M6 8.8 bolt fatigue failures (HELP!!!!) 4

[trikeflyer]

I am trying to improve a bolted connection (aluminum plate bolted to threaded aluminum case) which currently uses three M6 8.8 allen head cap screws. I have had all three bolts fail in fatigue, presumably due to vibration. I suspect that the problem may be that the bearing strength of the aluminum is lower than the tensile strength of the bolt, resulting in creep of the aluminum and loss of pre-load. Unfortunately, I don't have the option of using a different size of bolt or more of them.

I am looking for suggestions. I was considering replacing the 8.8 fasteners with 10.9 hex head cap crews and a stainless steel fender washer. I realize that the 10.9 fastener will require a higher torque to reach a working pre-load. The purpose of the washer was to spread this pre-load over a greater area in order not to exceed the bearing strength of the aluminum. I'm no expert (Obviously!), so I'd really appreciate an expert opinion.

Thanks for suggestions and direction!.

Dave C.

 

[MintJulep]

The first step would be stop suspecting what the problem is and identify the problem.

From there we can help with solving it.

 

[hydtools]

What evidence do you have to allow the fatigue failure conclusion? What is the nature of the capscrew failure? Through the threads? Under the head?

The threads in the aluminum may not allow higher fastener preload without aluminum thread failure.

If you believe you are losing preload, reduce the diameter of the capscrew shank to the thread minor diameter to allow the capscrew to stretch more under preload and perhaps not unload as much if you suspect aluminum creep. Are the capscrews short compared to their diameter? Can you increase the capscrew length?


Ted

 

[JJPellin]

If there is room, add spacers to increase the free-length of the bolt. This will allow for more stretch and can prevent a loss of preload in a similar manner as the suggestion by hydtools. I would not expect creep. If there is vibration that could cause any relative motion between the two parts, it could quickly relax the strain from the bolt. The greater free length prevents this. Make sure that the mounting surfaces are smooth and have good contact. Blue check for contact between the surfaces. The greater the contact area, the lower chance of wear sufficient to relax the strain. If there is not room for spacers, consider Belville washers to prevent loss of pre-load.

Johnny Pellin

 

[trikeflyer]

Thank you for your replies. I'll try to be as clear as I can in identifying the problem.

The application is in the engine driving a pusher propeller in a small ultralight. The plate in question attaches to the crankcase of the engine. Mounted to the plate is one of two pulleys, the other pulley mounts to the crankshaft of the engine. A belt between the pulleys provides approximately 2.1:1 speed reduction. The propeller is mounted to the driven pulley. The motor develops a maximum of 27hp at 6250rpm and a peak torque of 35 Nm at 4500rpm.

The plate is approximately 23mm thick. There are three screw locations attaching the plate to the crankcase. These are arranged in an equilateral triangle approximately 120mm on a side centered on the crankshaft. There are 6mm deep recesses in the plate for the screw heads in each of the locations. In the third location a bushing is used to fill the recess, and the screw at this location captures a mounting bracket for the electric starter motor. The screws in the two recessed locations are 35mm long. The screw in the bushed location is 45mm long.

Facing the crankshaft of the motor as normally mounted, the driven pulley is directly above the crankshaft. One of the two recessed mounting screws is also directly above the crankshaft. The other recessed screw is in the lower left location. The screw that holds the starter bracket is in the lower right location.

Of three separate cases of failures that I know of the top screw failed first, the lower left failed next, and the 45mm screw failed last.

Now, to try and answer some questions:

The reason I think it is fatigue failure is because of the relatively small load and the definite vibration due to the motor and slight propeller imbalances. The static load is minimal, there will be some shear load due to the torque transmitted to the propeller, but since this is a pusher application there should be minimal strain on the screws The one failure I examined in detail showed signs of fretting corrosion under the heads of the two 35mm screws. the 45 mm screw location was bright and clean.

It appears the screws failed near the surface of the crankcase. This would also be near where the transition from shank to threads would have been. I did not recover any of the broken screw tops, so I can't say for sure. In one other failure that was related to me it was observed that the failure was in the un-threaded shank area.

Regarding increasing the pre-load, I am also concerned about the strength of the threads in the aluminum, however, the other crankcase screws which are also M6 are specified at 15Nm of torque, so I feel confident increasing the pre-load up to this level if it will help.

The capscrews are relatively long compared to their diameter, and it is possible to lengthen the two 35mm screws to 45mm as I was hoping to do. It would be difficult to increase the length of the top screw much more, as the screw would interfere with the pulley. Certainly it would seem that lengthening them would be a good idea, since in all cases I know of the longest one was the last to fail.

Belleville washers definitely seem like a good idea. Would it be appropriate to use these on top of the stainless fender washers? Again, I am wondering about the bearing strength of the aluminum.

I understand the principle behind reducing the shank diameter to the thread minor diameter. If such fasteners are commercially available in M6 x 45mm I would be willing to try this. Would I be better off with 8.8 or 10.9 grade (each with their appropriate pre-load torque)?

Here are some of my thoughts (that certainly may be misguided!)

The location where the starter bracket mounts consistently fails last. I felt that by going to 45mm screws with bushings and stainless steel fender washers in all locations I would be most closely replicating the starter mount location.

I felt that the more total pre-load I can apply to the plate, the better. Since It would be a significant change to try and use larger diameter screws (or more screws), I felt that by using a higher grade of screw properly preloaded with a torque that I have evidence that the crankcase would withstand would be the next best thing.

One more question: Should I use Loctite on these screws?

Thanks again for your quick and helpful replies.

 

[CoryPad]

Depending on the alloy, you may be critical on surface pressure and axial thread engagement for the aluminium parts. Do you know what your preload is? If you are using torque to estimate, then you need to measure or estimate the friction coefficient. Here are some hints on calculations:

faq725-536

thread725-35222

Your failure probably is fatigue. You should consider increasing preload by using a 10.9 screw. You should not consider using a hex head fastener (too small bearing area). You should consider a hex flange fastener. A waisted screw (shank diameter equal to thread minor diameter) is worth considering, but may not be available in small quantities.

 

[trikeflyer]

This is great stuff, I wish I wasn't so constrained by what I have to work with.

Estimating the thread friction coefficient and bearing friction coefficient at between 0.25 and 0.5 (steel to aluminum) I get an estimate of pre-load of the original joint with a screw torque of 10Nm of 2.7 to 5.2kN. With an equivalent 10.9 grade fastener at a tightening torque of 15Nm I get 4.04 to 7.75kN of pre-load.

I hadn't considered the axial thread engagement before. Examining the existing crankcase bolts I see that they have a thread engagement of 15mm and a tightening torque of 15Nm. If I were to use 45mm screws to mount the plate they would have a thread engagement of 22mm. As I result I feel fairly confident in using a tightening torque of 15Nm in these locations.

I am able to get M6 x 45 10.9 hex head bolts locally. I can order M6 x 45 10.9 flange bolts with a 50 piece minimum. I am not able to find waisted bolts.

Based on suggestions so far I am considering using the 10.9 bolts torqued to 15Nm with stacked Belleville washers. I would stack the washers to achieve just less than 100% deflection of the washers at 15Nm of torque on the fasteners. If I am thinking correctly, this would give me maximum retention of pre-load. Stated another way, if I were to deflect the Belleville washer 100% then I would lose some pre-load in the joint before the washers came in to play. In this scenario, does it make sense to use flange bolts? If the Belleville washer is deflected less than 100%, then it seems there is no advantage of flange bolts over hex head or allen head screws. Am I thinking correctly about this?

Thanks again for the great help and suggestions.

Dave C.

 

[CoryPad]

You should reduce the friction to ~ 0.1 using a lubricant rather than rely on 0.25 to 0.5 range. This will enable you to maximize joint preload.

I forgot to mention in the previous reply, but a thread adhesive appears to be a good idea. If you use this, you don't need a lubricant. With this, the friction coefficient likely will be between 0.15 and 0.25.

Why do you want to use Belleville washers? You need as much preload as possible and don't need to worry about extra joint surfaces that can reduce preload due to settling at the surfaces. Stacked washers is for low preload joints or joints with short clamping lengths, which doesn't apply here.

Go with the flanged screws tightened as high as possible, somewhere close to the proof load, which is 16.7 kN for an M6 with property class 10.9.

 

[trikeflyer]

Thanks, CoryPad. I'll order those 10.9 flanged screws.

Just to summarize, I'm going to bush out the recesses in the plate, use M6 x 45 hex head flanged screws with medium strength thread adhesive (blue Loctite), and I will torque the bolts to 15Nm (The highest value that I have evidence that the aluminum casting will tolerate)

Thanks again for your help.

Dave C.

 

[MikeHalloran]

If I'm understanding the geometry correctly, the 'top' bolt is reacting all of the belt pull, in tension. Which may be why it fails first.




Mike Halloran
Pembroke Pines, FL, USA

 

[trikeflyer]

Brilliant! I hadn't even considered this. Yes, I think you are understanding the geometry correctly. Tension on the belt will create flex in the plate which will be countered by tension in the top screw, thus reducing the contact pressure between the plate and crankcase in this location. Thus this location is the most susceptible to fretting, loss of pre-load and fatigue. The question, of course; is there anything I can do to counter this?

Dave C.

 

[MikeHalloran]

Longer, stronger bolt, more preload, as has been suggested.
High strength stud and nut with thick washer.
Same, necked stud.
Thicker plate.
Narrower belt and pulleys.
Smaller offset of belt and pulleys to reduce the moment.
...





Mike Halloran
Pembroke Pines, FL, USA

 

[dik]

Can you use a type of Loctite, if the problem is caused by loss of preload?

Dik

 

[trikeflyer]

Yes, I can (and will) use Loctite, although as I understand it Loctite by itself will not prevent this type of failure, since the loss of pre-load can occur without the screw ever turning.

Dave C.

 

[dik]

How are you losing the preload?

Dik

 

[trikeflyer]

I believe that the initial pre-load is insufficient to prevent relative motion due to vibration. This causes fretting corrosion which causes loss of surface material and thus further reduces the pre-load. As the pre-load drops the surfaces are more free to move putting more fatigue stress on the screws, which are then more susceptible to fatigue due to their lower loading. This continues until the screws fail.

That's how I understand it, anyway.

Dave C.

 

[dik]

Fretting corrosion caused by electrical contact? Is there a material that will isolate the two materials and also 'glue' the parts together?

Dik

 

[trikeflyer]

Quoted from The ASM Handbook on Fatigue and Fracture (as found on Wikipedia). Fretting is:

"A special wear process that occurs at the contact area between two materials under load and subject to minute relative motion by vibration or some other force."

I don't believe there needs to be an electric current or potential involved. Certainly there is not in my case.

If there are surface treatments that would help I would like to know about them. I am not aware of any. Welding the parts together or some other form of permenanent bonding that would prevent motion would presumably work, but I would prefer to keep the motor in a configuration that would allow for periodic disassembly and maintenance.

Dave C.

 

[dik]

I didn't think that that was considered as corrosion... if it is, then mechanical wear might be eliminated by using the Loctite material... For mechanical wear, there has to be movement?

Just a thought...

 

[CoryPad]

Fretting is mechanical wear, not electrochemical corrosion. A thread adhesive like Loctite only prevents fastener rotation, but the clamped joint members can move at any time when the friction forces from preload are lower than the applied external forces. Thus, fretting still can occur and result in preload loss and joint failure.