Necessity for Thread Reliefs - And Comments on thread relief issue 2

[johnchrc]

Are thread reliefs still necessary when the parts will be made on CNC machines?

I have been allowing threads to run out to optimize strength through threaded connections instead of using thread reliefs. What are the manufacturing implications?

I have a case with an SA Box where I could make min full thread 2" and run out to 3" (Linear from face). See attached pdf. Instead, I gave into pressure to add a thread relief at 3" with 1/4" length. The part still has the tensile strength through the relief (just barely) so I allowed it. (1)Should I have made the change considering this will probably never be made on a manual lathe? (2)Would a machinist on a manual have a problem pulling out with 1" before shoulder?

One last question about design on pdf... now I have a case where there is a stress concentration at the relief corner. How hard would it be to run the 45 degree face all the way to the thread relief ID and would a radius larger than .03" make it harder?

I would appreciate any comments. Thanks.

 

[MikeHalloran]

I'd give the relief pocket a larger corner radius, at least at the end near the 45 degree face. I wouldn't try to make them contiguous unless they're generated by the same tool.

But first, I'd take a hard look at milling the threads, in which case I don't think you need the relief pocket at all.







Mike Halloran
Pembroke Pines, FL, USA

 

[hydtools]

You allow 1" thread runout without a thread relief. Give them at least a 1/2" wide thread relief. Make the inside corners of the relief with as large a radius as you can, maybe 2x the thread depth. Like Mike suggests, take the 45deg on the right up to the thread relief and eliminate the sharp corner. Add a 45deg chamfer to the left edge of the thread minor diameter to remove the razor edge of the thread runout. Look at the thread relief as removing material that does not need to be threaded to reduce thread cutting effort from a manufacturing point of view.

Somewhat concerned with you comment that the tread relief leaves barely enough material for tensile strength. Reduce stress risers as much as possible.

Ted

 

[johnchrc]

Thnanks for the feedback.

I always thought the thread relief was to allow machinist time to stop insert moving foward. If programming without a thread relief, would you taper out after defined perfect thread or pull out at some point past that at same distacne every time?

As for feathered edges, our default is 30-45 on entrty and exit bevels or a note the Block first thread to remove that incomplete thread.

I like running the 45 to the relief diameter. I wasn't sure what radius would be on the insert and the complications a larger raqdius would have. I just want to know. I know they can put the radius I want in I just don't want to understand how manufacturing will do this and what complications the change would make and why.

 

[berkshire]

johnchr
Wrote

"I always thought the thread relief was to allow machinist time to stop insert moving foward"

It is put there to allow the cutting tool to rotate around the shaft as the lead screw is disengaged.
B.E.

The good engineer does not need to memorize every formula; he just needs to know where he can find them when he needs them. Old professor

 

[johnchrc]

Berkshire,

Isn't that the same thing. A profile is blanked out so a machinist has time to disengage the lead-screw and keep it from moving forward into shoulder or surface.

Is the addition of thread reliefs really a Manual lathe practice... not a CNC practice/requirement?

I realize most connections require a relief to accept the mating thread. However, if running on CNC, why would a relief be required if the mating connection will make-up?

Do you not gain a tensile area advantage by running thread out and eliminating relief?

Is it bad practice to run threads out where possible instead of adding reliefs?

 

[MikeHalloran]

On a manual lathe, you never disengage the lead-screw until the thread is completed, because of the difficulty of re-engaging with the original relation between spindle and lead screw rotation.

Again on a manual lathe, big threads, especially Acme threads, require multiple passes, so you run the tool in, stop the spindle, reverse the spindle to withdraw the tool, then feed the tool radially a bit, and repeat until you reach full depth.

While doing all of that, you have to be careful to stop the spindle for each succeeding pass just a little before the point where you stopped the previous pass, so you don't accidentally double the depth of cut and break the tool.

The relief pocket makes that care about exactly where you stop unnecessary; just run the spindle until the tool starts cutting air in the relief pocket, stop the spindle, and back out without disengaging the half-nuts.



Mike Halloran
Pembroke Pines, FL, USA

 

[racookpe1978]

Seems like you were (originally) proposing also to thread an extra 1 inch on every part... Wouldn't reducing that excess length of thread be a real savings in time-per-part?

 

[johnchrc]

racookpe78,

No, I was discussing the part in the first attachment before the change. The end of thread had 1" to the shoulder so there was plenty of room to taper out. To add the relief, on this particular part, the thread has to be extended to 3"-thd relief length.

I made a quick example of the callout in "No Relief" attachment. A relief would reduce the tensile capacity of the connection.

 

[hydtools]

The reduction in tensile capacity would be very slight. The thread relief diameter would need to be only slightly larger than the thread major diameter. The tensile capacity would be slightly less than that at the thread major diameter wall thickness.

Ted

 

[johnchrc]

Thanks you for all the feedback.

hydtools,

The thread major (I assume your talking about a box) is in a helix so the cross section is larger than the difference between the major diameter and the relief. If you have a coarse pitch thread the area gained can be a fair amount. On a fine pitch thread this advantage drop.

If your working to bring a design in under the required limits it can make a difference. It could be the difference when you have very little wall thickness to work with.

I still believe that if your running parts on CNC, or as someone mentioned milling the threads, why not leave relief out unless connection requires to make-up.

 

[berkshire]

johnchrc
Yes essentialy : Depending on the lathe you would Make the screw cut, stop at the cut out, back the tool out, take it back to the other end, preset the next depth of cut, then drop the saddle in using the thread dial indicator (The thread dial indicator made sure you engaged at the same point in the thread.) without ever stopping the lathe. On other lathes you had to do it Mikes way or the thread was garbage.
B.E.

The good engineer does not need to memorize every formula; he just needs to know where he can find them when he needs them. Old professor

 

[johnchrc]

I have come to conclusion that I should include thread reliefs unless a very special situation arises. Sometimes I get crazy thoughts and throw them out to get shot down or fly. This helps me when I think I understand something but really don't.

Thanks all for taking the time to post.

- CJ

 

[Tmoose]

Making the thread machinable IS a reason for thread relief, but sometimes it IS NOT the only reason.

Depending upon the loading, as others mentioned, an abrupt change in section, whether from no thread relief, or a thread relief in the form of a relatively sharp groove, as in your design, may via stress concentration increase some localized stress more than 3X higher than a basic area calculation predicts.

The failure in the attached picture occurred fairly recently when a trusted supplier's shop foreman decided on his own to modify the relief on several shafts to be used in a product that we have built and sold for decades. The failure happened within a year of installation, just as we were assembling a second batch of machines for the same customer.

 

[hydtools]

And I bet the foreman did not charge you extra for the modification.

Ted

 

[johnchrc]

Tmoose -

Thanks for the input. However, we are not talking about the same failure mode. What I do is run out the thread to a final scratch so there in no final thread root or relief with 360 degrees of material missing through any of the threaded section. The stress concentration can be no more than with the thread as it tapers out to nothing. I have used this practice for years and have never had a failure. Stress concentration due to relief is eliminated.

I work at a new location and they insist on thread reliefs but only say because there is 1% chance it could be made on manual.

Consider a 3.250-6 pitch SA 2G thread with 4" box OD and 1" dia pin ID with 1-3/4 thread engagement. The tensile difference can be:
434,800 lbf at end of thd runout (No relief) vs.
361,400 lbf at last box thread (Thread relief)
73,400 lbf difference (that's a 20% improvement).

Now I realize that not all of that material is available as the thread helix passes through each planer section but if it even 1/2 that the strength improvement is significant.

My question is what manufacturing problem is created by calling out threads this way on a CNC? I understand the need for a relief on a manual.


- CJ

 

[johnchrc]

Add on to previous note with additional info:

That is 95 ksi material (105 ksi K500 Monel at 270F)
Thread relief on box is 3.310 inches. Pin Relief is 3.035.
Thread shear is 411kips.

I 'm haven't tested a connection to failure so I have not confirmed what amount of strength is gained.

- CJ