How to test cable for broken strands?

[vk6zgo]

Thermal cameras will pick up the cable break if the cable is under enough load and can be done while it is operating

http://www.goinfrared.com/success/ir_image/1097/industry_id/1026/

If the circuit continuously draws enough current,as you noted above.
Many cables only have current flowing when a function is performed,& are idle at other times.
If the Thermal camera method works in this case,it ,& x-ray are the only non-destructive methods that will.

VK6ZGO

[vk6zgo]

maintenance itself is a big cause of failure

I second that. Isn't it in aviation where some parts are not touched unless they fail?

I'm more than a little doubtful about this.
If you wait for a part to fail in aviation,it might be hard to fix it when the aircraft is in a thousand little pieces spread over some farmer's back paddock,or at the bottom of the ocean.

VK6ZGO

[Kiriakos-GR]

the problems were with one of solenoid cores

I would expect to read more practical approaches, about the specific problem.
That is a solenoid core who operates at 12V or 24V, and needs almost 1A ampere so to arm it self.

One primitive and brave electrician    , he would replace the solenoid in a such of a problem,
and if this had no effect, he would replace the cable.

There is no actual need to ever test a cable for broken strands.

It is the fate of the control cables to die eventually due the stress conditions that they face.
Another example of where such control cables gets stressed at it max, are in the control boxes of huge movable cranes inside of a factory.
The worker gets the control box of the movable crane at hand  ( the cable is 8-9 meters long and is connected  with an crane of  11 meters high),
The in-door crane moves on metal rails that are on the side wall of the large building.

In order to protect those control cables from the adequate pulling, that causes the distraction of them, we use " steel wire " that connects the crane with the portable control box, in parallel with the control cable, so to be stressed the " steel wire " instead of the control cable.

The steel wire haves the tendency to extend it length, and need few readjustments until to stabilize.
No one does those fine adjustments,  the workers usually do not report anything at the maintenance department until something really stops working,  and eventually the control cable stays unprotected, and dies. 

In the last factory that I was working, he had  14 of those .
 



   

[phil_jp1]

I believe I had read a thesis about this issue. They was using a high frequency pulse (signals love to travel on the surface of the conductor, that why the antenna guides on tv sets are usually hollow) through the cable. When the pulse met an "anomaly" of the cable there was an echo back to the source. Measuring times, bla bla bla, they could spot the location of the problem.

They was testing long distance cables though.

Yeah, when I was studying in university, we used that technique to localize place where cable was broken. But for such a short cables like this it must be very expensive equipment to measure such a short time between pulses.

It's just my theory, but why don't you try to measure cables resistance? If it would be some broken strands, resistance of that cable should increase.
In order to measure resistance of that cable you can use 4-wire resistance measurement method. And on top of that you don't need to measure an absolute resistance. Instead you need to measure relative cables resistance - which makes it much easier.

You would need some, let's say 12V(5V, 24V, etc.) power supply and power resistor (let's say 10 Ohm 20W for 12V power supply).
Connect all that stuff in series with cable and then measure voltage drop across that cable. That way you will get relative cable resistance readings, using pretty much any multimeter, that has mV scale.

More voltage drop across the cable - the more resistance it has. Big voltage drop = big resistance = broken strands is found!

[alm]

I'm not sure if the extra resistance from the very short distance with less strands is going to be significant compared to variations due to cable length and connectors. The problem with broken strands is that the increase in resistance is very local, so the dissipation heats up a very small part of the cable. Not sure how to test this without a large amount of current and a thermal imager.

[phil_jp1]

I'm not sure if the extra resistance from the very short distance with less strands is going to be significant compared to variations due to cable length and connectors. The problem with broken strands is that the increase in resistance is very local, so the dissipation heats up a very small part of the cable. Not sure how to test this without a large amount of current and a thermal imager.

That's why I said, that it's just a theory.

[artakami@yahoo.com]

what kind of ride is it?

Its slow moving ride 200mm/s, nothing like a roller coaster. But smooth and continues running is critical as in any attraction business.

[artakami@yahoo.com]

Sorry for leaving abandoned the topic. Ive been away.

Ive found this page http://www.cirris.com/testing/resistance/broken_strands-2.html
It looks they have done lab research on subject.
There is described another method as well (http://www.cirris.com/testing/resistance/broken_strands.html)  but, as they say, not very reliable results.

Now I need to persuade management to buy a Thermal Imaging Camera (for around £2.500 (almost unimaginable))

[amspire]

I think the most likely simple method would be to set up a machine to measure the elasticity of the copper wire in the cable. So you clamp the probe at one end and the banana plug at the other somehow, apply a force and measure the stretch. For this to work, it would need the cables to be made in a specific way. If a solder termination at one end was done at a slight angle, that give some extra stretch and confuse the measurement. This method would probably require a force that can easily break strands, as the stress may not be equal on all strands. It wouldn't work on any Kevlar reinforced cable.

Other then that, the only other likely methods I can think of that might work would be to put some RF or fast rise time pulses through the cable, and somehow probe the cable for anomalies like small localized amounts of ringing. Sounds like something that could take a lot of research.

If a good enough test jig can be made, it may be possible to send a 10 to 50 picosecond pulse down the cable and look at reflected signals. A broken strand would show up as a small anomaly in the reflected waveform, and from the delay, you can identify the position. You would be needing a scope with a bandwidth of over 20GHz but it doesn't have to be a real time scope. A Sampling Scope that has an equivalent sampling rate of over 50Ghz would probably be OK, but even better may be needed. If it worked, it would be very expensive. A big problem is the more strands, the harder it will be to see any broken wire reflection on top of the background noise and good quality multimeter cable has lots of strands. An extremely difficult test rig to build and to make work.

Detecting broken strands is a common problem in steel cables, but they are able to use Magnetic Flux Leakage detection which unfortunately only works with materials that can be magnetized. If they made special wires for instrument leads that need to be tested where each strand was a copper coated steel wire, then this method could be used.

Overall, it would probably be much cheaper to replace probes every couple of years then it would be to test for broken strands.

Richard