14 replies to this topic

[mariner]

Nearly stripping a UNC thread in my Muncie gearbox alloy case started me thinking - mainly because the threaded hole isn't blind, it goes right through the case . So cleaning up the thread or an insert would require stripping down the whole box which is way beyond my skills.

 

So I just scraped out any chips VERY carefully with a fine pick  and shortened the bolt to only go into the undamaged part of the thread.- then I lockwired them to be safe!

 

The shortened bolt only has a few turns of thread but I recall seeing somewhere a formula which aid something like " you only need 1 1/2 times the bolt  diameter of thread to get maximum grip."

 

I think it is saying that since the thread length is pi D times the diameter and all the stress is in the top  part of  the engagement .

 

Has anybody else heard of that rule?

 

[Bikr7549]

If the shear strength of both the bolt material and the case material are the same then this assumption is roughly correct, in terms of shearing either the male or female threads. The ratio of the internal thread shear area to the external thread shear area is just about 1.5. If material strengths are different then the shear strength of the threaded joint will vary with that ratio applied to the shear area ratio. The typically low strength of cast non-ferrous material is one of the numerous reasons why threaded inserts provide an advantage, they increase the shear area of the female thread, offsetting the lower material strength. Sometimes you see high strength bolts threaded into an aluminum housing-generally this provides no real value except for wasting money.

Edited by Bikr7549, 01 March 2021 - 15:06.

[Canuck]

I used to make a product that had a pattern of 12 x  1/2"-13 blind holes, 1.75" deep in really knarly material. It was awful stuff to machine in general with really hard spots interspersed with really gummy spots. It was not unusual to have a double-digit scrap rate with a new vendor and more than one raised the 1.5x concept. The whole thing was kind of a nightmare actually.

[Bikr7549]

The load distribution along the length of an internal thread is not linear, far from it. A lot depends on the stiffness of the materials used, but an often seen analysis indicates that the first two engaged threads carry about half the load, for like stiffness materials. 'Engaged threads' means actual threads-be sure to leave out the chamfers at the start of the bolt and the hole entrance, as well as the thread relief under the bolt head-its not simply the length of the bolt

 

Check this reference out-about 1/3' rd the way down is a table on this issue  https://engineerdog....bout-fasteners/

 

There are often reasons beyond strength to have engagements over 1.5:1, but as mentioned by Canuck there can be problems in doing this-simply getting the threads to mate up for long lengths can be a problem at times unless the threads (internal or external) are varied from nominal size to provide more clearance. 

Edited by Bikr7549, 01 March 2021 - 15:59.

[Greg Locock]

This all rather changes if the bolt is a dissimilar metal to the nut, as i suspect is the case here.

[Keith Rolleston]

See post 2, line 3 where that case is covered.

[mariner]

Thanks for the info. GM being GM they use big bolts with UNC threads into  the alloy casting even though the loads are quite low. I think they reckoned it was worth it to get plenty of " meat" on the alloy female thread.

 

Funnily I think one of the best people at understanding threads is IKEA. Most of what they use is low grade particle board which is brittle so they deploy lots of tricks to stop it breaking. The round bolt things in the holes is one way but I bought a table where the legs screwed straight into the  particle board sides. The screws were thick with a really coarse thread so you could screw the leg on so tight it stays rigid. Hasn't broken yet in 5 years so I guess they did the sums right.

[Fat Boy]

I used to make a product that had a pattern of 12 x  1/2"-13 blind holes, 1.75" deep in really knarly material. It was awful stuff to machine in general with really hard spots interspersed with really gummy spots. It was not unusual to have a double-digit scrap rate with a new vendor and more than one raised the 1.5x concept. The whole thing was kind of a nightmare actually.

 

I'm not going to say anything if you spell 'color' with a 'u', but that is not the proper spelling of the word 'gnarly'.

[Canuck]

I'm not going to say anything if you spell 'color' with a 'u', but that is not the proper spelling of the word 'gnarly'.

Having never lived in a place where the use of gnarly was not an affectation (Great Slave Lake is a long way from the beaches of California), I will defer to the local expert and go back to my gnarly gnurling.

[Fat Boy]

Having never lived in a place where the use of gnarly was not an affectation (Great Slave Lake is a long way from the beaches of California), I will defer to the local expert and go back to my gnarly gnurling.

1. I won't steer you wrong.

2. Thank you for not apologizing somewhere in the response.

[Sisyphus]

The load distribution along the length of an internal thread is not linear, far from it. A lot depends on the stiffness of the materials used, but an often seen analysis indicates that the first two engaged threads carry about half the load, for like stiffness materials. 'Engaged threads' means actual threads-be sure to leave out the chamfers at the start of the bolt and the hole entrance, as well as the thread relief under the bolt head-its not simply the length of the bolt

 

Check this reference out-about 1/3' rd the way down is a table on this issue  https://engineerdog....bout-fasteners/

 

There are often reasons beyond strength to have engagements over 1.5:1, but as mentioned by Canuck there can be problems in doing this-simply getting the threads to mate up for long lengths can be a problem at times unless the threads (internal or external) are varied from nominal size to provide more clearance. 

True that the load in the threads varies significantly from thread to thread but I disagree with the table in the "10 tricks" reference that the max load is in the first thread.  Actually, photo elasticity and FEA modeling show the second thread typically carries the max load.  The first thread starts out with zero load and picks up load as you go up the ramp and the first thread gets engaged with its mate.  Whether this is important or not depends on what you're doing.  If the female threads are in a lower modulus material (like a steel bolt in an aluminum housing) then the load per thread curve flattens out with the threads picking up a more equal load.

 

1.5 times the nominal diameter is a good rule if you can do it but I've seen some critical joints with only 1 diameter (actually 6 full threads).  Gave me the willies but worked as long as you have good threads.

 

Measuring thread stretch directly is the best practical means of ensuring proper load, in my experience.  Turn of the nut is pretty good as it is also measuring the bolt stretch (but it doesn't directly measure the compression of the clamped parts).  Measuring torque, even with an electronic torque wrench is dodgy in my experience, and measuring torque with a manual wrench is really prone to error.  Probably no better than an experienced person doing it by feel.

[Kelpiecross]

 i may be slightly uneducated bolt etc.-wise  but I personally would like to see at least  3 times the the bolt diameter and maybe more for a really tight clamping.    

[mariner]

It just so happens that ARP have done an article on threads and fasteners. Partly a sales pitch but they should know what they are talking about .

 

https://www.enginela...reads-in-depth/

[Bikr7549]

Another good bolted joint reference is the "Bolt Science" website-their FAQ section has some interesting articles with lots of info. I have never been able unfortunately to get into one of their seminars, or to buy the software they offer-both look excellent.

 

https://www.boltscience.com/

[Greg Locock]

I've done the bolt science course, and twenty years (?) later have forgotten it all, but it was very good although it doesn't go much beyond what's already on their website.

Edited by Greg Locock, 05 March 2021 - 02:42.