Strength of Bolt vs. Strength of Nut

[jondon]

I'm designing a structure which requires high-strength stainless steel bolts. I have spec'd out 1" diameter ASTM A193 B8 Class 2 bolts, with a yield stress of 80 ksi and an ultimate stress of 115 ksi.

I'm now trying to spec out nuts to go with these bolts, and the information I'm getting from our supplier seems to run contrary to good engineering practice. Our supplier (as well as several sources I've been able to find online) recommends ASTM A194 Grade 8 nuts to be paired with ASTM A193 B8 Class 2 bolts. These nuts have a proof load stress of 80 ksi.

Everything I've read regarding bolt & nut design states that the proof load stress of the nut to be used should be greater than the ultimate tensile strength of the bolt, so that the bolt itself will fail before the nut threads will strip. This would seem to indicate that the proof load stress of the nut in this case should be greater than 115 ksi.

Our supplier does have higher-strength strain-hardened A194 nuts available, but they are nearly 4x as expensive as the basic Grade 8 nuts, and I don't want to make this assembly unnecessarily expensive.

I've spoken with both our supplier and their bolt manufacturer - they both insist that the basic A194 Gr 8 nuts are the right way to go for the bolts I'm using, but I'm still skeptical given the lower proof stress of the nut. Is this an acceptable combination of materials, or is it necessary to go with the higher-strength nuts in this case?

 

[racookpe1978]

In this single particular case, the threaded piece (the bolt) is as easy to replace as the nut. Thus, either nut or bolt could be stronger (not strip) when the other piece is destroyed.

Your supplier is right in the general sense, particularly if the threaded piece (the male threads) is an engine stud or machined part. Then, the "replaceable" (cheap) piece is the nut. Opposite if a tapped hole is used in a casting or machined piece: Then, the female threaded item is the expensive part, and you want the male threads to be destroyed if excessive force is enthusiastically overapplied applied by a nitwit.

 

[desertfox]

hi jondon

Well even if the nuts and bolts have the same yield strength, the nut will have a larger shear area than the bolt so I don't think you will have a problem.

desertfox

 

[Screwman1]

The yield strength that needs to be used is the part strength, not the material strength, so you need to determine the actual yield of the parts in question. Fox is right about the larger shear on the nut thread.

 

[jondon]

Thanks for the responses, guys!

I see the point about comparing the total part strength rather than just the material strengths. I think some of my confusion came about due to the wording in several of the references I found which seemed to imply that a simple comparison of the proof load stress of the nut material and the ultimate stress of the bolt material would be sufficient.

I've gone back and calculated the failure load in the bolt versus the proof load in the nut based on the equations referenced here. Assuming that both the nut and bolt are UNC thread (1", 8 threads per inch), and assuming that the total thread engagement depth is equal to the depth of a standard 1" hex nut (0.83"), I get a failure load on the bolt of 70,378 lb, and a proof load on the nut of 95,948 lb. This, of course, shows that the standard A194 Grade 8 nut is acceptable for use with the A193 Gr 8 Class 2 bolt. Can someone confirm that my assumptions are correct in these calcs?

I have two concerns with these assumptions:

- Can the thread engagement depth be assumed to be the full depth of the nut? It looks like a minimum thread engagement length of about 0.62" would still result in a bolt proof load that matches the failure load in the bolt, but I don't know how much of the nut depth I can account for in my calculations. I'm aware of the the rule of thumb for nut depth as 80% of the bolt nominal diameter, but I'm just curious as to how much of the nut is considered to have effective thread engagement.

- Do I need to account for uneven distribution of shear loads on the nut threads? I know that the first thread takes more of the total load than subsequent threads, so it would seem that the average shear stress calculations in the above link would be non-conservative.

 

[desertfox]

Hi

Can't see a link?

Check out the roymech site on fasteners

 

[jondon]

Desertfox, the roymech site is the one I used and was trying to link: [URL unfurl="true"]http://www.roymech.co.uk/Useful_Tables/Screws/Thread_Calcs.html[/url].

Not sure why the link didn't show up.

 

[jondon]

Oh, and one other assumption I'd like to have clarified: is the proof load stress specified for a nut a shear stress value, or a tensile stress value? I would assume that this is a shear stress value in the threads.

I only ask because typically if a material yield stress (σy) is specified for an isotropic material, it is usually a tensile stress, and the allowable shear stress can be calculated based on the distortion energy theory as σy/√3. If the proof stress in the nut is actually a tensile stress value for the material, then the allowable shear stress in the threads would be only 46,188 psi, and the nut proof load would be reduced to 55,395 lb, making it unsuitable for use with the high-strength bolt.

Perhaps I'm over-thinking all of this, but I'd just like to make sure I have a good understanding of this subject.

 

[bcd]

Hi. Strain hardend bolts or studs with regular nuts is the norm for stainless steel. Most common way of validating this is OK is to compared the minimum required proof load values in ASTM A194 to the strength of the bolt or stud in tension. You will find that the bolt or stud breaks in tension well before the threads shear in the nut. The proof load is a test done on each manufacturing lot, so you have some verification beyond your calculations.

 

[CoryPad]

The recommended nut for bolts according to ASTM A193 Grade B8 Class 2 is ASTM A194 Grade 8. See here:

http://www.portlandbolt.com/technicalinformation/astm/ASTM_A193.html

The nut proof load is the applied tensile force in the mating externally threaded member. Testing is according to ASTM A962.

The engagement length can be considered the nut height minus the chamfers.

You do not need to account for unequal stress distribution through the threads.

 

[jondon]

Thanks again for all your responses. CoryPad and BCD, I wanted to verify your answers for myself, so I tried to dig up copies of the ASTM A962 and ASTM A194 specs.

I was finally able to find a copy of ASTM A194, and the spec lists a proof load for a 1" Gr 8 heavy hex nut of only 48,480 lb, which is significantly lower than the calculated ultimate load on the bolt (70,378 lb). The spec says that this proof load is based on a "proof stress" of 80,000 psi, which corresponds to the proof load stress listed in various other references.

I'm now confused as to what exactly this "proof stress" is referring to. When I calculate the proof load directly based on the shear area of the threads and the listed proof stress, I get a proof load of 95,948 lb - almost double the value listed in the spec. As it stands, I can see two possible explanations for this discrepancy:

1. The "proof stress" referred to in the spec is not a stress measured in the nut threads, but rather a tensile stress measured in the testing mandrel. The size of the mandrel is known, and the applied load is back-calculated based on the mandrel stress.

2. The "proof stress" in the spec is a stress in the threads, but it is defined as a tensile stress in the material, and we therefore have to apply the distortion energy theory to get the allowable shear stress: σy/√3 = 80,000/√3 = 46,188 psi. If we calculate the proof load using this new allowable shear stress, and account for the chamfers on the nut to get a more realistic effective thread engagement, we end up with a proof load closer to what is shown in the spec (0.73" total thread engagement length with max allowable shear stress of 46,188 psi gives a proof load of 48,662 lb).

Can someone help shed some light on this discrepancy? The data provided in the spec would seem to directly contradict BCD's assertion that a standard Grade 8 nut is acceptable for use on a B8 Grade 2 bolt. I'm hoping I'm just reading or interpreting something incorrectly.

 

[Screwman1]

Proof load is a quality check not a material property. It is a load set at about 92% of the minimum yield and the fastener must withstand that proof load without nay permanent deformation. the intent is to 'prove" that parts have at least a minimum load capability. Proof and yield are not the same thing.

 

[jondon]

I understand that proof and yield are not the same - I'm primarily confused as to the discrepancy between the listed "proof stress" and the applied "proof load" in the A194 spec. If I use the listed "proof stress" to back-calculate the "proof load" based on the shear area of the nut threads, I get a "proof load" value nearly double what is listed in the spec. This makes me think I may be applying the "proof stress" incorrectly in my calculations for the reasons listed in my last post.

My main concern is that the spec lists the proof load for the nut I am interested in at 48,480 lb, while the breaking load on the bolt I am interested in is 70,378 lb. This would indicate that the nut threads will strip out well before the bolt breaks, which runs contrary to good engineering practice as I understand it.

 

[desertfox]

Hi

Try this link

http://books.google.co.uk/books?id=...how to achieve proof stress on a bolt&f=false

 

[jondon]

Thank you, desertfox, that reference had what I was looking for: "The [proof] load is equal to the proof stress multiplied by the tensile area of the same size external thread". So the proof stress is not the shear stress in the nut threads under the applied proof load, but rather the tensile stress seen in the corresponding bolt under the applied proof load.

While this clears up the data listed in the A194 spec, it still doesn't quite address the problem of the nut proof load versus the bolt ultimate load. As I've mentioned in previous posts, all literature I've found (including the handbook linked by desertfox above) indicates that the proof load of the nut should be greater than the breaking load of the bolt, to avoid the potentially disastrous issue of stripping the nut threads during installation.

As BCD has mentioned, it is typical to use normal stainless nuts with strain-hardened bolts. Why is this an acceptable practice, when the proof load of the normal stainless nut is more than 20,000 lb less than the breaking load of the bolt?

 

[desertfox]

Hi Jondon

It's not always required to tighten a fastener to its strength capacity, the preload you put on the joint should be suitable for the materials your clamping because if you preload to the fastener capacity you might well yield the clamped components.
I think screwman mentioned this earlier.

 

[CoryPad]

If you want to have the nut proof stress to be equal to the bolt ultimate tensile stress, invoke supplementary requirement S1 from ASTM A194.

Regarding your question about acceptable practice, there are many factors involved. The A193/A194 parts (especially Grade B8) have a lot of applications in the boiler and pressure vessel arena instead of the typical industrial usages for other fasteners. The elevated temperatures, joint configurations (e.g. flanges with studs and nuts), tightening methods (hydraulic vs. wrench), and many other items have brought the industry to its current condition.

 

[tbuelna]

Most good quality commercial bolts use rolled threads and have a body diameter that is slightly greater than the thread root diameter. When nuts are installed on these bolts there will always be at least one bolt thread pitch beyond the nut face. When a tension load is applied to the bolt, there is a large stress concentration created at the root of this first bolt thread pitch. The bolt will yield at the location of this first thread pitch root fillet long before any other part of the bolt or mating nut begins to yield. The stress level that causes the bolt to yield at this point is usually a combination of preload tension and torsional shear from friction at the thread contacts. The nut also is subject to some radial hoop stress created by the 30deg pressure angle at the thread flank contacts.

 

[desertfox]

Hi Jondon

Another way of looking at this bolt breaking before nut stripping is to consider what happens when you use a high strength fixing into a tapped hole whose material strength is inferior to that of the bolt.
Under the above circumstance you can due several things to avoid the tapped hole from stripping:-

1/ Increase the thread engagement of the bolt (but after about 1.5 x bolt dia engagement there is little advantage).

2/ Use a helicoil insert.

3/ Take advantage of the increased shear stress area of the tapped hole and choose a bolt size which can easily
handle the preload.

Basically what I am saying is that, if you can design a joint in which the bolt fails, then that is preferable however given certain materials, space constraints etc. its not always possible.