Force testing with a magnet

[MAC3382]

Hello, All:

I hope that this is the appropriate forum for this question.

I have a job in-house which requires quantifying the break-away force of a magnet assembly (~ 6 N +/- 2%) through an air-gap with a tolernace of +/- 0.015mm. The magnet attracts to a precision work-piece of a specific alloy.

To date we have a motorized test stand, a 10N transducer, a linear bearing guided jig, and various methods of "fixturing" the magnet.

The problem is that the fixturing of the magnet corrupts the measurment. The system/method will not pass a gage R&R because it is wholly dependent upon the operator. (We have tried clamps, clips, carriages, etc.)

We have also had a XY knife edge coupler built for a near frictionless coupling. The thought was the compliance in the coupler would account for the minute mis-aligmnets and symetry issues, but the compliance in the coupler actually exacerbated the problems. (Any break-down in teh geometry symetry would cause an unbalanced force condition and the magnet field would establish force compenents to futher skew the system.

Instead of compressing a material to act as an airgap and pulling the magnet away, we would like to suspend the magnet a set distance (0.5mm) above the work-piece. The work-piece would be connected to a lab grade scale. (Common scales are at least an order of magnetude btter than a commercial load-cell) The problem is that, like the load-cell in our transducer, the scale works on displacement. So, the airgap is reduced and the reading is incorrect. we have tested several Acculab and metler scales and found as much a 0.010" of deflection for our target mass/gf.

Finally the question:
I have found some literature for a type of scale style that uses "Forced Restoration Field" or magnetic compensation to return the "pan" back to the initial equilibrium position. I have. however; not found this to be common and most scale manufacturers have no working knowledge of this. I have a Technical datasheet from metler on this subject, but it is from 1985.

Does anyone have any knowldge on this style loadcell? I would appreciate any and all leads.

thanks,

Mac

 

[JR97]

Mac,

Maybe I don't fully grasp your situation but could you establish your 0.5 mm initial air gap by using plastic shim stock between the magnet and the piece? The plastic should not affect the strength of the magnetic field in the piece. Then lift your magnet and record minimum scale reading.

JR0097

 

[MAC3382]

Thank you for your response JR0097.

Yes, that is what we currently do (leaxn 8010 film stock). That requires us to fixture/clamp/etc on the magnet assembly. This seems strait forward, but it is not. Also to meet the 6N tensile force we need about 25N of compressive force. This begs the question, why not use some mechanical element/feature that supports the bottom of teh magnet? Well the answer is "squareness." The area that we can "support" is inbetween two protruding elements and we can not produce the surafce square enough to the working surfaces such that the "pull" will not be corupted with rotational forces.

We have lloked for a fixturing solution for about a year. constant optimization.

Are you familiar with the self correcting loadcell?

thanks,

mac

 

[israelkk]

Can you explain the 25N? A sketch may make everything clearer. As to the (~ 6 N +/- 2%) this kind of accuracy need calibration for every piece of magnet. Magnet properties tolerances are far more larger than +/- 2%.

 

[MAC3382]

Hello, Israelkk:

About a 25 newton compressive force is required by one of our many clamp solutions to "fixture" the magnet. The magnet is Pulled "normal" to the work-piece and the compressive fixture device "grips" the magnet assembly from the sides.

Regarding the magnet. The "magnet" is actually a magnetic assembly. It was deisgned to operate lower on the Force vs. Gap curve. Lower meaning less exponential portion. We also pre-qualify the magnets based on teh magnetic moment. Over th epast year we have been able to achive this level of accuracy, but the current measurmenet method is very cumbersome.

Sorry, no sketch or images. Quite propriatary. Sorry.

Mac