Cable Length Measurement

I have a reel of cable approximately 1000 ft long attached to a robot. I need a way to measure the length of cable that is off the reel in real time (it is being pulled out and reeled back in constantly).

My currently method is to measure revolutions at the center of the reel using a quadrature encoder. While this works, what I'm really looking for is a method that can be retrofitted onto the reel and placed outside the reel.

Cost is a factor, and the cable itself operates in a wet environment, so non-contact would be ideal.

Any ideas? Something like this seems like it would work well, but I'm going to guess it will be way too expensive.

http://www.laserspeedgauge.com/index.php/en/wire-a-cable

• No ideas-- but I have the caveat that one reel revolution when the reel is full will let out much more cable than when the wheel is empty. You'll have to correct based on filled circumference Commented Aug 8, 2013 at 14:14
• Let the cable run over a spool (or clamped between a pair of spools) as it leaves the reel. Then measure revolutions of that spool. (Of course, this only works if you get enough friction between cable and spool to avoid slipping.) Commented Aug 8, 2013 at 14:34
• How accurate does it have to be? Commented Aug 8, 2013 at 14:50
• us2012's answer is the same one I was going to give. I've worked with machinery that worked just like this. What is the problem you're having with the setup you are using at present? Commented Aug 8, 2013 at 15:54
• Like us2012 and lyndon said, I think a double spool would be the best idea. You could even place in some spring clamps on each separate spool to press the spool wheels against the cable. This is what pipe inspection cameras have so plumbers can tell how far into the pipe their camera is! Commented Aug 8, 2013 at 15:59

Possible solutions:

1. Use a sensor from "modern ball-less mouse". Essentially this is a image sensor + processing that can detect motion. Point it at the cable and follow the motion data that it sends. Possible problems: Cable might be too thin to be detected properly, or too uniform in color, or move too fast. Could also get confused if the lens gets dirty/wet.
2. Put a magnet on the reel, and use a hall-effect sensor to detect the magnet. It could be done "outside the reel", to retrofit existing reels. Also, it's non contact and can be completely sealed against water. Downsides: Might not work depending on how the reel is shaped, and has limited resolution/accuracy.
3. Mark the cable with either bar-codes or colors or stripes. Then detect these optically as the reel spins. It could be something as simple as changing color every foot or meter. This could be super easy to implement, or super hard. It all depends on how complicated you want to get. Reliability could range from super reliable to super not.
4. Do what US2012 suggested, which is to have some sort of measurement wheel/pulley that the cable moves over, and the measure the rotation of the wheel. This could be problematic if the cable gets wet and slippery, but there might be ways around that. IF it works, then it should be fairly cheap and reliable. But it might not work.
5. Have a different sensor that detects the depth of the cable remaining on the reel. This could be a simple arm that slides over the cable on the reel as it turns. The angle of the arm changes depending on how much cable is left on the reel. This could be a fairly cheap and robust way to measure, but has the worst accuracy of them all.

Of all of these, and without knowing the details of your situation, my preference would be for #2-- putting a magnet on the reel. It has reasonable accuracy, high reliability, simplicity, and can be easily retrofitted to an existing reel.

• #6: Some sorf scheme to weigh the spool. This might be easy or hard depending on your setup. Accuracy should be pretty good once calibrated. Commented Aug 8, 2013 at 16:12
• @OlinLathrop I was thinking about that, or weighing the cable that has been played out. But when people talk about reels, cables, and water I immediately think of a tethered underwater robot. When you factor in the waves, currents, and boat movement then weighing anything becomes difficult. Of course I could also be reading way too much into this, and it really is super easy.
– user3624
Commented Aug 8, 2013 at 16:25

Imagine this set-up: -

Left Hand side: The cable, as it is withdrawn from the reel passes through a metal/conductive sheath that grounds the cable capacitively.

Middle: Reel (drawn square because I can't do circles as neat in "paint.exe")

Right hand side: An oscillator capacitively couples its signal to the cable via a conductive sheath.

Idea: As the cable unfurls from the reel it's inductance gets smaller and the oscillator (tuned LC type) increases its frequency. This can be measured and equated to length of cable remaining on the reel.

Alternative idea - Send a pulse from the oscillator sheath and receive a series of pulses at the earthed (now measurment end) sheath. The length of time to receive the last pulse largely is dictated by the length of the cable on the reel.

• Interesting. It won't work with non-insulated cable, like a bare steel cable. But for insulated stuff this is quite innovative. The right side would be difficult, since that part is spinning. An interesting alternative would be to turn the spool itself into the primary side of a transformer (single turn), and the cable on the reel is a secondary with a varying number of turns. Put a pulse into the primary and see what comes out of the secondary.
– user3624
Commented Aug 9, 2013 at 2:03
• +1, this is a neat idea. Did you check for existing patents? :-) Commented Aug 10, 2013 at 16:44
• @AnindoGhosh I've got three patents in metal detection, one more aint gonna make a big deal LOL. You are welcome to it. Commented Aug 10, 2013 at 16:52
• @Andyaka I wasn't referring to me getting a patent, nor you, but whether some smart person already had! In any case, publication of an innovation in a public platform precludes patent protection :-) Commented Aug 10, 2013 at 17:27
• @AnindoGhosh well, if someone has already patented it then my "idea" shall be rightly relegated to the dark corner of legally unimplementable answers. At least now nobody can patent it!! Commented Aug 10, 2013 at 17:58