Servo Failures
 

Hi All,

I'm fortunate enough never to have had a servo failure (yet) and I'm wondering what is the most frequent type of failure i.e. do they mostly seize, do they just stop working but the servo arm remains free to move, half and half or something eles?


I ask the question in the context of dual servos driving a one piece elevator. It's possible to create a 'sliding bar' mechanism so that if one servo seizes the other will still give a degree of elevator control. However, if a servo fails but does not seize, then the amount of remaining control is dramatically reduced as the 'failed' servo cancels out much of the movement in the functioning servo i.e. there's nothing solid for the functioning servo to push against.

I don't want to open up a new debate re 'redundancy'! I want to understand better than I do what happens to servos when they fail.

Thanks for any insights

Bob

Bob, search for the threads on the Skygate Hawk elevator setup options - I reckon that will help answer what you're looking for ;)

Hi Mark,

Not sure it does. I know how to set it up but the LMA insist that I use a sliding bar arrangement (as per the original Skygate setup). The more I play with this I realize its limitations. What I'm wanting to understand is "how likely is it that a servo will seize on failure". If servos seize more than 50% of the time when they fail then the sliding bar system makes sense. If it's more likely that they just stop working but the servo arm is free to move then the sliding bar system makes less sense. I just don't have any experience of servos failing. I have had occasion to change a couple on the ground but nothing in the air so I'm finding it hard to judge this issue.

Thanks

Bob

I'll take a stab at it. The main thing you see failing in servos is the gear train, in which case, they flop around loose or skip. Or I have seen several cases of linkage failure. Broken horns, broken solder couplings and in one case a broken clevis. This leads to the bad situation of floppy control surfaces and possible flutter of the loose control surface. I am not sure I have ever seen a servo that seized. Also seen a few cases of bad electricals that made the servo appear dead but was either connector or in one case broken lead. Once saw a melted extension. So the overwhelming number of cases I have seen are either loose or dead servo(as in unpowered). Not seized. my 2 cents.

Bob, one of the threads has about 2 pages on the elevator setup, the options and the ways servos have failed before. At the end of the day, the view was the SG method was the best so I'd go with that. To be honest some of the worst built/installed models I've ever seen have been LMA models so I'd go with what you know works and is proven the world over.....

The other option is to mount the 2 servos directly to the stab in the tail under the stab cover. In that scenario you want a failed servo to go 'loose'. There is no ideal or perfect setup. It's in the lap of the Gods - where it happens, if it happens, how it happens and if it's your lucky day i.e. your chosen method tallies with the servo failure type......

I believe this is the answer your looking for when I was setting up my ultra flash I was working on the elevators I did not realize that my gear servo was Amping out. It ended up burning up. When it did it actually locked in the position it was In i had to break the gear train to get the servo arm to move

To get the real info as to failure modes you would need to pry that info out of one or more of the manufacturers, possible one of the larger dealers. The sample size that any other hobbyist is exposed to is to small to draw a conclusion.

I have had servos fail where they drove hard over to one side. I have had them fail where they simply stopped working as if the power was turned off making them free running. In 45 years in R/C I can only remember 4 in flight servo failures. It is just not real common. 3 Of those were on Helios, the other one was an elevator servo on a 40 size 3D aircraft. I have had other failures but they were all done as a result of stupidity during setup. Maybe I have just been lucky to have such a small failure rate, maybe it is because I never use just enough servo for the job, I don't know.

Will anyone that has connections to a servo manufacturer or distributer chime in here with some real data? I would like to know the real answer also.

I've only had two in-flight servo failures in many years of flying.

One was a mini servo on an electric model fitted to an aileron, it failed jammed at about 1/2 travel,
flew it home on the other aileron servo. :D

The other was on a friend's ARF model I was training him on, failed free wheeling (like there was no power on).
It was on the elevator, all I could do was cut the throttle & watch it crash. :(

A relevant question for me, my next model will have a prominent, visible single pushrod & horn on a one piece elevator. With the two examples of failures I had a sliding bar system of some sort would save the model if the servo seizes, albeit reduced throw depending on where the bad servo seizes. If it free wheels I would be better linking the servos solidly together so the good servo could drive the elevator & (hopefully) the dead servo.

John.

I had a strange failure (right Elevator) last year and was very lucky, not to loose my Ultraflash. The reason was probably not the servo but a device between receiver and servo but later I exchanged everything including the servo. What I've learned, even with the redundancy of two elevators and two channels you have no chance, when one elevator goes fully down and stays in this position. In my case it went back to normal after a few seconds. I could reproduce the failure on ground, the servo intermittantly went to full down position.

Regards
Walter

I had (thank goodness on the ground) a JR digital Titanium gear servo, go un-commanded to it's limit and just quit. Turning the power off and manually adjusting the servo back to neutral and re-power the same thing occurred. I guess the pot lost it's centering capability. Anyway it is in the scrap pile.

Glenn

Great thread! Would be a great idea to Poll it. I think Schroedm is pretty much correct, but I would love to see the true percentages over time.
Jay

Hi All,

Thanks for responding and filling a gap in my education. While trying to come up with a sliding bar solution, I kept banging my head on the obvious weaknesses of such an approach. It's better than nothing in that it will help in one of two failure scenarios but wont help with both. Now that I know that I can move on.

Thanks for getting me past my blockage!

Bob

Old cheap servos, so I don't really blame them. 2 failure modes were nothing, no movement at all, and no movement from center to one end. Travel in the other direction was OK.
Remember the Kraft servos which were so expensive that we sent them to Mr. Pullen for repair rather than just replacing them?

I think that all you can do is use the best quality servos, leads, and servo arms and connectors and try to reduce the odds of a catastrophic failure. If you have one just hope that it is not on the "next of kin" control surface the elevator.

================================================== ================================================== ====================

My company was involved in servo manufacturing for 17 yrs (1968-1983). I was aware of all failures reported by customers and in house failures. We cycle tested every servo before shipment and there were a very few that failed in test. I remember one field failure and we found a grain of sand that locked the gear train. Another customer reported a servo reversed in the air and he landed with reversed elevator. My first reaction was this was impossible but later found Mitsumi had changed the way they anchored the magnets in the servo motor and the magnet came loose and rotated 180 degrees under vibration.

I feel vibration is the main cause of servo failure today. If a power connection or signal wire connection breaks the servo is going to be dead. If a pot connection breaks the servo is probably going to drive full travel. I suspect that most servo amplifier failures are caused by over voltage or prolonged high loads from mechanical linkage.


I have been flying R/C models since 1952 and have never personally had a servo failure in the air.

WOW!!
I'm not sure if many of us (including myself) would've "figure" that out if it happened while flying our jet..., but maybe slightly easier to ascertain with something slower with plenty of altitude!

Great info. Sid.

In this case, as most have mentioned here, the source of failure and its effect can be anything. By putting 2 servos on the same linkage may not buy you much because the linkage could fail (sliding mechanism, pushrod, control horn, clevices, etc..). Also, you are introducing more parts that could fail, the slider mechanism and the second servo. Having 2 separate elevator halves with individual servos might save you if the servo or linkage fails in a non-hardover condition. The servo going hardover to one extreme is only one of the many failure modes, so you have statistics on your side.
But in the end, because these are toys and we don't want to get into super complex and heavy failure mode mitigation, I have to agree with JJP and say do the best possible selection of hardware: good servos, good pushrods, clevices and control horns. And check them everytime you go flying. My two worst crashes I ever had were not because of servo failures, but because of stupid distractions causing me not to check that the servos were reversed!!!

Wonder how something like this on elevator would work out. If it had a form of circuit breaker on each servo, then in theory the good servos could possibly overpower an un commanded full deflection until the breaker tripped, work a dead free swinging servo, or possibly strip the arm of a frozen servo. I'm sure it'd have to be tweaked with the right equipment (possibly plastic gear servos).
I understand that no matter what you do with your elev setup, you still only have 1 TX & 1 power distributer, but I think full scale is so safe today because known problems are addressed one by one.
Jay

You would still have a non-redundant connection to the flight control surface that could fail. Which fails more frequently, the servo, the linkage to the flight surface, or the flight surface itself?

Have to agree with the statement on vibration being he hardest thing on the onboard electronics. Vibration can cause all sorts of stress related wire to individual component breakage. Minimizing the complexity of the system is sometimes the best thing. I do like the split elevator idea and try and incorporate when possible but on the other hand when you get a hard over on any one servo you are pretty much "cooked" !

Agreed. Now if there were some "smarts" that could be added to servos to detect hard-over failures and shut off power to itself, that would really help. Then the servo would be partly or totally re-centered from aerodynamic forces on the control surface, maybe.

I have had three servo failures, two in flight. All were hard-over failures - electronic failures - with no damage in the gear train.

================================================== ================================================== ======

JP,
I would like to know what voltage your were running on the servos and what voltage are the servos rated for?
Sid

Good idea for next generation R/C systems. Either with an additionel sensor built in the servo or external pickups. As soon as there is a difference between commanded and measured position, the servo will go into damping or dead mode.

Regards, Walter

And it doesn't seem like a very hard thing to do either. Just a simple 8-bit microcontroller and some sensors for each servo.