Lift Assist Torsion Bar 1

[JacobP12345]

What would be an ideal material for a torsion bar assisting lifting a 200 pound lid with the rod being subjected to a displacement of 90 degrees? Hardened 4140 has been attempted but it yielded and is not ideal. Looking elsewhere has suggested higher carbon steels such as 1074, but before trying this I want a better idea of what would work. What kind of spring steel if any would work best? The bar will be 36 inches long and of 1/2 inch diameter. From my calculations using the material properties of 4140 I am getting a maximum surface stress of 240 ksi. Any Suggestions?

 

[desertfox]

Hi

Here is a link regarding the theory you are using and one of the assumptions is that the angle of twist of the round bar is small. It might be you are having problems because your angle of twist is large.

https://www.oreilly.com/library/view/strength-of-materials/9789332514829/xhtml/ch6sec2.xhtml

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[FEA way]

Sure, regular mechanics of materials formulas always require many simplifications with the biggest one being linear material and geometry. In this case the bar experiences large strains and mechanics of materials analytical calculations are not enough. In case of torsion also warping plays important role but this mainly occurs in thin-walled open sections. That's why nonlinear FEA is advised in such cases. It has its own limitations but eliminates pretty much all the assumptions of hand calculations. As you can see the results are very different. Simple formula says the stress is 120 ksi while nonlinear FEA gives 3.4 times lower stress. That's a huge difference and of course correct result is the one from FEA. For such stress it will be much easier to find appropriate material. It only needs to have slightly higher yield strength than typical steel.

 

[JacobP12345]

I fully trust your analytical calculations and have always wanted to learn FEA methods better. Our issue is that after testing a 1/2 inch bar of low carbon steel with a yield of 54 ksi and a hardened bar of 4140 with a yield of 110 ksi we still experienced plastic deformation and rapid fatigue. Is there a way to analyze where the bar will fail, such as an angular displacement? Our low-carbon bar experienced plastic deformation at around a 45 degree twist, of which worsened steadily under a few cycles. Our hardened 4140 bar (not tempered) experienced permanent plastic deformation at approximately 70 degrees. When we attempted an experimental design during rudimentary RnD we found that if the angle of twist was kept under the 60-70 degree mark (test was performed in a weldshop with minimal equipment of measure so angle was approximated by eye) the bar did not deform, but of course due to a lower twist much less torque was provided to help lift the lid. Is there a way to enhance the analysis by utilizing the angle at which it plastic deformed at? Thank you for the help.

 

[FEA way]

Nonlinear Finite Element Analyses are solved in such way that full load/forced displacement is divided into steps (increments) and applied gradually. It is possible to examine results for each of these steps. Thus if you plot plastic equivalent strain (PEEQ, used to determine which regions yield and to what extent) for all the increments you will see when the plastic deformation starts. In my analyses yield started at about 26 degrees.

 

[JacobP12345]

Would there be deformation considering a much stronger heat treated 4140 bar?

 

[FEA way]

It depends on the plasticity data of such heat treated 4140 steel. I assumed yield strength of 415 MPa and ultimate stress of 655 MPa with corresponding plastic strain of 0.26. You would have to obtain accurate properties of heat treated 4140 to compare it with regular one.

 

[JacobP12345]

Considering the relationship of UTS to yield strength provided earlier, the new UTS should be in the ballpark of 1375 MPa and the yield strength should be roughly equivalent to that from my plot of the relationship. Thank you for the help

 

[EdStainless]

We used to test shafting and couplings with a little home made rig that gave us an out put of torque vs angle. It was a torque meter and a simple rotary encoder.
The response did not need to be fast and the encoder was very rough.
I question if your 4140 has been Quenched and tempered at all. Perhaps it is normalized and surface hardened?



= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[JacobP12345]

It was, we're trying a tempered and quenched bar next as per your suggestion. The bar we had was rated at 100ksi. Will report in with the results of the tempered bar.

 

[TheTick]

I haven't seen anything about desired spring rate or preload or other characteristics that define the performance of a torsion bar.

How much of the load are you actually trying to lift? (Certainly not the full weight.) Where is the CG of the load w.r.t. the pivot?

Problems like this typically start with a graph of lid torque vs angle and spring torque vs angle on the same graph. The lid torque will be a cosine curve and the spring torque will be a straight line determined by spring rate (k) and preload. Resting points are where the curves cross.

 

[Jughandle]

If you can go larger than the stated 1/2-inch diameter, try steel tubing slit lengthwise; make wall thickness to suit.

 

[Jboggs]

Have you considered using a mechanical reduction (such as gears or levers) to reduce the amount of angular deflection required of your torsion bar?