3

I'm working on a project which involves measuring the distance between two pins which are nominally 3-5 cm apart. These pins can move up to 1cm in either direction. I need to take a measurement of the distance between the pins which is accurate to about 500 μm.

Currently we take a measurement once a day using calipers and record the distance manually, but I need to automate the process.

Update (2012-02-21)

Okay, I figured the implementation details were unimportant, but as a lot of people are asking for more detail, here goes. We want to measure the opening and closing of a crack in a brick wall over the course of a year or more. We currently drill holes and insert two metal pins into the wall either side of the crack and take measurements between the two pins. So long as we can measure the opening and closing of the crack over time, then the implementation is unimportant.

stevenvh
  • 145,145
  • 21
  • 455
  • 667
littlecharva
  • 564
  • 1
  • 6
  • 12
  • 1
    "Small" is for hand waving, not engineering. And 0.05 cm is a really silly unit. You might as well tell us what you need in micro-furlongs. In engineering we make sure there are 1-3 digits left of the point and adjust the x1000 multiplier accordingly. 0.05 cm is half a mm or more properly put, 500 um. That's about 2.5 micro-furlongs, but I don't recommend saying it that way. – Olin Lathrop Feb 16 '12 at 16:22
  • I see a lot of very short answers to this question that all just sound like brainstorming. Maybe this question is more of a brainstorming question and would be better for chat. If not that, then maybe a community wiki. – Kellenjb Feb 16 '12 at 21:25
  • @Kellenjb - Excellent observation. However, I still have hope that this can be salvaged. I've added a post notice, cleaned up the comments, and incorporated some of the information into the question. I'd much rather see this get a good answer with a comparison of distance measurement methods than relegate it to some community wiki brainstorming abberation. – Kevin Vermeer Feb 17 '12 at 15:16
  • 3
    @Kellenjb: What else can one do without some real information? What are these "pin" things? Can they be attached to mechanically? How much force is allowed? What is the overall purpose? – Olin Lathrop Feb 17 '12 at 16:35
  • Not sure if this will meet budget or long-term mechanical requirements, but try a web search on "wire draw encoder". – The Photon Feb 20 '12 at 17:57

5 Answers5

2

A vibrating wire transducer can automate this process. http://www.slopeindicator.com/instruments/jmeter-crack.html

They are expensive but it may give you an idea on how to build your own. The challenging part will be calibration.

MarkSchoonover
  • 1,113
  • 10
  • 17
1

This sounds like a job for a LVDT (linear variable differential transformer). These are transducers that measure small absolute displacements from some nominal position. Usually the total movement is rather small, like a mm or two, but that is apparently all you need. There are also dedicated chips to drive these things and extract the position information. I know Analog Devices makes some, and there are most likely others.

Added:

One reason LVDT came to mind is that they sense absolute position. A strain guage system could possibly work, but those usually have significant drift. Power shouldn't be a problem since it only needs to be powered up for a short time once a day.

"LVDT" is a standard term in the industry, so you can find lots of stuff written about these devices. For example, here is one page I found by searching "LVDT position sensor".

Olin Lathrop
  • 310,974
  • 36
  • 428
  • 915
1

If you're allowed to connect them together, something like a strain gauge could work: it's a variable resistance that increase its value if strained; however, requires a mechanical connection and gives very small signals with huge common mode; Wheatstone bridge circuits with differential amplifier (InAmp are better) are usually used.

Another way could be measuring the capacitive coupling between the two pins: but the feasibility depends on shape and distance of the pins, as well as if you can put signals in the pins themselves.

Update

I've read your additional informations, and I think that both strain gauges and LVDT can work (note that you have to measure differentially, this is good). Since you are almost free to play with those pins, I would suggest also to create a structure (maybe with curved pins) to reduce the distance, and make the variations more significative.

clabacchio
  • 13,481
  • 4
  • 40
  • 80
1

You say you are using calipers manually. Can you attach anything to the pins? If so, could you use two electronic dial indicators with serial data output? They will handle the resolution you need. They are typically spring-loaded so the probe will move back and forth with the pin. Here's an example

lyndon
  • 4,366
  • 1
  • 16
  • 11
1

I'm not sure if this will meet your budget or long-term mechanical requirements, but try a web search on "wire draw encoder".

This is a wire that pulls out of a mechanism to measure a length. As the wire is pulled out, the mechanism can produce an analog voltage proportional to the withdrawn length, or a digital signal. The digital signal is probably better for your use. This emits a pulse for each length of wire withdrawn (say each 1/10 or 1/20 mm).

Wire draw encoders are apparently available with resolution to 50 microns at lengths of up to 5 meters. Overall accuracy is specified as a percentage (like 0.05%) of the measured length, probably limited by thermal expansion of the wire itself. In any event, the resolution is probably more important than the overall accuracy in your application, unless your wall is in a location subject to wide temperature swings.

The Photon
  • 126,425
  • 3
  • 159
  • 304