Formula to predetermine length growth from thread rolling

[Patrick618]

Is there a formula to predetermine the linear growth from rolling a thread? I have a part made from 4340 steel that will have a 3/8-24 thread rolled for a length of 1-3/4" and I only have .006" tolerance on the overall length which I am afraid will get longer after the thread is rolled. The thread roll will be sublet and obviously be done after the blank has been produced so I must compensate for any growth on the blank.

 

[rb1957]

why not trim length after rolling ?

why not start from thread stock (rather than bar) ?

is the application fatigue sensitive ? or corrosion sensitive ? ... why not machine the thread ?

 

[Patrick618]

The part is a machined hex head bolt run on a CNC lathe. The print calls out rolled threads only. We do not have an axial thread roll attachment and the run quantity does not justify purchasing one so we must sublet the thread rolling.

 

[rb1957]

i'd suggest trimming the part after rolling the thread

 

[btrueblood]

Plastic deformation is (to a reasonable approximation) an incompressible process, i.e. the volume of material before and after cold work should stay the same. Talk to the thread rolling shop, they should be able to tell you what starting size bar you can use to minimize length increase (if your print allows a starting blank small enough...).

 

[hydtools]

I'd expect little to no length change since the material is displaced into the peaks of the thread. Is not the preroll material diameter less than the major diameter to allow material displacement out towards the major diameter?

Ted

 

[CoryPad]

As alluded to by hydtools, if the preform diameter is equal to the pitch diameter (which is usual), then the displacement is radial and the length change is minimal.

 

[ornerynorsk]

I agree, length change SHOULD be nil, but have them run a couple of test blanks if there's any doubt. The end may appear "cupped" after the rolling process is finished, depending on the allowable chamfer, and this could put you outside of the tolerance permitted.

 

[KENAT]

.006 is fairly tight and I'd expect in the realm where slight 'deformation' of the final thread could push you outside of the limit.

Heck, unless you're being fairly precise with your starting diameter I could imagine this kind of length variation might come from if your start bar is near max tolerance.

Trimming the part after rolling sounds like an idea. Is the geometric or other reason not to do this?

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[OSIRIS7]

I do not expect any elongation, as the rolling will indeed only affect the threading.

http://www.youtube.com/watch?v=3kxcw08p_oY

At 2:10 threadrolling starts.

 

[dinjin]

Are you rolling the full length of thread?
and is the preform diameter equal to the body diameter
rather than the pitch diameter?

 

[Patrick618]

The part looks like a typical hex head bolt and is 2.800" OAL. The part grew approx. .010-.012 on the OAL after thread roll!! As I said before the thread is 1-3/4" lond but there is a .125 wide diameter smaller than the thread minor on the front end (no threads) so the front ind is not cupped or distorted. The company that did the thread rolling is not aware of any formula but said the growth we have experienced is not unusual. The parts have been heat treated already so any rework will be difficult. I just thought that there must be some formula to estimate this for any similar parts in the future.

 

[btrueblood]

So what was the blank diameter prior to rolling? Scootch the blank diameter inwards for the next part, enough to cause the volume of a .011 length extension to be made up by radial expansion required to fill the threads.

 

[Patrick618]

The thread roller ground the blank diameter to the pitch diameter and then rolled the threads. The pitch and major diameters on the parts are correct so there is no room to play with the ground blank diameter.

 

[hydtools]

I disagree, there is room to play. However, you are now faced with sizing the bolt after thread rolling. That should not be difficult even on a heattreated bolt. Use a ceramic cutting tool.

Ted

 

[PeterStock]

As an alternative, make the length before rolling .011 shorter.

Peter Stockhausen
Senior Design Analyst (Checker)
Infotech Aerospace Services

http://www.infotechpr.net

 

[KENAT]

Yeah, if the .010-.012 is consistent then just compensate. It may not be as simple as just .011 if it's more of a % length increase issue but you should be able to work something out.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[dinjin]

It would appear from your data that
.011/1.75 would give you the expected
expansion per inch of thread rolling.

 

[Cockroach]

Very interesting. I would of done as RB1957 suggested and simply face the shank of the thread by 1/64 inch after the rolling process, and not worry about it.

But from an engineering point of view, I wonder if a guy could get away with Possion's Ration for the three dimensional case of cross sectional diameter reduction? Since the cross sectional geometry is circular, then the two axis orthagonal in the plane of the circle are equal, I only have the longitudinal axis to worry about. The the strain is deformation over original length, maybe something small like 1/64 inch is not out of the ordinary!

Looks like a nice problem this evening over a few glasses of red wine. I will post a solution if I get one!

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[hydtools]

Cockroach, you may have to apply some fudge factor since some of the displaced material goes outward to create the thread peaks.
Fudge and wine. That could work.

Ted