NAS6605, MS21042-5, Torque 9

[blue_lateral]

Hi, I am John, and I am new here. I have a shear bolt question.

I am using this hardware to bolt an automotive differential ring gear on (this will not fly), I just want to get the best from the hardware. The gear is hardened steel, and the flange it bolts to is something a bit softer, probably cast or nodular iron. The loading is primarily in shear.

I bought NAS6605-14 close tolerance shear bolts, and MS21042-5 stop nuts. I intend to install one washer under the nut and torque the nut. If I have measured correctly I will have one thread below the nut and two threads out the top of the nut, sticking out. I have 2 different thickness washers (AN960-516 and AN960-516L) in case a slight adjustment is needed to prevent thread bottoming. The bolts and nuts both have UNJF threads, 3A and 3B.

This all sounds perfect to me, but I am way out of my area here.

I have been poring over pages online for the past few days. Some published by the FAA, some by race car builders who regularly use aircraft hardware, some by experimental aircraft builders, and so on. I can not find a table that addresses this particular combination. I now gather that since the nuts are tension rated, they are not normally used with shear bolts, even though they are thin enough to not run out of thread.

One table I keep running across says that MS20033 through MS20046 nuts should be torqued 100-140 inch pounds, but the bolts listed do not include NAS6605. Another column lists NAS6605 but only with thin shear nuts, and specifies 60-85 inch pounds. This is all far less than I would expect even with hardware store grade 8 bolts. Another online source suggests that 160 KSI 5/16 hardware can be torqued to 330 inch pounds. Someone in another forum claculated the tension in the bolt and also came up with 330 inch pounds. I also wonder how much I am limited by the thin heads of the NAS6605 bolts. Tables seem to address only nuts.

Another complication is that I would like to use Loctite 271. This assembly will never be disassembled. It was rivited originally, but I can't seem to get it riveted in today's world. I see the MS21042 nuts are intended to be torqued dry, and I expect Loctite to behave as a lubricant.

To summarize: NAS6605-14 bolts, MS21042-5 nuts with AN960-516 or AN960-516L washers under them, Loctite 271, Loaded in shear. Not for aircraft. Will never be disassembled. How much torque?

Thank you for any light you may be able to shed on this.

John

 

[rb1957]

did you try googling "recommended fastener torque" ? grade 8 5/16" bolts … 20 ft.lbs

as a non-aircraft application, I'd suggest posting in the "mechanical engineering" forum.

another day in paradise, or is paradise one day closer ?

 

[blue_lateral]

Thanks for responding! Yes, I did and that was one of several conflicting answers I came up with before posting. Since this is aircraft hardware, rather than grade 8, I thought people would be more familiar with it here. If this thread doesn't go anywhere, I'll try over in Mechanical Engineering. Thanks for the suggestion.

I also just realized I didn't link the specs. Here they are:

http://www.meracing.com/Userfiles/Archive/1339/pdf/NAS_6603_-_NAS_6620_Series_Specifications.pdf

http://www.coastfab.com/uploads/4/1/4/3/41432063/ms21042.pdf

http://everyspec.com/AN-SPECS/AN100-AN999/download.php?spec=AN960_REV-21.027234.pdf

John

 

[Sparweb]

It's not quite the right combination for what you're doing, but it should work. Can't guarantee the "never come apart" thing, despite the loctite, when we can't see the exact configuration of what you're putting together. Presumably if the previous install had a rivet driven in it, then I hope all issues related to the clearance between bolt and its hole do not create new problems.
Tension bolt & nut combination, 5/16" UNF: 90 in-Lb

quote: "Tables seem to address only nuts." There could be a good reason for that.

Are you familiar with the difference between "running torque" and "installation torque"? The self-locking nuts put up resistance to turning (that's the point) and your installation torque is additional to that torque. The sum being what you apply with your wrench.

How come no rivets? I don't consider my little shop well equipped, but I think I could drive a few 5/16" solid rivets if push came to shove. Aluminum rivets, that is... You aren't talking about steel rivets are you?

No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
STF

 

[blue_lateral]

Yes, steel rivets.

quote: Are you familiar with the difference between "running torque" and "installation torque"? The self-locking nuts put up resistance to turning (that's the point) and your installation torque is additional to that torque. The sum being what you apply with your wrench.

Yes, in fact I bought a couple extra bolts and nuts to experiment (with a beam-type torque wrench) to see how much drag the stop nuts cause, and how much it changes with uncured Loctite 271 on the threads. I will do that and add it to the 90 in lb.

I did wonder how much the drag of the nuts on the washers would change due to the uncured Loctite (as lubricant) as the fasteners came up to torque. I couldn't think of a way to quantify that.

Originally this was assembled with steel cold-formed rivets. Hot riveting was discouraged, though I have heard it was once a common way to do this repair. I have not been able to find anyone who will rivet it, cold or hot. I have been looking for a year.

Today people typically use grade 8 internal hex capscrews. I was attracted to these bolts instead because they are close-tolerance, have a radius on the thread roots, and come in enough different grip lengths that I could keep almost all the threads out of the hole.

Thank you

John

 

[WKTaylor]

BL... am very uncertain about this particular automotive application using bolts/nuts VS originally specified rivets.

NOTE. I have to be brief... due to work constraints... hope this is intelligible...

There are Aerospace grade, 'steel', monel, A286, etc solid driven rivets that would better-suit this automotive application that could be 'squeezed' by hydraulic press for install.

I get twitchy/doubtful RE conventional NAS bolts and MS nuts used in long-term rotating applications with randomly cyclic loads in mechanical components... NOT specifically designed for these applications.. Solid driven RIVETS are 'best' for this for a number of reasons... some obvious some not-so obvious [from my helo experiences].

NOW to the Bolt/nut combo itself...

NAS6605-14 shear bolts and MS21042-5 self locking nuts are ideal for structure... but be advised, this is 'classic shear' structure.

NAS6603-6620 Bolts are intended for primarily cyclic-shear load at 95-SKSI + static torque-tensile loads for structure clamp up with some transient cyclic tension-tension above the static preload.

HOWEVER these bolts will simply NOT meet their potential tensile load capacity of 160-TKSI with an 125-TKSI-rated MS21042 Nut. Think of it this way... IF The reverse were true... IE there was a 125-TKSI rated bolt [~75-SKSI] with a 160-TKSI nut that that would appear to be a 'bad combination'... but not so. Tensile preload [induced by torque-turning] has to be chosen based on the complex of factors... related to what is the strength rating of the LOWEST strength part... so choose a torque value which induces good clamp-up tensile pre-stress for shear application.

IN GENERAL a tensile preload of ~30%-up-to-40% of the lowest rated tensile strength is common for shear joints in primary aircraft structure. So in either case where the bolt or nut is rated to 125-TKSI... then tensile preload should be 38-to-50-TKSI. Several torque-tables that I reference relate to 125-TKSI nuts... hence recommended torque-range for 5/16 [0.3125-24UNJF-3A] would be 100-to-140-in# for 'dry-cad plated parts'.

NOTE. There are 160-TKSI [or higher] tensile-fatigue rated bolts [IE: MS20005] that... when combined with a 160-TKSI [or higher] strength rated nut [IE: NAS1804-5]... can-and-should-be torque-turned to much higher tensile pre-load such as 165-to-175-in# or as high as 195-to-205-in# [carefully].

CAUTION. Torque turning is usually a combination of values: (a) free-running torque for a plain-nut to attain intended pre-stress + (b) 'running torque' which is parasite 'torque-drag' due to self-locking features, tight threads, etc... WHEN NUT IS FULLY ENGAGED on the bolt threads.

WHERE EVERYTHING GEST EVEN MUDDIER is when: (a) the threads have added lubricant or lubricant-like media... other than dry cadmium such as wet paint, sealant, oil, antiseize compound, thread locking compound, etc...; or when (b) the torque turning occurs NOT on the nut but on/thru the bolt head; and (c) if any material is inserted into the joint such as sealant or crushable gaskets, etc. There are 'special rules' to deal with these 3-factors...

HOWEVER, here is another concern of mine... when it comes to threaded fasteners in rotating machinery in-lieu-of solid driven [small/compact/simple/permanent] rivets or very low-protrusion threaded fasteners intended for rotating machinery... installed mass/inertia and PROTRUSION both sides can play unexpected roles in vibration... especially if the fastener passes thru [or close to]... or may be made 'wet' by... any lubricant [gear-oil, etc]. In helicopters all of these factors are critical for operating balance.

CAUTION any lubricant that is 'pounded by periodic mechanical protrusions'... such as bolt heads, tails/nuts... as opposed to low-profile rivet heads/tails... may under certain conditions beak-down the lubricant and/or induce 'foaming' at very high temps.

Another factor to consider is the cadmium finishes... will they be exposed the environment and/or fluids such as oil or fuel. Cadmium plating films are NOT very durable in the circumstances and tend to break-down... washing-off... exposing the steel to possible corrosive factors... and-or the cadmium ions/particles contaminating oil and fuel systems.

Just saying... I tend to over-think stuff like this...





Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[SwinnyGG]

Out of curiosity, is this for something old (Cadillac??)

In any event, you do not want to think of this as a shear bolt application. This should be re-assembled under the assumption that bolt tension serves only to maintain contact (i.e. friction) between the ring gear and carrier, and that the ring gear/carrier interface is what will carry the load.

Patterns of multiple shear bolts only work to full capacity when hole diameters and hole pattern geometry are very well controlled.

If the holes in the ring gear, carrier, or both are either not sized precisely or their positions within the pattern are imperfect, than any shear loading applied to the bolt group will not be shared equally between all fasteners. This can very easily result in a few or even only one bolt actually taking any shear load, meaning the assembly can fail very quickly.

Because rivets expand during setting to fill their holes completely, a pattern of rivets will much more reliably share shear load across the whole group- meaning that replacing rivets with rivets will always avoid this problem.

Unless you are re-machining your carrier and ring gear so that the holes are very closely matched to the shear bolt diameter (slip fit or better) and the bolt pattern is very precise to current standards, you're setting yourself up for a fall here.

You need the strongest fasteners you can get, including maximum strength nuts, and you need to torque them to 80% or so of proof load, as would be done in a modern application. If it were me, loctite and safety wire would be the order of the day.

 

[Sparweb]

Ah, yet cotter pins. Good idea!
MS14144-5 or MS14144L5
And "NAS6605D" bolts to go with them.


No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
STF

 

[Screwman1]

One of the issues related to ring gear applications that I have dealt with at the OEMs is lateral movement when the preload in the joint is inadequate to freeze the the ring against the carrier. This is a single shear joint that also sees a significant reversing load in service. The rivets work well in this joint because when squeezed they expand across the rivet body to fill the space between the rivet body and the joint to help prevent lateral movement. There have been lots of cases where hollow dowels have been used in the carrier and bolt hols that are a light press fit to prevent lateral movement.
The point of this is that you are going to need the maximum amount of clamp load possible in this joint and I would recommend the use of tensile bolts and nuts for that reason. For testing purposes you can take your two extra bolts and torque them to failure and then set your seating torque at about 75% of that failure torque and you'll be getting about 85% of the yield strength of the weakest link in the chain.

 

[WKTaylor]

jgKRI & SparWeb... good info, RE this specific automotive usage and good safetying practices in mechanical equipment...


Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[WKTaylor]

Screwman... good explanation for value of rivets in this application.

However... I'm not so-certain about Your torque-testing guidance for the specified aerospace hardware.

Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[SAITAETGrad]

Generally speaking, bolts should be sized so that grip length just exceeds the total material thickness. You then select the washer thickness required to prevent bottoming. You've effectively done this - just thought it might be a useful tip for next time.

 

[Screwman1]

Wil,
I tend to agree with you about the torque test, but if you have nothing else to go by in a "commercial" quality joint, it at least get's you into the ballpark , but it doesn't tell you one bit about the achieved clamp load. In this case, I would almost take a bunch of solid rivets, a torch and anvil and hand rivet the thing together rather than trying to bolt it unless you can get some sort of match drilling done so that there is just a slip fit between the ring gear holes and the bolt shanks. We have, and continue to do a lot of power train testing using ultrasonics on just this joint for various auto companies; it is an ongoing problem joint for them as HP levels continue to increase.

 

[blue_lateral]

Sorry I didnt get back to this sooner, I have been a bit distracted with taxes....

Out of curiosity, is this for something old (Cadillac??)

Yes, 1936 Pontiac.

This should be re-assembled under the assumption that bolt tension serves only to maintain contact (i.e. friction) between the ring gear and carrier, and that the ring gear/carrier interface is what will carry the load.

Patterns of multiple shear bolts only work to full capacity when hole diameters and hole pattern geometry are very well controlled.

If the holes in the ring gear, carrier, or both are either not sized precisely or their positions within the pattern are imperfect, than any shear loading applied to the bolt group will not be shared equally between all fasteners. This can very easily result in a few or even only one bolt actually taking any shear load, meaning the assembly can fail very quickly.

I was and am concerned about that. The holes are not perfect. The gear is extremely hard, making it difficult to do anything about that.

Because rivets expand during setting to fill their holes completely, a pattern of rivets will much more reliably share shear load across the whole group- meaning that replacing rivets with rivets will always avoid this problem.

I wanted to rivet it and still do. I have found it extremely hard to find someone who can rivet it. The automotive forums all tell me to "just bolt it" "we bolted one and it was fine", etc. etc.

I had someone lined up to rivet it, but he backed out because his tooling couldn't get close enough to the differential case. He was going to make some dies to do it, but eventually backed out, as he thought the tools might walk and chip the teeth.

Months ago, when I thought I had riveting arranged, I installed the gear with loctite under the flange (an old automotive trick to prevent motion and loosening on more modern designs that are not through-bolted and have threads in the ring gear). I had to heat the ring gear a little to slip it on, and I bolted it quickly with grade 8 hardware store bolts, torqued to their maximum allowed torque, to hold it tightly together while the loctite on the flange cured. The bolts I used had grip through the entire hole, to get the ring gear aligned correctly. They are too long to actually use.

The plan was the riveter would remove a bolt and install a rivet, one hole at a time, in a criss-cross pattern, as outlined in the shop manual.

This is how it currently stands, bolted together with too-long grade 8 bolts, torqued to their maximum.

You need the strongest fasteners you can get, including maximum strength nuts, and you need to torque them to 80% or so of proof load, as would be done in a modern application. If it were me, loctite and safety wire would be the order of the day.

Maybe I have been asking the wrong questions. Is there a firm, anywhere, who can still install steel rivets?

I am not opposed to the idea getting different fasteners if it would be a little better. Do you have any suggestions on what to get?

I initially avoided the idea of safety wire in favor of stop nuts and Loctite, because I thought the holes would weaken the fasteners. It is through bolted, so would need to be wired at both ends. I could wire it if it would be better.

 

[blue_lateral]

Ah, yet cotter pins. Good idea!
MS14144-5 or MS14144L5
And "NAS6605D" bolts to go with them.

Looking into those now, Thanks.

One of the issues related to ring gear applications that I have dealt with at the OEMs is lateral movement when the preload in the joint is inadequate to freeze the the ring against the carrier. This is a single shear joint that also sees a significant reversing load in service. The rivets work well in this joint because when squeezed they expand across the rivet body to fill the space between the rivet body and the joint to help prevent lateral movement.

Yeah. I really don't like the idea of changing the design.

For testing purposes you can take your two extra bolts and torque them to failure and then set your seating torque at about 75% of that failure torque and you'll be getting about 85% of the yield strength of the weakest link in the chain.

I'll probably try it since I have the extras.

In this case, I would almost take a bunch of solid rivets, a torch and anvil and hand rivet the thing together rather than trying to bolt it

I have an acetylene torch, and a small anvil. I have no rivet sets, and no experience hot riveting.

John

 

[blue_lateral]

NOTE. There are 160-TKSI [or higher] tensile-fatigue rated bolts [IE: MS20005] that... when combined with a 160-TKSI [or higher] strength rated nut [IE: NAS1804-5]... can-and-should-be torque-turned to much higher tensile pre-load such as 165-to-175-in# or as high as 195-to-205-in# [carefully].

Would these be a good choice?

 

[Sparweb]

When concerned about components that may move around in the bolt-hole clearance when loaded in service, there are things you can do about it, depending on how much extra material is available to work with.
The example that comes to mind is done for bearing pillow blocks that need to be disassembled and re-assembled with perfect alignment. When off-the-shelf pillow blocks are used, there aren't any features to help accomplish this, although the bolt holes might be fairly tight clearance. To be sure, though, one can align the bearing, tighten up the bolts, and then drill a new hole through the flange to install a transition-fit dowel pin or a spring pin (or called "roll pin"). The pin maintains alignment and transfers shear load, while the bolts keep the assembly held tightly together. If you do this in your application, I believe you will need enough dowel pins to transfer all of the shear load, and enough empty space on the flanges to drill.


No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
STF

 

[WKTaylor]

BL... RE You last post to me: For many reasons, I am NOT convinced the cited fastener combo would be acceptable.

In my corporate-professional-engineering capacity, I would specify appropriate strength solid/driven Steel, CRES or Monel Rivets for an application like this.

NOTE. The use of threaded fasteners to attain temporary hole-alignment and high-clamp-up prior to replacing these parts [one-at-a-time] with solid rivets is not a bad idea.

Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[blue_lateral]

Well, I assembled it with the parts mentioned in the original post.

I used WKTaylor's quoted 100-140 in/lb for dry parts, and added the turning torque. The turning torque was about 40 in/lb with new bolts and nuts,(and maybe 20 with used ones). The hardware is all new. The loctite as lubricant only made about a 5 in/lb reduction, as the cad is apparently quite slippery. I torqued them with wet loctitie to 170 1n/lb. That should leave me about 5 in/lb under the acceptable limit for these fasteners.

I tried tightening one of my extra bolts as Screwman suggested to see where it would fail. I stopped at 330 in/lb, and I could see no visible damage, but past about 300 in/lb something was clearly wrong. Either the bolt reached yield or the threads were failing.

The final assembly as expected resulted in 2 threads sticking out the end of the nuts.

WKTaylor: I too think this should be riveted. I was a whole year into failing to get that done before I ever posted here. I hope it holds.

Thank you to everyone who responded.

John

 

[hpaircraft]

...but past about 300 in/lb something was clearly wrong. Either the bolt reached yield or the threads were failing.

Been there, done that, got the NAS20142 T-shirt. These low-pro nuts are light and small, but there are definitely better choices for high tensile loads or if you need lots of clamping force. What we found in tests was that these nuts have so little hoop strength that they would actually expand and hop over threads under bolt tension. My guess is that's the wrongness you rightfully sensed as you torqued them.

Here's the VAF thread on the hardware tests that I did with my friend Steve Smith (Dr. Smith of Mythbusters fame):

http://www.vansairforce.com/community/showthread.php?t=42630

TL;DR: Use NAS1804 instead.

Thanks, Bob K.