Thought Experiment; Detecting Distance

Question

We have two blocks, A & B shown below, moving away from each other. How can one detect that the separation has reached a certain point, distance X in this case? The most obvious solution is two rods, one on each block, that make contact when the distance is X. But this is a crude and cumbersome option. It's 2010, how about a high tech solution? There is no right or wrong, just an excellent chance to work out the prefrontal cortex and showing off some creativity.

Ghamble · Answer

How about using a laser Interferometer system. See Page 3 of the pdf here for a basic system block.

albertosombri · Answer

Hi Cabe, if the distance is not so wide the best approach to this experiment could be an UltraSonic system. Just like car parking system, I built up an EPS system some years ago, and it had a range from 10[cm] to 2 meters. Alberto S.

Former Member · Answer

Of course, if you're doing this with a reflector, 3.33nS would give you 50CM resolution, as the light will travel there then back - an increase of 1M distance then means 2M extra round trip. Again, doubling the time makes it a bit easier to detect perhaps. My apologies - I came straight to the 2nd page! Laser interferometry too expensive then... Ok, well using a laser pointer as a previous poster suggested, with a reflector, might be the way forward. As the blocks are also fairly close, an old barcode scanner might be an ideal source of parts; The internals often use a sensor that's effectively a single line cmos / ccd camera. This could be used to determine the angle of reflection quite easily, and also be an easily adjustable / configurable solution. Ultrasonics might be a good way forward too, but you could also use an LED driven by a tone generator - a phototransistor might then be used to determine distance based on either amplitude of the reflected signal, or angle, dependent on placement / focus / surface reflectivity. Hope that helps some more! Further to the above, the PIC24F64GB004 (and others) have a CTMU peripheral. This is capable of 1nS resolution & generating pulses asyncronously w.r.t. the internal clock. So, laser ranging may still be possible on a budget! (1M = 2M round trip of 6.6nS, therefore 1 / 6.6 = 15.15CM resolution @ 1nS) Perhaps not for the 250 microns you'd ideally like for your setup though.

Former Member · Answer

Your picture is pretty much what I thought, except the laser would need to be set at angle theta, on the "source" block. Movments on the target block away / towards source would then change where the reflected laser point hit the target. Either way, you've got a mirror / reflector / angle problem to solve. Setting theta for the laser shouldn't be too hard - simply accurately measure a right angled triangle inside a large room, the triangle is of known proportions, bingo, you've got a known theta. The other way is to get a reflector set at angle theta on the target block. A little bit harder, and if it's moving, it may change the angle and destroy accuracy. If you use a flat mirror on the target, it's probably simpler all round. Anyway, it looks like the final calc is going to be distance = (h/2) / tan(theta) as your diagram would appear to show the beam reflecting off a surface with a mirror set at theta degrees. So, detecting where the laser pointer hits is the last problem to solve. Preferably, I suppose you'd use as few optical components as possible as that way is going to be expensive, due to accuracy depending on quality and ability to set correctly. Simplest option I can think of is to print off a scale on a laser printer / A4 paper, and use this for the beam to hit and for a camera to see. Software can then determine where the beam is hitting, maybe even without a printed scale, and convert to distance. In my opinion, using software to do the bulk of the work is going to be the best / cheapest way in terms of physical cost, apart from perhaps time. All you'd need to do though is detect at what point the colour thresholds exceed a certain value, in the X and Y planes. A camera looking at the back of the sheet of paper should do this ok. Classes are available in both .NET and Java to do this, depending on your preference. For more accuracy this way, the simplest method is to make angle theta more obtuse, therefore affecting h more over a given distance. To calibrate such a device, measure distance D at two points with calipers, and then mark the spot positions (in software / on paper / however) as required. Determining distance between these two points should be a simple matter of interpolation.

Former Member · Answer

What about String-pots? http://www.unimeasure.com/ With a little bit of calibration/interpolation, you can probably achieve the accuracy you're after Or LVDTs? http://www.lvdt.co.uk/pdf/AML-E.pdf Or finally magnetic encoders; http://www.renishaw.com/en/magnetic-linear-encoders-and-magnetic-ring-encoders--9802 Any of those are simpler/easier to implement that optical or US solutions, and should give you the kind of accuracy you're after.

Thought Experiment; Detecting Distance

Top Replies