How does "bearingless" work in rotary torque transducer?

[BrianE22]

I've found a torque cell that would work for me:

http://www.interfaceforce.com/index.php?T11-Bearingless-Rotary-Torque&mod=product&show=66

I'd like to understand how it supports the spinning shaft without bearings. I'm a bit familiar with hydrostatic bearings but it's not that. I thought maybe magnetostatic but that looks to be big and expensive.

The base model for this sells for $4900. Any ideas how they support the shaft on these?

 

[Nescius]

Are you sure that they support the shaft at all?

 

[drawoh]

BrianE22,

Either your device has bearings of some kind (magnetic? air?), or it is modular (frameless?), and you need to provide the bearings in your design. This is not necessarily a bad thing.

--
JHG

 

[SwinnyGG]

That design requires the rotor to be supported by the two shafts attached to either side (via couplings).

It comes with tools which hold the rotor and stator housing in alignment during installation. After the couplings are installed and the motor and load are connected, the tools are removed. The rotor is then free floating within the stator.

 

[BrianE22]

Yes, I believe you two are correct - no support of the rotor. I'm impressed that it works. With the rotary transformers there's a lot of mass on the rotor that spins at 30,000 RPM. It must be balanced very well. Also, I would think the rotary transformers would require a minimal gap between the rotating and non-rotating pieces and that gap would have to remain constant.

The brackets and the base would have to be machined to tight flatness and squareness to avoid any movement when tightening the mounting screws. Very impressive that you could align all the parts, tighten the mounting screws and couplings and still control the small gap between rotor and housing.

 

[SwinnyGG]

They actually don't require all that much precision. The gap being imperfect around the radius doesn't have much effect on the output, and it's constant (rotor and stator relative positions are fixed) so it doesn't induce any transient error.