Hi there SimPilots,

This poll is once again related to the ongoing development of the new site and the Job Portal (and associated processes integral to company flying). This one should be a tad bit less controversial than the last one I hope.

Would you be kind enough to post ...

"What setting(s) do you use for Random Failure (On/Off)"

"What value, if On, do you use for 'Mean time between failure' - in hours?"

If you also have a short anecdote/suggestion/advice, please share as well.

Thank you very much,

bc

Bonus Question:

For any of you longtime X-Plane enthusiasts.

Has there been any/significant changes to the Failures between Version 9 and 10.?

I haven't flown enough in V10 to be sure and would appreciate any commentary on that.

Thanks muchly,

bc

I'm the chief windshield washer and I do not run with "failures" on.

Perhaps it's because I've had so many RW failures over the years. I never really thought about it.

What do you run with bc? I'll give your settings a shot and check them out in XP10.

I find flying challenging enough with out Random Failures turned on, pilot error and Mother Nature cause enough problems.  As my competance level improves I will likely enjoy the added thrill of mechanical failure, but not yet.

Dirob

I do not remember ever setting it, but I do have failures. I guess I am using system default and I have never changed it to my recollection.

It is on and set to 10,000 hours, which seems like Never, but I do have failures.(XP9)

haven't checked XP10 as yet, but should be same story since I don't change default.

As it was explained to me, the failure matrix is random based on the probability level of the dart hitting a failure on a target. The original program had very few failures on the dart board, but many have been added over the years and updates, leading to a higher probability of SOME failure happening than in the days of old. The recommendation was to ignore the "hours between failures" aspect and bump the number up to what at first glance appears impossibly high. I think I bumped mine to something like 75,000. Failures are now realistically rare, but they do sometimes occur. I like being kept on my toes, but I didn't like having a failure every three or four days. When I snapped a prop shaft on decent with minimal power applied, that was the last straw (ending a 3 hour flight just short of the airfield) and I took the sage advice of increasing the odds in my favor. In the real world I would use fanatical attention to maintenance to accomplish the same end.

Mine are on. About one failure every 10,000hrs... almost never get any failures.

Perfect guys. This is the kind of data that is useful to me.

@Zymurg - alright sir, I shall take you up on that. Thank you for the Test Pilot offer

I have refined my failures in V9.7 to...

8,800

which seems perfect for GA and bush craft.

Please put that number into your failures in Version 10.25 and if you would, let me know any time you have a failure that is *not* associated with the developers built-in failures (like the Carenado C90, the Red Eyes AN-2 etc.). It would be great if, when you report a failure, you could let me know approximately how many hours you have flown since the last failure (any aircraft - don't need failures per craft). Thank you very much. Once you have some data to report, I will compare it to mine (which I have dutifully logged each occurence) and then share any differences so we all know.

Everyone else... please keep the numbers coming - the more I have the better of an understanding I shall have. While I certainly agree with those of you who turn failures off, and understand your reasoning, I am building a process that will be highly enhanced by the refined knowledge of how and when across the versions.

Thanks to all who have posted. Much appreciated,

bc

I have mine set on 5000 in v10.25, but I am also running xHobbs which make that number vary somewhat. I changed to that setting around the time we were building Beaver Creek in conjunction with me installing xHobbs in v10 (I had it in v9). To date, I don't think I have had any failures in a non-Carenado aircraft in that time. I'd have to go back through the forums but I think I've had two major failures in Carenado planes in that time. Flight times probably have been running at a ratio of 2/3 Carenado and 1/3 Others.

I turned mine on and input 8800 hours.  V10.25.

 If I crash I'll prosecute to the greatest degree of root beer justice.;-(

I have it turned off in XP10, which is the only version I have ever used.  I don't know what its default setting was, but I turned it off in annoyance, figuring I had enough to do in learning the quirks of the sim itself without adding failures into the mix.  I'm pretty sure the only Carenado I had back then would have been the 152 Cessna, and I *think* I remember encountering a failure with it. And that would have been with v.1 of the 152, but I don't know whether that makes any difference.  

I have mine off as well, for the same reason stickybaby cited.  I find X-Plane exciting enough already, and doubly so with EPOCH's activities within it.  Someday, if the flying ever gets boring, I'd consider switching it back on.

Great anecdotes folks . Allow me please...

@jetjerry: Thank You very much. In an odd twist of 'oldguyitis' on my part, I incorrectly identified Zymurg (who is up to his elbows in Ultimate devving right now) as offering to Test Pilot the numbers. I do thank you for rejoining the thread and offering and I look forward to any reports from you.

@Carenado fliers: Well, according to Dan K. at Carenado, the only one of their aircraft with custom failures is the C90. That's it. Now I have had a couple of failures over the last year in their aircraft and would have sworn it was driven by the craft and not the default processes 'armed' by the Failure menu, but apparently I'm wrong.

@Everyone with Failures=Zero: Totally understood, and for the most part, you are right on the mark by turning this off. But there is value in this feature if, and only if, it is used correctly and in very specific instances, which I will propose down the road a ittle ways. The math bears me out and I hope the practise supports it. We shall see.

Thanks again Pilots. It makes the build much easier when receiving expert opinions and reports, not to mention the aspect of an enjoyable discussion with friends

Cheers,

bc

A friend of mine has the C90, and if you over temp/trq the engines they will quit.

So, guys and gals, watch your Redlines!

All right, bc, this down-the-road "proposal" might just pique my interest, there.  I'm nothing like an expert in any of the XP craft I horse through the air, and my level of interest in the sim in general doesn't run to knowing all the systems in complex aircraft and their attendant failure recoveries.  But if we're talking about losing the odd instrument or function while flying one of those EAA missions that might have a few tufts of hair on it, well sir, just count me in!  Makes the refreshments taste all the better after landing with a, um, laundry problem, don't ya know.   

Well said Sticky.

I'm going to get a little long-winded here, but I would like to expand upon this 'idea' that Austin had (random fails), but obliquely, as I'm not quite ready to breif on my ideas yet.

I'm also going to ask Cheif Pilot jetjerry to ponder this post, as he's the one citing experience with RW failures. Also, if any other current or retired RW pilots are reading this, perhaps a comment from you if you have a moment would be telling.

So, I'm going to go back up this thread first of all. Dirob stated (Post #4) that he would enjoy the challenge of mechanical issues after he feels competent dealing with flying-related challenges and the weather. So true... these are the primary issues that pilots face daily. Mechanical/electrical issues, especially in modern aircraft, are very few and far between.

Now, let's go to Zymurg in Post #5, who states that he has experienced failures but doesn't expand upon them. I do believe I was even on TS flying with him and the other Nightfliers once when he had a failure and had to 'let down'. Zymurg seems to consider failures as part and parcel of the experience, as he never indicated any personal issue with them.

Then down to Mr. Dare in Post #6. May I quote...

"When I snapped a prop shaft on decent with minimal power applied, that was the last straw (ending a 3 hour flight just short of the airfield) and I took the sage advice of increasing the odds in my favor. In the real world I would use fanatical attention to maintenance to accomplish the same end."

I wonder if he was flying an old piston pounder round engine during that descent (like the Classic Otter or, the doozy of all, the DC-3). Why, you ask? Because a descent with minimal power can result in the engine/prop going "below square", where the prop is driving the engine instead of vice versa. Guess what you can snap? Most of the time damage to the crankshaft occurs, but you can snap a prop shaft if the damaged crank sets up a sympathetic vibration. Here's a RW DC-3 pilot about flying 'square'...

"The rules we live by are:

Above 140 kts IAS... MP at least 5" above square.

Above 100 kts IAS... MP at least square.

Below 100 kts IAS... MP can be reduced below square.

I never found any reason to go below square until flair for landing."

So, could this be what happened/was simulated. I, for one, cannot get a straight answer from anyone - not devs from Version 5 onwards nor Austin himself. Some go on about just how involved the failure coding is, with variables spanning flight model interactions, atmospheric conditions through to type of equipment. Others just say 'hogwash', that it is a simple pseudo-random generator that, as Mr. Dare suggests, is nothing more than a dart thrown at a failure board. BTW... Austin says that a significant chunk of the flight loop tests for errors, to the point where he even pondered writing the routines in Assembler to realize a speed increase over compiled "C". Like I said though - it's the big question when nobody has the same answer for sure and I'm never quite sure about Austin .

vsully has 10,000 set. But only states anecdotally that he 'almost never gets any' failures. At 8,800 I can assert that I average 1 failure per 84.39 flight hours. Aircraft specific or biased... not yet sure, but I am analyzing that data, as I have it all in my logbook.

Based on rick.hudd, sitting at 5,000 for a value (and expanding/projecting a bit upon his post) he has had 2 majors in Carenado craft (unknown species) and 1 in a non-Carenado. That's 3 but over what period of time/flight hours we're unsure. Not all that important either right now, so don't anyone go logging just for the sake of having verifiable data. Fly... have fun - it's my job to slog through these things to make us the best bush VA ever

Sticky and Rocky are following the "No Fail" practise which again is probably the most accurate (almost) for modern models but not so much for the older ones (if the 'heavy and subtle code' people are to be believed). Both who have very valid reasons for not wanting to be saddled with such an annoyance during fun and educational - remember this word flights.

That leaves those of us flying models where you can't turn off failure modes... the study and near-study sim models like the AN-2 and the C90. I wonder how many simpilots avoid these craft for the very reason of the fails. I do recall a certain 'groan' even around here when we were all dabbling with the freeware AN-2 and it's annoying habit of insisting on the right engine management or it would unceremoniously break out in fire.

So now... to what I wanted to really say about failures and why I am studying it.

1. There can be no regular EPOCH flier that would state that the flying job tests, contracts and missions have been no more challenging than flying the pattern at the local aerodrome on the plains of New Zealand. It's tough and hard flying - hard on the pilot and even hard on the machines. Are all of your landings under 150 fpm at touchdown? I thought not.

2. Alaskan operators that I have either studied deeply or in 3 cases interviewed all agree that the hiring criteria (especially the check rides with the company CP) are more rigorous than most small and medium commercial operators in the "easy" part of the country (lower 48 )

But, despite all this, I wonder if you would take a moment and reflect on the following...

Failures:

So why do pilots plan for them, practise for emergencies as a result of them, simulate them and then hope they never happen?

Because even your brand spankin' new Cessna can have a failure. Even your highly maintained old warbird can shudder and fall. Based on an entire night of reading and tabulating accidents from the Aviation Safety Institute's database, it happens more often than you might think. About 6% of the time it seems.

Note that my failures are set to 8,800 - a setting that seems to provide a random failure less than 2% of the time (always in GA aircraft for the most part).

... just one of those things that make you go "hmmmmmmm"

bc

I leave it turned on, at one point I had set it to 100 (no zeros missing) and as I took off I would have a failure, sometimes before take off. Didn't take long to jump it up. Default is 10,000 I believe. I think I recall early beta XP-10 v with def set I had many more failures. Not sure of flt hrs/failure time currently but, it happens and seems like 1-2 in 20 flts or so? I see lots of blown tires. For carb ice, icing or other systems like turning on gens etc. we can control these by (check sheet) turning them on or not and are act builder gen.. 

Your last sentence is surely telling gofigure.

If in fact the "random failures" are tied to a probability table that increases the likelihood of a failure, then it behooves us to at least attempt to stay within manufacturer recommendations.

If however the "random failures" are just a dice toss during the flight loop, then no matter what we do, we will either fail or not, based on 'luck'.

But let's look at the difference, in a very simplified paper airplane example and with a bit of pseudocode to illustrate.

A38000 Paper Airplane Model by Sammy Hagar - Nickname, "Take a Ride..."

Definable Failures = AOA outside of bounds, Weight out of limits, Weather (airframe stressors).

Flight Model Table

AOA: 5 units to 10 units.

Weight: 20 units to 30 units.

Weather: Wind 2 units to 5 units.

Okay, now the two types of failure randomizers...

First, the simple crap shoot -

[Flight Loop Begins]

if RandomFail_occur_now = Random(seed number) then RandomFail (Definable Failure)

... other stuff for flying....

[Flight Loop Ends]

then the more complex 'edge of the envelope'...

[Flight Loop Begins]

if AOA < 5 units or >10 units then Definable Failure = AOA(related equipment):FailCheck

if Weight <20 units or >30 units then Definable Failure = Weight(rel. equip): FailCheck

if Wind <2 units or >5 units then Definable Failure = Wind(related equipment): FailCheck

end the ifs

FailCheck:

if RandomFail_occur_now ~= Random(seed number) then RandomFail (Definable Failure)

return

... other stuff for flying....

[Flight Loop Ends]

note: " ~ " = approximately equal to

While that code is near-silly and way over-simplified, it does count out as being about the same size in bytes as the real code, with differences.

The simple model has 1 test, 1 function and 1 assignment.

The complex model has 9 tests, 3 functions and 6 assignments.

6 times the 'complexity' (read: system resources required).

Multiply that by a hundred failures and then double it for complex test algorithms and you may consume 1,000 to 2,000 times the "time" within the flight loop that the one simple random failure test uses.

But so what. Ever looked at the flight loop in a Carenado aircraft. Dig around and see how the test for drag on floats/gear/skis works. Add up all of the Carenado flight loop additions and then look at the difference between it and a high-quality, no-plug-in model and you will only see maybe 10 frames per second difference at the worst. And that's with an interpreted scripting langauge, not a low level compiled langauge like X-Plane is built in.

So I believe that the potential for a complex failure mode test environment exists. How complex or how simple is the real question, but I honestly believe that there is more going on here than just a simple roll of the failure dice. I think

Does Austin play dice with our universe?   Maybe only Bill Gates knows for sure

Just some food for thought.

Thanks for the post gofigure!

bc

As I said above, I had changed my settings around the time the Beaver Creek Hub construction flights began and since we were building Beaver Creek and the subsequent scheduled routes before my new years resolution of keeping a true logbook, I will have to estimate my hours. All of the flying at Beaver Creek, the Winter Resupply, the New Zealand Vacation, and USAK Season 2 would probably add up to around the 50-60hr mark. So that is 2 major failures in 50 hours (both Carenedo planes but not the C90 - I think it was the Beech A36 and the C340).

I will also be the first to admit that my knowledge of flying (or lack thereof) could contribute to a higher mortality rate on ALL aircraft parts. (just ask the 5 C185s floating upside down in Turner Lake)

So a failure 4% of the time. Interesting and I've marked this in my book. Thanks.

The last statement you make here is the one that really interests me Rick (and I don't know what it is about everybody's last words, but wow).

You cite "lack" of flying knowledge contributing possibly to a higher , let's call it STRESS rate on your equipment.

I submit that in the RW, the pilot's lack of attention to the systems, or lapses in attention, or complete lala land in his mind is what allows many small failures to begin the chain of events leading up to the big crash. Not their knowledge... but like the old saying goes...

You can lead a horse to water but you can't make it drink.

Your assumed lack of knowledge probably gives you a bit of an edge, since you tend to be perhaps hyper-vigilant watching the aircraft 'dance' in anticipation of anything that may 'look' or 'feel' wrong.

Thanks for the post and further clarification.

bc