# When to use a Castle nut and cotter pin ?

Discussion in 'Workshop Tips and Secrets / Tools' started by Pops, Feb 26, 2019.

### Help Support HomeBuiltAirplanes Forum by donating:

1. Feb 26, 2019

### Pops

#### Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,929
5,842
Location:
USA.
I know you need to used a castle nut and cotter pin when bolting something when the parts move (rotate) as in the Cub type landing gear legs, etc. Control cable shackle, etc.
Been looking for the answer , nothing yet. How about when using a rod end where the bolt is torqued down and the rod is turning on the rod end ball that is stationary ?

2. Feb 26, 2019

### TFF

#### Well-Known Member

Joined:
Apr 28, 2010
Messages:
11,437
3,176
Location:
Memphis, TN
Depends. Flight controls? It use to be ok by the FAA to just have a nylock nut. For flight controls on new aircraft the FAA requires two forms of safety, so many use castle nylock with the cotter pin as two. Also depends on if I can see it or not. Can’t see it, I’m sticking a cotter pin in it.

Victor Bravo and don january like this.
3. Feb 26, 2019

### Victor Bravo

#### Well-Known Member

Joined:
Jul 30, 2014
Messages:
5,886
4,710
Location:
KWHP, Los Angeles CA, USA
A spherical ball joint usually qualifies as allowing movement without trying to loosen the nut/bolt. If the ball joint freezes or rusts or gets pitted, and it doesn't want to rotate easily... you could make a case that in an extreme situation it could try to impart a rotational force on the fastener, which could maaaay-beeee loosen it. Unlikely but certainly possible.

HOWEVER, if you had a flight or engine control using a ball joint and a Castle Nut-Cotter pin, you also have the equally valid possibility of the end of the Cotter Pin catching on something (electrical wire, control cable, edge of a bulkhead) and jamming or impeding a flight control because of that.

Because of this, in the "real world", I believe a good, new, tight locknut with 2 threads showing is likely as low-risk as a Cotter Pin. Many factors could slant the odds one way or another.

4. Feb 27, 2019

### Pops

#### Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,929
5,842
Location:
USA.
Wish all the DAR's would agree on the answer one way or the other. I agree VB, using a new All Steel locknut with 2 threads showing and in an area that can be inspected, I believe its good. But some DAR's would disagree and that is who you have to satisfy.

5. Feb 27, 2019

### Victor Bravo

#### Well-Known Member

Joined:
Jul 30, 2014
Messages:
5,886
4,710
Location:
KWHP, Los Angeles CA, USA
Where is this attachment and what is it for? If it is a control linkage and there is something that a Cotter Pin can catch on, then you can make a case with the DAR that you wanted to eliminate that problem because it is more likely than any other type of problem. If it is in an area where there is nothing that it can catch on, then it is hard to make that same argument. If it is something you are never going to see again once the structure is closed, then the Cotter Pin can be explained as be ing more fail-safe over a longer period of time.

So the question is what environment is this joint going to be in.

6. Feb 27, 2019

### cvairwerks

#### Well-Known Member

Joined:
May 12, 2010
Messages:
173
65
Location:
North Texas
In the production world, any rod end gets a castle nut and cotter key or safety cable.

7. Feb 27, 2019

### Pops

#### Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,929
5,842
Location:
USA.
So to cover all bases and DAR's it looks like going to the castle nut and cotter key is the way to go.
Thanks.

8. Mar 1, 2019

### Marc Zeitlin

#### Well-Known Member

Joined:
Dec 11, 2015
Messages:
450
617
Location:
Tehachapi, CA
I will respectfully disagree that this is the only way to go. While it is possible to use a castle nut and cotter pin to fasten a rod-end to <something> properly, I rarely see it DONE properly. See AC43.13-1B Section 7 paragraph 40(f), which discusses castle nut torquing. This does NOT address when to use them - just how to install them. It clearly indicates that there should be torque on the nut/bolt, but many times when castle nuts are used in control systems (I've seen them on rod ends and engine control pushrods, as well as other control system components) they are installed essentially loose or at best finger tight, with the cotter pin being the only thing keeping the nut from running off the bolt. In many of these cases, vibration has caused substantial motion and wear of the bolt holding the rod-end onto the <component>.

Next, see Section 7, paragraph 64, which discusses nuts of various types. Section (b) says that castle nuts with cotter pins can be used in ANY system, but obviously the caveat is that they must have drilled bolts and must be torqued appropriately per 7-40(f). Section (a) says that self-locking nuts (and the first part of the paragraph states that they're referring to metal and/or fiber self-locking nuts here) shouldn't be used on parts subject to rotation. But the inside ball of a rod-end is NOT subject to rotation on its axis - it's SUPPOSED to be clamped tight against whatever it's connected to.

So as an A&P, given the above, I'm perfectly happy to see either metal or fiber locknuts on rod-ends (with a large area washer, so that if the ball gets loose in the housing, it's still captured and can't slide over the nut) in the cabin of an aircraft, and in the engine compartment, only metal locking nuts are acceptable. I'm also perfectly OK with castle nuts and cotter pins, as long as they're installed and torqued correctly.

If talking about parts that ARE subject to rotation by design, THEN I believe that AC43.13-1B implies that only castle nuts and cotter pins are acceptable, because they're not considered "self-locking" nuts, even though they're in the **** paragraph labeled "self-locking nuts".

Sigh...

Pops and wsimpso1 like this.
9. Mar 1, 2019

### wsimpso1

#### Well-Known Member

Joined:
Oct 19, 2003
Messages:
5,727
2,958
Location:
Saline Michigan
I have used FMEA a bunch in my career since being trained by DuPont in 1980, so when I look at a system, I first marvel over how it was designed (sometimes the marveling is at the beauty, and sometimes at the lousiness) then at how well the failure modes are shoved out.

I have always looked at rod ends and other linkage items for how it can mess up. It seems that with a spherical rod end, its intended function is covered wonderfully with a fiber locking nut. But how can it fail and do we worry over it? Well, if it is flight controls or engine controls, we definitely worry over it. So, failure modes:

Containment of Ball in Housing is Lost - If it is sandwiched on both sides with the arm, we are done, but if it is open on one side, we apply a large washer to keep it from getting away;

Ball Seizes within Housing - If the pilot continues to exercise the control, it will move and one or more of the surfaces may get loosened (whichever joint is slipping). Which options do we have?
Ball Slips on the Arm - the control action will get sloppy, and it may exercise the bolt and nut;
Ball Slips on the Washer (not sandwiched) - control action will get sloppy and the nut may become loosened, and;
Washer slips on the Nut - Control action which apply torque to the nut.

Now if the Ball is Seized in the Housing, I can easily feature that it is also Seized to the Bolt - If this joint is then exercised, the nut will have torque applied to it.

To actually do an FMEA, we would need to assess failure mode severity and liklihood - I WILL not attempt that. But just looking at things that can happen gets my attention. So what to do about them?

We already either sandwich the ball between arms or use a big washer to keep the ball somewhat contained, but the other modes all seem to include loosing the nut. We definitely do not want the nut coming off. Keeping it tight once the other modes commence is not going to be assured, so we have to be happy with retaining the nut. We all know that the cotter pin will resist a decent amount of torque, but only a little rotation of the nut relative to the bolt will shear the pin. If the pin shears, the nut may continue to rotate - there will be some drag from the remnants of the pin within the bolt, but I dislike counting on that. Use a self locking nut has its good point in that it will have a prevailing torque required to unwind it until the self locking feature comes off the bolt.

In my mind, seized parts appear to be possible to loosen a self locking nut and to shear the cotter pin. The important thing would seem to be how much torque and how much energy is needed to commence loosening and then to unwind the nut in each case. Got me tempted to run a few tests out in the shop with a direct reading torque wrench...

Once other thing to think about is that when we use the rod end with one side open, the bolt will wobble about, the hole in the arm will open with continued action, and we may well keep enough load and thus torque on the nut to continue loosening and wind the nut past the retaining features (sheared pin or self locking portion). But if instead, we sandwich the ball between paired arms, once the axial preload on the nut is lost, the torque to further unwind the nut is lost.

From a failure mode management perspective, it appears to this engineer that the best scheme is probably to sandwich the rod end between two pieces of the same arm, as the rate of failure progression drops with it once the preload is lost. Once you are committed design-wise to single sided arms (it is done a lot and I do not like changing designs on known airplanes) I do not know which will stay put more often once bearing failures commence. Yeah, I am thinking about the eight places in my aileron circuit that are single shear on ball-type rod ends...

Enough of the over-educated voice - Let's get the voice of experience on how this stuff all really runs out in the world. What failure modes do we actually see? Which schemes stay put better when the failure modes do occur?

Billski

BoKu likes this.
10. Mar 1, 2019

### ScaleBirdsScott

#### Well-Known Member

Joined:
Feb 10, 2015
Messages:
953
611
Location:
Uncasville, CT
Not to go off the trail too far but, what about Nord-lock type wedge bolts/washers/etc? I like the theory behind them, seen their videos comparing them to traditional self-locking nuts and lock washers and so on, but havn't bothered to use them personally, and so I'm wondering if they used in aviation to any degree? Are they just an expensive gimmick that doesn't truly solve a problem, or a pragmatic solution to certain cases?

11. Mar 1, 2019

### Pops

#### Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,929
5,842
Location:
USA.
I'll still stand by my quote, if the DAR will not sign it off, you don't fly. So if you want a sign off, the DAR is the boss.

But, I agree with you on the rest of your post. Castle nut just snugged up so the part can still rotate ( like the Cub landing gear leg) install cotter pin. On a rod-end the bolt and nut need to be torqued up and the rod end rotates on the captured ball.
In this case, I always use an all metal self locking nut, because it seems to take more torque to turn it.

The reason I ask the question. A friend of mine just had his third homebuilt turned down by a $800 DAR inspection because he didn't use castle nuts and cotter pins on the rod -ends. He told him to fix it and let me know when you wants another inspection and left. Will he have to pay another$800 ?, we well see.
Almost forgot, The DAR also wanted to see that all the AD's were done on the Experimental Cont - C-85 engine.
The owner is upset to say the least.

12. Mar 1, 2019

### Hot Wings

#### Well-Known MemberHBA Supporter

Joined:
Nov 14, 2009
Messages:
6,212
2,186
Location:
Rocky Mountains
Probably way off topic:

This is just another example of why the FAA needs to change the way they approve DARs. It is perfectly with in the FAAs mandate to require persons providing this service to be qualified. IMHO for them to control the number and distribution of DARs is not. Safety is one thing. Market control is something best left to the market.

gtae07 and Pops like this.
13. Mar 1, 2019

### cvairwerks

#### Well-Known Member

Joined:
May 12, 2010
Messages:
173
65
Location:
North Texas
Scott: Nords require the one washer face to impart damage to the structure as part of the locking mechanism. Same reason you don’t see star locks or split lock washers on aluminum structure.

BoKu likes this.
14. Mar 1, 2019

### ScaleBirdsScott

#### Well-Known Member

Joined:
Feb 10, 2015
Messages:
953
611
Location:
Uncasville, CT
I did recall seeing that it left impressions on the part. Pretty sure I wouldn't use them on aluminum for that same reason as you mention. But as I havn't really used them personally I wasn't sure if they were just leaving some marks on a finish, or actually cutting into the parts. I suppose that's a matter of torque and material hardnesses. Whether on a steel structure if those marks are significant enough to compromise the part hard for me to know. But nevertheless that is a major downside for our applications in aviation.

15. Mar 1, 2019

### Dan Thomas

#### Well-Known Member

Joined:
Sep 18, 2008
Messages:
4,786
2,017
You wouldn't believe the number of times, and failed components, we've found due to incorrectly installed fasteners. It seems that far too many mechanics don't understand what the designed intended; the mechanics are just wrench-turners that didn't get the education required.

I've seen control surface hinges destroyed by loose hardware, left like that by mechanics that think the bolt is the bearing surface. Apart from the aforementioned Cub landing gear and the like, which uses substantial lengths of tubing of doubled or tripled layers of thick 4130 sheet to bear the loads, many hinges involve light aluminum brackets with a bushing trapped between them by the bolt, and that bushing rotates in a bearing pressed into a lug on the mating surface. The bolt needs to be torqued up to clamp that bushing tightly so it can't have any relative movement against those light aluminum brackets. Cessna, where much of my maintenance experience lies, does this a lot, in numerous places, and they use nylocks or all-metal locknuts, not cotter-pinned castellated nuts and bolts. This is common in the light-aircraft production world and I have yet to see a locknut missing. In some places, like the aileron bellcrank inside the wing, a castellated nut with its cotter pin and longer, drilled bolt, can snag on adjacent structure and cause control problems. I've found that more than once. Pinned assemblies require longer bolts that often can't fit into confined places, and getting the pin in (or out) can be a nightmare.

Loose bolts in such assemblies rapidly wear out expensive brackets. Oh, the bracket might only be $200, but its replacement requires removal of the control surface, drilling out of rivets, and installing the new part and then reassembling everything. Sometimes you might have to disassemble the surface to get at stuff. Ugh.$. In the 180/185, the stabilizer hinges are retained by bushings and bolts at the aft end of the tailcone, and those bolts pass through the ends of aluminum angles that run about five feet forward in the fuselage on both sides, and aluminum brackets picking up the inboard sides of the hinges. Those angles are VERY expensive and VERY difficult to replace, and it's all because some guy left the bolts loose so the stab could rotate for trim function. The bushings are supposed to be clamped by the 1/4" bolts at 70 in-lb, as specified by the manuals, which many mechanics don't read, maybe don't even have. And the bolts use MS21042 locknuts, which Cessna specifies in a lot of places. No pinning.

Engine controls usually use all-steel locknuts or pinned nuts, with pinned nuts required on throttle controls for many aircraft. Nylocks ahead of the firewall are frowned upon. Yet I have found nylock nuts on the exhaust pipe clamp bolts! Duh. Nylon long gone. Nuts were still tight. Aircraft exhaust systems typically run very hot, often glowing red at full power.

Seized rod-end balls and the like are a result of irresponsible maintenance, like never changing the oil in your car. There's no need for such stuff at all, but we see it anyway bceause too many owners are cheap, or own far more airplane than they can afford. They push the shop to keep the labor times down, and inspection detail suffers. In the end, the airplane is a worthless mess of junk once the guy goes to sell it.

Last edited: Mar 1, 2019
D Hillberg, Marc Zeitlin and wsimpso1 like this.
16. Mar 1, 2019

### Victor Bravo

#### Well-Known Member

Joined:
Jul 30, 2014
Messages:
5,886
4,710
Location:
KWHP, Los Angeles CA, USA
Marc Zeitlin's post reminded me that you often cannot perfectly torque a nut when you are using a Cotter pin, because you usually have to loosen or tighten it a little to get the bolt holes to line up with the castellations in the nut.

So if proper torque is important (bolt stretch, slack removal, clamping force, or pre-load on a structure, etc.) then you may be better off with a locknut.

On a truly life-critical component where you have zero tolerance for the fastener coming apart, you can "peen" or "stake" the threads downstream of the nut, to providee a last line of defense. This makes it into a semi-permanent fastener instead of a removable fastener, usually requiring you to throw away the hardware if you ever have to disassemble the joint.

17. Mar 1, 2019

### Pops

#### Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,929
5,842
Location:
USA.
When I am building, I temporary assemble with non-locking nuts. Then when assembling for the last time I put the proper nut on and torque and mark with a red paint mark.

gtae07, wsimpso1 and BoKu like this.
18. Mar 1, 2019

### Marc Zeitlin

#### Well-Known Member

Joined:
Dec 11, 2015
Messages:
450
617
Location:
Tehachapi, CA
I see this on occasion and it's not something I like. I get concerned that the temporary non-locking nut will be forgotten or missed and stay on the assembly for flight. I've seen it happen - people THINK that they're not assembling something for the last time, but then don't check it or reassemble it. My position is, you either assemble the bolt and nut with the final hardware any time there's the slightest chance that the plane will fly, or else you don't assemble it at all. No partial measures that can set up a failure.

Of course, if you're perfect, and NEVER forget anything or forget to check anything, then my point is invalid. But the people I've met that fall into that category are few and far between (me included)...

19. Mar 1, 2019

### Pops

#### Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,929
5,842
Location:
USA.
For a simple airplane it easy to go over everything and check for the red paint mark. That is also why I check other peoples work and they check mine.

MadRocketScientist and Hot Wings like this.
20. Mar 1, 2019

Joined:
Apr 4, 2007
Messages:
8,513