Free falling hammer

[vcb]

Hello,

First I am not sure where to put this question, it is a difficult one.

I have a free falling hammer that falls vertically and crushes specimen being tested. The weight of the hammer is about 250 kg and it's speed 10 m/s. The deformation of the specimen is about 0.15 m.

I need to measure the force and the position of the hammer. Force measurement is an easy thing, problem starts when I try to measure hammer's position.

Linear encoder is precise enough, resolution 40 nm, total accuracy 1 micro m.I need accuracy about 0.01mm.

Problem- I cannot measure hammer's position from it's start but I start to measure just before collision. So the hammer hits the liner's encoder scale which starts to vibrate.

The encoders manufacturer specifies that encoder's redhead ridehight variation must be within 0.05mm.The scale's vibration makes measurement impossible.

My idea is to install a shock absorber to decrase scale's acceleration. Shock abrorber that has 0.3m stroke will decrease the acceleration to 333m/s^2. Seems acceptable.

The scale must be install on some kind of linear bearings or linear guideway. The tolerance of this must be smaller than 0.05mm what is difficult to buy.

I want to ask you for help, some ideas how to solve this problem.

Best regards,

przemmo

[johnsold]

My first thought was to go to an optical measurement.

How often do you need to make measurements of the position?  At 10 m/s the hammer travels 0.01 mm in 1 microsecond. Do you need to know the position every microsecond or just to 10 micrometer accuracy at a few times?

Lynn

[smercurio_fc]

I'm not sure if this will work in your setup, but perhaps an LVDT might be an option for you. Don't really know if you can get one that has an accuracy of 0.01mm, though. Additionally you'd need to provide some sort of stop to prevent damaging the LVDT if the hammer travels the entire distance of the LVDT plunger.

[unclebump]

This is the fastest distance measuring laser that I know about. I would think you want to go noncontact for the distance measurement.

http://www.keyence.com/products/measure/laser/lkg5000/lkg5000.php

[przemmo]

Thank you for reply.

I need to kont it every microsecond. We make about 5 tests a day, sometimes more.

Optical measurement will not give this accuracy. And has a high price. Inrerference laser sonsors could work good, but they are too expensice.

My problem is how to dump/ decrease the acceleration to prevent scale from vibrating?

Thank you all!

Przemmo

[Simon_T]

Hi Przemmo,

Just to run with the optical idea a little further:

Install a precise scale on the hammer, and a large zoom objective on the camera. A high speed camera on a very large zoom level can capture as the scale (markings) on the falling hammer move, similar to a quadrature encoder. The inspected area (the scale) could be 10mm, in which case you need a 200px vertical resolution to measure 0,05 mm. This is available. Make the markings little triangles, use visual inspection, and you can measure 10m/s at 0,05mm accuracy at 1000 FPS. Add a unique markings (ie a serial number for each triangle) and you can measure the 0,05mm precise analog position at any FPS you need.

All you need is a very acurate scale on the hammer, and a very stable fixture for the camera.

Best Regards,

ST

[SnowMule]

Could you use a rotary encoder on a rack-and-pinion type of setup? 

[przemmo]

Hi Simon,

This idea looks most reasonable to me. I do not know how expensive such a high fps camera can be, but from some time I am considering it. Can you help me to choose?

SnowMule- with a rotery encoder you have the same problem- acceleration. With any contact measurement the problem exists, also LVDT. The accaleration is up to 3000 m/s. If the ecncoder weight is 0,1kg (encoder + cables+ fixing ) the force is F=m*a = 0,1 * 3000=300N-> 30 kg.

I do not know encoder that could withstand this acceleration. If there is one I will consider. Have you any information.

Przemmo

[johnsold]

Przemmo,

I think the approach proposed by Simon is a good one.  The precision and speed you require are not going to come inexpensively.  The non-contact optical methods eliminate all the force and acceleration issues.  If you find the camera is vibrating, record two images.  One from the hammer and one from a fixed object, preferably one isolated from the impact.  The image from the fixed object provides information about the motion of the camera.

Lynn

[przemmo]

Hello!

Ok, the idea with high fps camera is good. As long as you want to assess the exact position of the hammer on every frame. You can do it (according to Simon's calculation) with 0,05 mm accuracy. But...

The hammer moving 10m/s travels 10 mm in 1ms. So between two frames the position will change and you do not know how.

To have information of its movement with 0,05 mm you need 20 000 fps.

Still do not have any idea...

Thank you for help.

Przemek