Jet-nut life / reuse

[Rick Iverson]

Don't have an excessive # of threads exposed after a locknut is tight.

Gents;

Yep!

IIRC, Carroll Smith recommended three (3) threads exposed.

[patman]

so you pick a nut and bolt that when properly installed , creates the correct clamp load and has the correct retaining required by the application. " so here we are using jet nuts on CV joints on typical Grade 5 grade 5 bolts which are all AN bolts are. whats the torque ? whatever the A+P manual says it is which is the defined by how far you stretch the bolt in order to create the clamp load before you get near yield or its elastic limit. somebody else already did that math for you ... so whats the question ? if you changed those bolts to grade 8 then you need a different number because to stretch a grade 8 bolt takes a greater torque.
'

not disagreeing with your general sentiment, but this is NOT how (most) bolted connections work. Very few bolts are torqued into the plastic region. As long as we are in the elastic region, all steels have essentially the same stiffness, so a higher grade bolt at the same torque does not offer any additional clamping force.

When one designs a bolted connection properly, the bolts are used to apply a clamping load to the parts. Friction on the interface between parts actually holds the connection together. So if we design a connection for a 5/16" gr5 bolt torqued to 15 ft-lb, and you change that to grade 8 and torque it the same, the clamping force will be the same, the connection will be the same. The only thing you have changed is the strength of the bolt. If you want, you can torque the gr8 tighter before it will yield, but if the connection was properly designed in the first place, it won't do anything useful and will just increase the bearing stress under the head of the bolt. If you are torqueing your bolts anywhere close to their yield point, you are doing something wrong.

Personally, I use jetnuts because they are lighter than SAE all metal locknuts, and because they are a different hex size than the bolt, so I don't have to carry two sets of wrenches. I am not concerned with the higher material strength at all, as most applications I am using them for were originally designed for lower grade fasteners anyway.

Let's also keep in mind that most of our cars are not engineered to the level of aircraft either. Aircraft designers are specifically calculating all load cases for any bolted connection. Half of our cars were just bolted together with whatever size the designer had a good supply for.

As for CV loosening, red loctite seems to have eliminated that issue for me.

[problemchild]

Personally, I use jetnuts ........ because they are a different hex size than the bolt, so I don't have to carry two sets of wrenches.

Really? Five extra wrenches?
I don't consider this a persuasive consideration when selecting fasteners for any race car that I work on.

[Mike B]

not disagreeing with your general sentiment, but this is NOT how (most) bolted connections work. Very few bolts are torqued into the plastic region. As long as we are in the elastic region, all steels have essentially the same stiffness, so a higher grade bolt at the same torque does not offer any additional clamping force.

Respectfully, Steve didn't say a bolt is torqued into the plastic region. His words were "create the clamp load before you get near yield or its elastic limit" That is exactly how bolted connections work.
Totally agree with the rest of your post (except the red loctite thing... )

[DaveW]

...As for CV loosening, red loctite seems to have eliminated that issue for me.

IMO, as I've often said, use of Loctite in this type of application is not good. It will keep the nuts from rotating on the bolts, but it will not assure that everything in the assembly remains tight and under sufficient clamp load, which is what is needed to prevent failure. I never use any variety of Loctite in this sort of application.

[patman]

Respectfully, Steve didn't say a bolt is torqued into the plastic region. His words were "create the clamp load before you get near yield or its elastic limit" )

Sorry didn't mean to imply that he did. I was just stating that outside of the plastic region all steel bolts will provide the same clamping load at the same torque regardless of grade.

[patman]

IMO, as I've often said, use of Loctite in this type of application is not good. It will keep the nuts from rotating on the bolts, but it will not assure that everything in the assembly remains tight and under sufficient clamp load, which is what is needed to prevent failure. I never use any variety of Loctite in this sort of application.

Not sure where the physics of this is coming from....

Most cvs on race cars I have seen are a hardened steel flange and a hardened steel race bolted together. How exactly is that going to loosen without the bolts loosening?
6x5/16 bolts are unlikely to stretch unless you have some very strange loading. So where does the alleged reduction in clamping force come from?

Actually come to think of it, I have never built one with nuts either, just tapped holes in the drive flange so maybe you guys are doing something different than me

[DaveW]

Not sure where the physics of this is coming from....

Most cvs on race cars I have seen are a hardened steel flange and a hardened steel race bolted together. How exactly is that going to loosen without the bolts loosening?
6x5/16 bolts are unlikely to stretch unless you have some very strange loading. So where does the alleged reduction in clamping force come from?

Actually come to think of it, I have never built one with nuts either, just tapped holes in the drive flange so maybe you guys are doing something different than me

OK, to help explain the logic, here are 4 possible scenarios (Loctite vs no Loctite):

1. Fasteners torqued and Loctited to enough tension that assembly is clamped sufficiently to not have anything move or slide
2. Fasteners torqued (no Loctite) to enough tension that assembly is clamped sufficiently to not have anything move or slide
3. Fasteners are torqued and Loctited to enough tension to seem tight. However, working loads exceed clamping load, so ass'y moves or slides against itself
4. Fasteners are torqued w/o Loctite to enough tension to seem tight. However, working loads exceed clamping load, so ass'y moves or slides against itself

What is likely to happen in each case:
1. Nothing comes loose, everything performs as desired
2. Nothing comes loose, everything performs as desired
3. The ass'y slides, etc. ("works"), against itself and so begins to become loose. However, retorquing Loctited nuts/screws shows the proper torque and gives the false impression that everything is still tight, so the ass'y continues to "work" and fret until something fails from fatigue
4. The ass'y slides, etc. ("works"), against itself and so begins to become loose. In this case, retorquing non-Loctited nuts/screws shows low torque values indicating the ass'y is getting loose, warning that the ass'y has issues and needs attention

So Loctite gives a false sense of security in these situations, whereas, if there is no Loctite, you can usually detect the issue before it causes a failure.

[billtebbutt]

With truly no disrespect intended to any posters here, 47 posts on a 20 cent part (28 cents Cdn )?

Are we trying to count how many angels can dance on the head of a pin?

cheers,
BT

[tstrong]

Not sure where the physics of this is coming from....

Most cvs on race cars I have seen are a hardened steel flange and a hardened steel race bolted together. How exactly is that going to loosen without the bolts loosening?
6x5/16 bolts are unlikely to stretch unless you have some very strange loading. So where does the alleged reduction in clamping force come from?

Actually come to think of it, I have never built one with nuts either, just tapped holes in the drive flange so maybe you guys are doing something different than me

Patman your points are all valid but are missing the specific most common failure mode of this particular joint when the CV is through bolted most commonly with AN5 bolts and MS21042 Jetnuts. The max recommended torque is around 18 lb-ft and result in a clamping force that is right on the ragged edge of being able to prevent relative motion between the cv and drive flange. As to your comment about how the joint loosens it is typically a combination of deformation of the dead soft cv boot flange and/or the common use an960 washers. If as Dave discusses you use hardened (small diameter) washers you can deform the boot flange pretty quickly to the point the joint preload stabilizes at just an acceptable level to prevent relative motion (all things being clean and perfect). What really benefits the joint is more clamping force developed by additional torque on the nuts. The problem is the an5 bolt will not accept much additional torque without plastic deformation and eventually stripped off threads. So the common solution is to go to a 160-180ksi bolt to match the strength of the jetnut which can then be reliably torqued to around 25lb-ft. This additional clamping load prevents the relative motion of the joint and eliminates the loosening issue without the use of locktite which indeed can lead to a false sense of security in this application.

[patman]

Actually I was just checking that very thing this morning...

Clamping force from a 5/16 bolt at 18ft-lb is approx 3450lb. So a 6-bolt CV would have almost 21000lb of clamping force.

Assuming typical dry steel/steel friction coefficient of 0.2 that should provide about 6000 ft-lb of torque bearing capacity. Maybe on a FA or something this would be the ragged edge, but most of our cars are only making a maximum of ~1200 ft-lb of torque at the CV, and that's even assuming all the torque going to one wheel.

In all honesty I don't know exactly what kind of shock loading would by typical, but I would think that the 5:1 safety factor would be sufficient.

So I am back to the conclusion that the most likely failure is not insufficent clamp load as per DaveW's scenarios, but more likely vibration loosening of the bolt. I have never had a CV come loose, but before I started using loctite, I did paint mark them (still do), and was able to confirm that the bolts were working loose. As mentioned all theoretical discussion aside this has eliminated any loosening for me on multiple cars.


I did not think of the soft boot flange thing, I agree that that could be an issue such as you guys are describing. But tightening the bolts more would just exacerbate that problem. It would be better just to use a hardened washer or backup flange to reduce the bearing load.

[DaveW]

I think the most likely loosening cause is shock/impact torque loading as the rear wheel skips over bumps/curbs (either acceleration or braking) or non-perfect dog engagement during clutchless shifting.

In any case, as long as you are not having issues, it's all good.

[jrh3]

Anyone using AN castle nuts and cotter pins?

[jchracer]

Having worked on bolted joints in the aerospace field for some time, we have come to the conclusion that friction is never your friend and relying on an assumed friction coefficient for critical bolted joint applications is risky at best. If you are experiencing repeated issues and really want to solve the problem, I would recommend using one of the modern torque wrenches that measure both torque and angle-of-turn (AOT) simultaneously. You could then "over torque" a few nuts to determine the AOT that results in ~85% of the yield strength if the bolted joint. You could then use use a quality anti-seize on the threads without risking thread damage or inadequate preload.

[DaveW]

... You could then use use a quality anti-seize on the threads without risking thread damage or inadequate preload.

I use some form of anti-seize on almost every joint that requires a bolt/nut/screw tension that is critical in preloading the joint properly. IMO, there is just too much variation in clamping force vs torque to have the dry torque value be a reliable indicator of tension. To rely on dry torque values to indicate a tension value, one needs to know the exact surface conditions involved, and unless one is working in precisely-controlled situations, that is not likely.

[jchracer]

I use some form of anti-seize on almost every joint that requires a bolt/nut/screw tension that is critical in preloading the joint properly. IMO, there is just too much variation in clamping force vs torque to have the dry torque value be a reliable indicator of tension. To rely on dry torque values to indicate a tension value, one needs to know the exact surface conditions involved, and unless one is working in precisely-controlled situations, that is not likely.

Agreed. I have data on coefficient of friction for coated nut/bolts that range from .03 to over .20 with dry showing even higher variability. That is why torque is such a lousy measurement of clamp load/bolt tension and AOT is used instead for critical aerospace applications. AOT takes the friction out of the equation.

[DaveW]

...AOT takes the friction out of the equation.

That, knowing modulus, area, and threads per inch, relates directly to the stress and tension force or preload. Much better than guessing friction coefficient.

[stonebridge20]

I use some form of anti-seize on almost every joint that requires a bolt/nut/screw tension that is critical in preloading the joint properly. IMO, there is just too much variation in clamping force vs torque to have the dry torque value be a reliable indicator of tension. To rely on dry torque values to indicate a tension value, one needs to know the exact surface conditions involved, and unless one is working in precisely-controlled situations, that is not likely.

This ^

[DaveW]

...Clamping force from a 5/16 bolt at 18ft-lb is approx 3450lb. So a 6-bolt CV would have almost 21000lb of clamping force.

Assuming typical dry steel/steel friction coefficient of 0.2 that should provide about 6000 ft-lb of torque bearing capacity. Maybe on a FA or something this would be the ragged edge, but most of our cars are only making a maximum of ~1200 ft-lb of torque at the CV, and that's even assuming all the torque going to one wheel.

In all honesty I don't know exactly what kind of shock loading would by typical, but I would think that the 5:1 safety factor would be sufficient...

Actually the torque capability in lb-ft resulting from that clamp force x radius x CF = torque. I.e., 21000 x (1.53/12) x 0.2 = 536 lb-ft.

So that is about 1/3 the engine-applied torque if there are no spike loads from rough shifting, wheel hop over bumps, etc.

That is still insufficient, and everything needs to be perfect to get that. So contact with the CV bolts does most of the work.

See the next 2 posts - error corrected.

[Lee Johnson]

Actually the torque capability in lb-ft resulting from that clamp force x radius = torque. I.e., 21000 x (1.25/12) = 2188 lb-ft.

So that is about a 1.8:1 safety factor to the engine-applied torque if there are no spike loads from rough shifting, wheel hop over bumps, etc.

So that is still marginal, and everything needs to be perfect to get that.

Dave -- I think your correction to the predicted torque capability brings in a needed factor (radius) but I don't understand why you apparently did not include the previously mentioned dry steel-to-steel coefficient of friction (the value 0.2 in patman's post of 09.06.22, 8:02 AM seems like a good rule of thumb value).

Including that factor, the equation (with clamping induced friction only) seems to me to be ---

torque capability = clamping force [lb.] x coeff. of friction x radius [ft.] = 21,000 x 0.2 x (1.535/12) = 556.5 [ft.-lb.].

This torque capability result assumes-
a) effective friction "radius of action" is @ the radius of a 78mm (3.07 inch) bolt circle diameter CV,
b) only one wheel-side of the car.
c) takes no account of shear resistance across the area of 6 each 5/16 diameter steel bolts (0.45 sq. in. for 6 bolts that could come into play once a bit of slippage eliminates CV hole to bolt shank clearance).

I don't know enough about various engine torque outputs, gear ratios, ring & pinion ratios, side-to-side torque splitting or tire torque capabilities to develop a safety factor with any confidence.

Lee

[DaveW]

Of course you are correct - I had a brain fart and didn't go far enough in my correction. And your added details are of course necessary to get the whole picture.

I left out the shear strength of the bolts even though they are the limiting factor in how far everything can move, but with alternating torque loads things will still move (there is always clearance to be able to put things together) and fret/fatigue until the ass'y starts getting loose and eventually fails unless the bolt tension is high enough.

Experience and logic says more clamping force is better, so ~18-20 lb-ft on each fastener seems to be the magic number.

Thank you.

See the following post.

[DaveW]

Of course you are correct - I had a brain fart and didn't go far enough in my correction. And your added details are of course necessary to get the whole picture.

I left out the shear strength of the bolts even though they are the limiting factor in how far everything can move, but with alternating torque loads things will still move (there is always clearance to be able to put things together) and fret/fatigue until the ass'y starts getting loose and eventually fails.

Experience and logic says more clamping force is better, so ~18-20 lb-ft on each fastener seems to be the magic number.

Thank you.

A related fact - the way this works:

Similar to the wheel drive pins on a center-lock wheel, CV's contact surfaces are most of the time going to slip to the limit provided by the metal-metal contact of the drive pins, or, in the case of CV's, to contact with the attachment bolts. If the contact friction (bolt tension) is high enough, they will remain permanently in the direction of the greatest torque applied by the wheel or half-shaft due to rear wheel or front wheel hop over bumps. If they are not tight enough, reverse torque loading will make them slip in the reverse direction, and this process will repeat until something fails or comes loose.

So the function of sufficient bolt tension is to lock the unit in the direction of maximum applied torque, and keep it there, preventing loosening or other failure.

[Lee Johnson]

A related fact - the way this works:

Similar to the wheel drive pins on a center-lock wheel, CV's contact surfaces are most of the time going to slip to the limit provided by the metal-metal contact of the drive pins, or, in the case of CV's, to contact with the attachment bolts. If the contact friction (bolt tension) is high enough,they will remain permanently in the direction of the greatest torque applied by the wheel or half-shaft due to rear wheel or front wheel hop over bumps. If they are not tight enough, reverse torque loading will make them slip in the reverse direction, and this process will repeat until something fails or comes loose.

So the function of sufficient bolt tension is to lock the unit in the direction of maximum applied torque, and keep it there, preventing loosening or other failure.

For many (most, now-a-days?) amateur class formula cars, i.e., those without in-board rear brakes, the largest torque loads thru the CVs will be forward drive in lower gears. This should result in the stable (while fastening bolt pre-load remains high) condition of Dave's described condition, that is, torque loading against the fastener bolts' shear capability plus clamping induced friction.

Lee

[jrh3]

[Larry H]

I did a rough back of the envelop calc and the connection has very little margin under worst case condition, even with 5/16 bolts stressed up to 180ksi. I assumed the worst case load would be generated by 3600 lbs of grip at one rear tire. 3600 lbs = 3g x weight of car @ 1200 lbs and a bolted connection surface coefficient of .2.

I found an interesting article on coefficient of friction. Where the paper claims the following:

The mean values of the dynamic coefficients of friction of the combinations in seawater are: “Rusted steel (D) / Rusted steel (D)” between 0.42 and 0.73; “Coated steel / Coated steel” between 0.14 and 0.24 with an influence of the amount of contact pressure.

chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.sintef.no/globalassets/project/eera-deepwind-2020/posters/poster_richard-pijpers.pdf

To be clear I'm sorta of throwing this up as tongue in cheek thought.

Larry

[DaveW]

That's exactly why the wheel pins or CV bolts are stressed in shear at the highest torque. The friction is what prevents alternating motion and loosening.

[Rick Kirchner]

but 3g on a single rear tire is probably more than 2x safety margin. IIRC the newer cars can get 3 but I've rarely seen more than 2.5, and that's with helpful track topography, high speed, and all 4 tires....

[Larry H]

3 gs seem excessive, I agree.

Does everyone run the rubber boot with the flanges? There was mention of this in this thread and the softness of the material. Maybe it would be good to have the pre-load by-pass the flange and directly bolt on to the hardened surface of the universal joint. Not sure if this can be done.

[DaveW]

3 gs seem excessive, I agree.

Does everyone run the rubber boot with the flanges? There was mention of this in this thread and the softness of the material. Maybe it would be good to have the pre-load by-pass the flange and directly bolt on to the hardened surface of the universal joint. Not sure if this can be done.

That's what the hardened 2-hole "washers" are for, but there is no room in my installation for those.

In my installation, I use 5/16" hardened "high" lock washers under the Jet nuts. These don't have a lot of contact area, but the nut clearance to the cover is so marginal that the only way I can get the Jet nuts to seat squarely on their end is to have these washers under them. This does dig into the cover metal, but that does not cause any issues with loosening or anything else.

Photo below: