection Pumps, Walbro;
    • SDS Fuel Injection;
    • Fuel Selector Valves, Traditional;
    • Fuel Selector Valves, Andair and Newton/SPRL
  • Sensors supporting electronic fuel and spark;
  • Avionics;
    • Avidyne GPS/NAV/COM;
    • Garmin GPS/NAV/COM
    • Dynon HDX Display/Processors;
    • Garmin Displays and Processors.
Thanks in advance for any help you can give us in estimating reliabilities in support of our efforts.

Billski
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,589
Location
Saline Michigan
21 of you have visited this page and not even one of you will share "I have X many hours in little planes and have had Y of this failure and Z of that failure"?

OK, I shall start. In 2200 hours I have had:
  • Two plugs on one cylinder foul and cut out the cylinder, with the first cutting out only a couple minutes before the second. The Archer maintained altitude on three cylinders;
  • Two Facet standby electric fuel pumps failed on the ground;
  • One alternator failure in flight;
  • One attitude gyro gave up in flight on an instrument training flight in VMC - that made for an interesting flight;
  • Two vacuum pumps failed - both in VMC;
  • One fuel sending unit lost its float while IMC, causing a divert to the nearest ILS in case the tank really was draining. It wasn't. And;
  • Miscellaneous light bulbs. They do not seem to be out when checked on the ground. They must burn out during flight.
No inflight battery or contactor or switch issues, near as I can tell from review of the logs, the switches and contactors all have 6700 hours on them. I will look back through the logs on the airplane for other failures.

Come on guys, let's hear your totals and all the things that have broken. A perusal of the log books for repairs won't hurt either.

Billski
 
Last edited:

mcrae0104

Well-Known Member
Log Member
Joined
Oct 27, 2009
Messages
4,164
Location
KBJC
Best estimate of hours, either PIC or right seat, approx. 450. The only failures I have experienced are:
  1. An intermittent Bendix com radio. I could not transmit or receive the tower immediately after takeoff for approx 30 seconds while departing my home class D. It had worked fine on the ground. After trying unsuccessfully several times to talk with them, and just as I was reaching to squawk 7600 and wishing I could remember light gun signals better, I received a transmission from the tower, and the problem went away. After the flight, another pilot said that plane had been experiencing weak TX/RX lately. I tested my headset in several other planes just to make sure that wasn’t the issue—it worked fine. The radio was fixed and I’ve never had any issue with it again.
  2. Spotty ADS-B reception on another flight (GNX-375). No big problem, just loss of traffic. I do not know whether it was broadcasting ADS-B. The GPS was working fine. Afterwards, just as I shut down, someone I’d never seen before sped over to me on a golf cart. The first words out of his mouth were, “Did you just...” and immediately I panicked—had I busted a TFR?—but he finished, “have any trouble with ADS-B? That plane’s been acting up.” Also fixed, no problem since.
  3. One burned-out position light found in preflight.

    A friend who has approx. 500-600 hours in his Bonanza has experienced one mag failure, caught on runup. I don’t know how many hours were on the mag. Unfortunately, it happened to him in a small town in Iowa while on his way to Oshkosh with three passengers. The FBO loaned him their crew car to complete the trip, but their mechanic was away for the week at OSH. Miraculously, he was able to locate another mechanic who had the correct mag on hand and was willing to drive an hour and a half to do the repair.

 

tspear

Well-Known Member
Joined
Feb 12, 2014
Messages
1,105
Location
Outside Boston
In the Aerostar, over roughly 300 hours I had the following.
1. Power spike. Mechanic was not sure where started in regulator or alternator. Both fried, took out over voltage circuit breaker and pretty much everything in between. About Less then 100 hours on new alternator. 100amp National (I think that was the name)
2. Lost GNS530W in flight. Garmin never said what failed, just pay the flat fee to fix. GNS had less than a thousand hours on it.
3. JPM, lost oil pressure sensor. Was a chafed wire that shorted out.
4. Engine starter, before flight.
5. Four mags failed, all found in run up. All first flight after IRAN from separate shops.
6. Wastegate controlers, turbos, failed in flight. (Aerostar has two per engine)

In Cirrus DR20/22, over 450 hours between two planes:
1. Lost MFD and PFD, Avidyne, both in flight. Back light failed.
2. Cracked case, the crack was larger than my hand, IO360. Less than a thousand hours. Mechanic stated it happened in flight, oil loss pre crack was normal, one quart between oil changes (30-35 hour hobbs time). Last flight landed with less than a quart or two. He said I likely had another ten to fifteen minutes at best before the engine froze. CMI replaced the engine, gave me full credit and I spent less then 16K total. :)
3. GNS430 failed; corrupted data on card. At start up. Had to replace the card.

Tim
 

Marc Zeitlin

Exalted Grand Poobah
Joined
Dec 11, 2015
Messages
1,067
Location
Tehachapi, CA
Please help by giving your estimates of in-flight failure probabilities... Thanks in advance for any help you can give us in estimating reliabilities in support of our efforts.
I've re-arranged your post to reply in the order in which I think things should be considered. First of all, anecdotal reports of failures are (as you know, Bill), completely meaningless from a statistical standpoint in trying to evaluate failure probabilities and MTBF's. So no matter who reports what here, in however many hours, it will not give you ANY useful information for failure probability calculations. Zero. Nada. Zilch. Nothing.

So given those facts, I'll describe the process we went through in developing the FMEA/SSA/FHA and Fault Trees for WK2/SS2/RM2 at Scaled, back in 2007 - 2009, with a not totally dissimilar situation, in that we had a boatload of components on the Mothership, Spaceship and Rocket Motor that we had invented out of thin air or had custom built for us, and so for most of them, had zero documented information on failure rates. Now, there were MANY components that we DID have actual failure rates on (wire, connectors, relays, hydraulic fittings and lines, fuel pumps, etc.) but that didn't change the fact that a substantial number of components had completely unknown MTBFs.

  • Likelihood of Failure
So the first thing you do is assume that EVERY item in EVERY fault tree has a MTBF of 500 hours. Why 500 hours? Because it's a freaking MISERABLE failure rate, and only the worst components on the planet will have a failure rate that bad (think magnetos and vacuum pumps, for starters). So you develop your fault tree logic, then fill in the failure rate for each component with an MTBF of 500 hours. If the fault tree meets your failure criteria, you stop - no need to know actual failure rates, because it'll only be better, and you already meet your criteria.

Let's assume, for the sake of argument, that you want the following failure rates for the various failure levels:
  • Catastrophic - 1e10-7
  • Hazardous - 1e10-6
  • Major -1e10-5
Don't bother to even create fault trees for failures that aren't in one of these three categories - you could even not develop trees for Major Failures, as they only lead to higher pilot workload, not aircraft damage or injury.

Now, if you have a fault tree that fails, you look at the critical path of the tree to see what components are leading to the high failure rate - usually, it will be a single string with one or two bad items that are "OR'd" together, rather than "AND'd" (redundant). If it's a common item (wire, connector, relay, mechanical or hydraulic component) you go to the relevant handbook (MIL-HDBK-217F, NSWC's "Handbook of Reliability Prediction Procedures for Mechanical Equipment", or the like) and find a failure probability for the one or two components in question. If that (better, hopefully, than 1/500/hr) leads to success in the tree, you're done. If it does NOT, then you need to reconfigure your system, usually by making the driving component redundant.

AC23.1309-1E is a good reference for the process, with a lot of good examples.

  • How likely individual items are to fail at engine start as opposed to in-flight?
Why do you care about non-safety critical failure modes? They're useful to think about and mitigate, but you don't create fault trees for them - they're mission failures, but not safety failures. Every failure in the FHA list should have "Phase of Flight" associated with it, as well as "Severity Level". If the severity level isn't catastrophic or hazardous (and except for maybe an explosion, it's hard to imagine a "engine start" failure leading to a catastrophic or hazardous severity level) then you're not even analyzing it.

  • Will we be able to detect the failures using common post-start and pre-takeoff checklist processes?
Part of the failure evaluation for each functional failure will be a description of the detection methodology - detection of each failure must be built into the system, if it's anything above Major. Whether by checklist, indicator light, EFIS warning, or airplane doing something obviously non-nominal, there needs be a detection methodology. Generally, an "undetected" failure will increase the severity level and lead to modifications to the system architecture.

  • How likely are detection measures in finding impending failures?
Impending failures are difficult in our small airplanes with minimal sensory inputs. Insight's engine monitors have the ability to have a real-time vibration analysis (FFT, I believe) displayed on screen, so can assist in predictive failure analysis. But they don't integrate with any of the EFIS MFG's equipment, so are totally standalone. They paid no attention to me 5 - 7 years ago when I suggested creating a module they could sell to the EFIS MFG's - looked at me like I had three heads.

With respect to electrical, hydraulic/fuel failures (pretty much the only "systems" subject to SSA on our planes), there's little in the way of predictive failure analysis, other than the yearly CI and monitoring things for change.

Hope this isn't just telling you crap you already know... Maybe I misunderstood what you were looking for.
 

proppastie

Well-Known Member
Log Member
Joined
Feb 19, 2012
Messages
5,454
Location
NJ
I happen to believe anecdotal data is somewhat relevant in the real world.....I have 3100 Hrs. 1200 in the Mooney aprox. 1900 in PA 28-140...1 hr. experimental.

(from memory):

Major crash totaled aircraft 1 hr. experimental (pilot error)

1 Fuel starvation off airport landing (pilot error) no damage

5 clogged injectors issues 1200 hr. Mooney (poor maintenance perhaps the cause )

broken field wires alternators 3

broken main wire at terminal alternator 1 in say at 2000 hr. Mooney

2 bad diode alternator say 1500 hr each

Bad starters 3

6 new batteries

Iran noisy electric fuel pump 500 hr. (Weldon) No issues Facet in PA 28

1 failed brakes (poor maintenance) at 1500 hr.
 

Marc Zeitlin

Exalted Grand Poobah
Joined
Dec 11, 2015
Messages
1,067
Location
Tehachapi, CA
I happen to believe anecdotal data is somewhat relevant in the real world....
Anecdotal data can be very useful in pointing out things about which it MIGHT be necessary to be concerned. "Hey - did you know that THIS can fail?" or "Hey - here's this failure mode that most folks don't consider".

But this is why, during the FHA stage of the SSA, you spend an inordinate time listing all the functions the aircraft must perform as well as the multitude of failure types of each of those functions in each phase of flight with a severity level attached to each. Only once you've determined all the failure modes (and even for a small, single engine GA aircraft, there might be hundreds of different failure modes of tens of different functions), do you proceed to the Fault Trees, where the statistics comes into play.

So from a statistical standpoint, which is what "Failure Analysis" is all about - how many failures are allowed within a given time period, calculated using MTBF (where the "M" stands for "Mean" - OWKA "Average") - anecdotes are useless. The plural of "anecdote" is not "data". So 50 people might answer Bill's request here with stories like yours about what failures they've had, and it in no way informs the calculations needed to create accurate fault trees using failure probability data. Is your experience normal? Do most pilots with Mooneys with <whatever alternator you have> have zero failures in 3000 hours, or 25 failures in 3000 hours? 5 people reporting here that they've had multiple alternator failures tells us exactly nothing about the population as a whole - Mooney alternator failures might have an MTBF of 50,000 hours, or an MTBF of 350 hours, and your experience would have been exactly the same.

Anecdotes may help Bill list possible failures (maybe), but they won't (not might, but won't) help in determining failure rates.
 

TFF

Well-Known Member
Joined
Apr 28, 2010
Messages
16,162
Location
Memphis, TN
Is the B&C alternator driven from the flywheel or accessory pad?

For regular alternators on Lycomings, I end up changing the brushes about three times before engine TBO. Continental forward mounted integral alternators lost 500 hrs which is the TBO. The pad mounted number 2 alternator on Cirrus last about 1200 hours.

Bendix mags if maintenance is done for both hour and calendar, they will work to TBO.

I have seen P mags hiccup on first start with about 500 hours. Off and back on would make work. The rebuild is pretty inexpensive that corrects it when it happens.

Starters depend on type.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
2,380
Location
Pocahontas MS
Please help by giving your estimates of in-flight failure probabilities.
If enough people answer, you're going to see failures for every possible item/widget, except the ones that killed the potential responder. And it will be completely unscientific data, since it isn't randomized and you're asking a selected group and responses are voluntary, instead of polling the entire pilot 'universe'.

IMO, this info is only useful at all if you're using the info to pick a particular brand/model widget. As I think I hinted in my response on the Aeroelectric Connection list, if the widget isn't flight critical, you're only shopping for best value/dollar. On the other hand, outside of things that just can't be backed up, like the airframe itself and the engine (assuming we're limited to single engine a/c), if the widget truly is flight-critical, there should be a backup, regardless of e10-x numbers. No one (well, almost no one) flies with only one magneto functional, even though mags really are pretty reliable and almost always have 'soft' failures that give us clues when they're on the way out. Or, what are the odds of a missing shoulder harness killing us? We still wear them, though.

I could give you my list (the ones I can remember). In ~1200 hrs of flying across 8 different a/c: loss of tailwheel on a Luscombe, leaky cockpit + clogged drain holes in the fuselage resulting is gallons of water shifting rearward, causing serious control issues (during solo pattern work as a student pilot), swallowed valve (Lyc in C172), rough running engine/high stress precautionary landing causing broken prop/bent gear (engine teardown found a cracked head on one cylinder, unrelated to the incident), Lyc injector/primer port plug backing out on one cylinder, causing extreme rough running at anything less than wide open throttle, lost part of the (epoxy) leading edge protection from a near-prehistoric Sensenich wood prop, lost the carbon fiber 'bent tip' on a Warnke wood prop that he was experimenting with, lost the alternator on a Lyc powered Swift, lost a couple of alternators on homebuilts, SLA battery went bad due to what was likely undetected shipping damage (tiny hole in the case), loss of transmit function on a comm radio (pin backed out of the connector), reversed sensing function in a VOR receiver/display (showed opposite radial of the one I was on), probably more that I don't remember at the moment.

The only ones in that list that I would tie to inherent product issues are the, wait for it, swallowed valve, and the cracked head, on separate Certified Lycoming Engines.

Charlie
 

Pops

Well-Known Member
Log Member
Joined
Jan 1, 2013
Messages
10,170
Location
USA.
In 51 years.
2-- charging failures . One was broken wire at the terminal on the alternator.
1-- Vacuum pump failure in IMC. Down to 300' in a hard rain.
1-- Crankshaft broke in a Lyc. Flew it on the rear 2 cylinders 7 miles to an airport. Broke at just in front of the center main bearing at an angle and the rear part turned 360 degs and hit against the front broken part and turned the prop.
1-- spark plug go bad 1.5 hrs on new plugs.
1 -- oil pressure sending unit broke off at the crankcase of the Lyc with losing oil. I was high on a long cross-country and with-in gliding distance of 4 airports. Had 2 qts in the engine when I landed.
1-- complete radios failure except for the transponder ferrying an airplane for the new owner over northern Minnesota in Feb with 8' of snow on the ground with a warm spell at -5 degs and almost dark, and didn't know how long I have been following 2 dead VORs. Aircraft with a new annual inspection and all radio's just back from the shop.

Lost a couple inspection panels while playing with my daughter's Piper Colt with the 160 hp Lyc and converted to a taildragger because I let the airspeed go about 15 mph over VNE.
Lost brakes on an Ercoupe and took about 5K feet to get stopped on a downsloped runway at a large airport.
All I remember at this time.

Added--- Nothing wrong with the airplane until another airplane came from above and behind and flew through my left wing at night with prop slices about 6" from my left shoulder and slicing down the side of the instrument panel and engine cowl. Only good part was the right landing gear. Mid-air and all the parts were never found. ( including part of my arm).
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,589
Location
Saline Michigan
What my colleague and I are doing is examining our airplane systems and electrical schemes, and comparing our intended equipment and electrical configurations to a traditionally configured airplane. We are not attempting to meet some target reliability - we are attempting to get an idea if we are actually reducing our likelihood of having severe results by adjusting aircraft configuration. Airplanes being considered are:
  • A traditionally wired and plumbed Cherokee with carburetor and magnetos - we are assuming an all electronic panel;
  • An RV-9 with Bendix fuel injection and Sure-Fly ignition and its base fuel system with wing fuel tanks;
  • My homebuilt with SDS EFII and a fuel system with wing fuel tanks and a header tank fed by lift pumps;
  • We are looking at traditional electrical system and AeroElectric Connection‘s Z-12, Z-14, and Z-101. My colleague had already decided that Z-12 looked appropriate for his RV-9, while I had independently decided that Z-14 seemed more appropriate for my bird;
  • The problem with coming to these conclusions was that while we each thought that our judgement was sound, we had no quantitative way of knowing how much better we were than the traditional plan, nor of knowing is Z-14 was worth the trip over Z-12, or if Z-101 was worthwhile.
I had commenced a system FMEA with the intent of looking at failure modes and their effects in each system. I had made a number of assumptions about reliability of each component in the system. Our analysis does not go down to the level of wires and connections, but it does include many of the components and the redundancy or lack thereof. I suppose I could include the wiring as well, but I was rather surprised to find that all three of the AeroElectric Connection's wiring plans are remarkably close in numbers of the various types of components and even of total wires in the ship power systems.

I am trying to get something in the way of failures seen and total hours to in some way validate our assumptions on component reliability. I appreciate the inputs and ask that everyone who finds this thread contribute their failures and the hours and if it was caught before flight or not. The purpose for asking folks to tell me if it was caught before flight was so I could separate the in-flight incidents from those detected before they would put flight at risk. The difference matters: Only the ones not caught before flight count in what I am doing.

Understand that until I started getting data from you guys, I was limited to my 2200 hours and the failures in our airplane, plus the things I had personally heard about within a flying club. Anything would be an improvement upon that VERY limited data set.

Marc Zietlin's suggestion of consulting AC23.1309-1E will be taken. I suspect that study of this tome will fill my breaks from fairing the fuselage over the next few days. I suspect that the data I extract from this AC will expand my knowledge base considerably.

Being as we are looking not to satisfy some minimum failure rates, but to estimate failure rates and severities of competing schemes, looking at the several systems in each plane, and attempting to place similar levels of skepticism on the analysis of each, I am running all the failure permutations. Severity of many of the effects of failure modes is vary considerable between the airplane configurations and wiring schemes.

I was taught FMEA and statistical processes in 1980 as a new du Pont employee, where all severities, failure probabilities, and detection levels were handled by order of magnitude, that is exponents only. Going into finer detail may be appropriate. We shall see. I have been using ever more “modern“ versions of FMEA ever since which makes what appears to be finer gradations on probabilities and detection. I also understand that historical data is often referred by duPont Engineering Services Division as "hysterical data", and can never take the place of test data and statistically valid approaches. Nonetheless, we can use the anecdotal history to check if our other information is “in the ballpark” or fails our basic laugh check. Consulting AC23.1309-1E, we shall see if we can further improve upon our numbers.

So please keep supplying the hours and failures you guys have. I shall tally the hours and various failures and attempt some sort of accounting of the data. After we finish our analyses, we plan to present the results to our local EAA chapters and perhaps write an article for Sport Aviation or KitPlanes. Sharing with the folks contributing data is also planned.

Billski
 
Last edited:

Pops

Well-Known Member
Log Member
Joined
Jan 1, 2013
Messages
10,170
Location
USA.
The only failure that I caught before a flight was the dead sparkplug on runup before takeoff. About 1.5 hrs on the plug.
 

Marc Zeitlin

Exalted Grand Poobah
Joined
Dec 11, 2015
Messages
1,067
Location
Tehachapi, CA
So please keep supplying the hours and failures you guys have...
Since you ask, I'll comment wrt known failures on one thing only that you specifically asked about.

I have been maintaining a list of Emagair product (i.e., P-Mag) failures on canard aircraft, since 2004 when they came out and I started using them. My list includes 34 total failures on 17 different "N" numbers (six of them mine - I'm a stubborn guy...). Most if not all of these 34 failures occurred within the first 200 hours of flight time (on the P-Mag) - many within the first hundred hours. Of course, I have no idea what the total # of hours of Emagair products that DIDN'T fail was, but to the extent I can tell, well over 10% of the total P-Mags installed on canard aircraft have failed at one time or another. That's a pretty miserable failure rate for a flight critical component. I have a COZY MKIV in the shop with two P-Mags that had one repaired twice within the first 100 hours of use, and the failure of which MAY have caused a failure of the accessory case drive gear on the left side of the IO-360. This was a known failure mode for which Emagair has issued an SB. Sucks for the guy with the $26K repair bill...

P-Mag reliability has been substantially better in the RV world, to the extent I can tell - there may be something about the vibration and/or heat signature on a canard pusher with which these units are unable to deal, but who knows - Emagair has never acknowledged that their failures were due to anything other than the customer's incorrect installation or an engine problem, even after they were presented with conclusive evidence of poor mechanical design, poor fabrication, or poor QC.

In any case, I STRONGLY discourage P-Mag products on canard aircraft, and even more strongly discourage using TWO P-Mags on one engine on canard aircraft.

And I have never heard of an in flight SDS failure of any type, either EI or EFII.
 

Chilton

Well-Known Member
Joined
Jun 28, 2014
Messages
185
Location
Jersey, channel islands
Hi Bill,

Looking back on my notes from operating in Tanzania covering about 8000 hours of Cessna 206 / 210 and about 20,000 hours of Cessna 208 which I was either pilot or company maintenance supervisor.

Bear in mind the environment we were operating in was hot and dusty so many electronic components were placed under heat stresses they were not intended for, Alternators and starters certainly suffered from the dust.

  • Electrical components:
    • Batteries; on the Cessna 206/210 I show an average of 18 months or 1000 hours between failures, all failing capacity check at maintenance rather than failure to start the engine, the C208 we averaged about 1200 hours between failures, in both cases I work on an average of 2 starts per flight hour.
    • Continuous Duty Contactors; one failure, failed open before start,
    • Continuous Duty Relays; one failure over 8000 hours on the piston engine types, 1 failed on C208 which I think my records cover 10,000 hours both were fail open and on first start of the day.
    • Alternators; Piston engine types 3 failures in 8000 hours.
    • Regulators; I don’t have a failure recorded
    • Diode packs and Diodes; I don’t have a failure recorded
  • Ignition Equipment;
    • Conventional Magnetos; On the Continental O-550 I had one mag go weak before TBO, found pre departure with a larger than normal mag drop still within limits, removed and checked as a precaution and subsequently reported to have a failing internal part, on the Lycoming 540 with the duplex mag the failure rate was about 200 hours between removals from the aircraft.
    • Fuel Handling;
      • Mechanical Fuel Pumps; I never saw a failure before TBO, I did hear of a company overhauling an engine without touching the fuel pump and having the pump fail about 500 or so hours later.
      • Electric Fuel Pumps, Traditional; 1 failed on a C206 discovered when it would not prime for start mid way through the day
      • Electric Fuel Transfer Pumps, Facet; 2 failed, 1 failed in flight failing to transfer fuel but running electrically, the other discovered failed before first flight of the day when it did not run when switched on.
      • Fuel Selector Valves, Traditional; Not really a failure in terms of not working but I had to remove three for leak rectification in 8000 hours on the 206/210 in all three cases they were found to be leaking during routine maintenance due to fuel staining on the skin below.
      • Avionics;
        • Garmin GPS/NAV/COM 2 failed one on departure with a major internal failure just after takeoff, one screen faded to black in flight. Some level of screen failure leading to rectification was about 1000 hour item for us across the piston and turbine fleets.
        • Garmin Displays and Processors. multiple failures on the G1000, almost all in the early examples of the G1000 C208, I never got involved with one in a C206. after the first 2 or 3 in Tanzania had 3 or 4 failures each the failure rate dropped rapidly to the point that now everyone thinks they dont fail!
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,589
Location
Saline Michigan
Is the B&C alternator driven from the flywheel or accessory pad?
Plan for both my ship and for my colleague's (this is irrelevant to this thread, but I will bite) are Lycomings with one B&C in the nominal position, and one B&C on the vacuum pump pad.
 
Last edited:

Tiger Tim

Well-Known Member
Joined
Apr 26, 2013
Messages
4,119
Location
Thunder Bay
Even if you can tell me how many hours you have and what failures you have experienced
I have around 7000TT and definitely haven’t tracked specifics but here’s some off the top of my head:

-blown electrical feeder relay. Happened right on rotation and even though the electrical system is split into a nearly redundant left and right system, I lost the side that powered the intercom and with it lost my headset.

-failed landing gear: no power to motor.

-failed landing gear: door sequencing issue, external maintenance switch left in wrong position.

-flap failure: FOD in flap drive

-engine failure: high pressure fuel pump drive sheared

-too many fuel boost pump failures to count. We used to run them right to failure but they’d give about a ten hour warning by getting noisier.

-roll control failure. A wing rib flange cracked, allowing the skin to flex up under air loads and jam the control surface. This was extra insidious since it couldn’t be easily detected on the ground.

-front crank case seal failure. Luckily in the circuit at the time but it made landing a bit more interesting with the windscreen blacked out.

-somehow I’ve never had a magneto fail.
 

Map

Well-Known Member
Joined
Dec 29, 2020
Messages
105
Location
California
What to me seems more productive than trying to estimate failure probabilities for all components on an airplane is to look for single point failures that can have catastrophic or hazardous effects. Eliminate them by building in redundancy. And look for components that tend to have gradual failure modes or give plenty of warnings that they are about to die.
 

speedracer

Well-Known Member
Joined
Feb 4, 2020
Messages
252
In around 2,000 hours on two different Long EZ's I had 6 mag failures. I always ran one magneto and one Lightspeed Engineering electronic ignition and had no failures of the EI's. 2-3 years ago the lightbulb popped on and after the 6th mag failure I tossed it in the garbage can and installed a second LSE EI. The last mag had 400 hours since rebuild. A shaft bearing was so worn I could wiggle the drive gear and watch the points open and close. Flying home single ignition while zigzagging over airports at 17,999' is a little nerve wracking.
 
Top