Torque applied to a system with belleville springs

[DaoBrabo]

This has something to do with thread404-239297.

Hello. I'm trying to calculate the torque I need to apply to a nut so that I can obtain about 2000 kgf of axial force. I have a component that can't take more than 2000 kgf. At the same time, the sealing of the system must be guaranteed.

Does the equation T = K*D*F (in which K = 0.2) apply in this case or is it just for bolts? I know whis formula isn't very accurate, since it doesn't account for pitch, but I don't need much accuracy for this.

Also, there's a large set of belleville springs in the bottom part of the system: 6 sets in series of 3 springs in parallel (each set has 3 springs in parallel, and there are six of those in series = 18 belleville springs in total).

I was wondering if the springs somehow change the way the force is transmitted to the bottom of the system, where the delicate part is located. Since the system will be pretty much static, I believe the formula could apply.

Another way I'm thinking of determining the "torque" is by measuring the deflection of the springs.

Thank you for helping me

 

[moltenmetal]

You can't use torque to determine the tension in a fastener with a Belleville spring. Instead, you must measure the compression of the Belleville spring itself. Using the torque after the fact as an indication of the compression of the spring may be possible if you use a washer between the nut and the spring washer, but I doubt it will be very repeatable...You'd have to test your own situation to see, with measurement of spring compression being the true arbiter of fastener tension.

 

[Screwman1]

Molten is right- for this type of situation you are better off trying to measure compression rather than trying to chase the torque/ tension relationship. If the threaded part has no prevailing torque feature, then you might be able to use a torque / angle strategy, where you set the threshold torque at a low level when the bellevilles just start to compress and then switch to angle control of the spindle to get the number of rotations that you need to get the required compression.

 

[DaoBrabo]

You guys are right, we'll just measure the compression of the springs. One of our guys really wanted to do it by means of the torque, especially because he has a torque sensor and it'd be easier for him. What we might do is create a simple empirical relationship between spring compression and torque measured for a few values.

Appreciate the help!! You guys are amazing

 

[3DDave]

I handled this situation by getting the springs into a spring tester and measuring their combined stack height at the required load. Then they were removed and put onto a string loop to preserve the order and labeled with the height. It produces a very small variation in load, without regard to torque. All that is required is a caliper.

It was a result of noting that minute variations in belleville springs retult in large variations in spring height at load. Even smashing them flat a few times was required to even out the forming residual stresses.