Measuring the speed of electrons?

[bigal_scorpio]

Hi to all especially Nigel and Eric,

I am trying to figure out if it would be feasible to make a device that could test the length of cables with any degree of accuracy.

My idea was that I may be able to measure the speed the electrons take to make a circuit on a given length of wire and with a bit of maths then use this as a formula to work out how long an unknown length was, single core would need both ends connecting to the device of course so the time would be right and with multiple cores any two could be joined and then the result halved. Or is it that simple?

The problem I have is what would I need for accurate measurement of such tiny times or is it even possible with ordinary run of the mill ICs like PIC or such.

I have seen meters on line for doing this but I assume they must work by capacitance or something as they can detect open circuits and breaks.

Anyway if anyone has any ideas or further info they can point me to I would be grateful.

Thanks for looking........Al

 

[Hero999]

Such a device already exists, it's called a time domain reflectometer.
Time-domain reflectometer - Wikipedia, the free encyclopedia

 

[ThermalRunaway]

Could you not accomplish this by measuring the phase difference between an input signal at the source and the output signal at the receiving end?

It would be crude, but you should be able to use some mathematics to calculate the approximate length of the cable based on the phase difference.

You would need to terminate the end of the cable so that you don't get reflections.

Would this work? Not sure - never tried it myself. Try it on a scope with a length of cable and see what happens! There should be a phase difference, and if there is you should be able to measure it and use the difference to calculate a cable length.

Brian.

 

[gaspode42]

A lot of IT cable tester can also measure cable length.

 

[Hero999]

The beauty of a time domain reflectometer is that you only need access to one end of the cable.

EDIT:

You could probably do with with a 'scope and pulse generator: send short pulses down the cable any pick up the reflections with your 'scope, measure the time between the pulse and reflections and work out the electrical length of the cable. If you have a cable of the same type with a known length you could easy calibrate your results using it.

 

[bigal_scorpio]

Hi Hero and Brian,

Yes, just as I assumed that method relies on something other than just measuring the time travelled by the electrons. I would prefer not to get into anything as complex as waves and wave comparison as to be honest most things nowadays seem to be beyond my small (and dwindling, doh!) knowledge.

So I thought if I went by the simplest route I may have at least some chance of succeeding!

Al

 

[ThermalRunaway]

The beauty of a time domain reflectometer is that you only need access to one end of the cable.

The downside is, they tend to cost a lot of money :-/

But regarding your pulse suggestion, yeah I guess that could work. It's similar to the suggestion I made above, except you're measuring the reflected pulse instead. This has the advantage that you don't need access to the other end of the cable, although you would need to make sure it was O/C at the other end.

Brian

 

[ThermalRunaway]

Saying that... how would you differentiate between the source pulse and the reflected pulse? Would the reflected pulse not distort the shape of the source pulse? And what about multiple reflections?

Brian

 

[bigal_scorpio]

A lot of IT cable tester can also measure cable length.

Hi Gaspode,

Yeah, I have seen similar ones but I wanted to build one if possible, just to feel that I have accomplished something.

Al

 

[ericgibbs]

hi Al,
As others have stated, a common method is to pulse the cable and measure the time taken for a reflected pulse to return.

A break or short in a cable causes a reflection of the pulse.

Transmission speed is close to the speed of light so its not cheap.

The actual speed the individual electrons move down the wire is measured in seconds per metre, quite slow.

For an open circuit cable, why dont you use your capactance meter across the cable.? use a 1 metre or so of exactly the same cable as a reference capacitance.??

If its a shorted cable,do the same tests with a resistance bridge.

 

[ThermalRunaway]

hi Al,
As others have stated, a common method is to pulse the cable and measure the time taken for a reflected pulse to return.

If you use this method, what happens if the pulse is reflected multiple times? Then the result would be a distorted pulse, caused by the addition of multiple reflections?

Also, you would need to be careful with the pulse width because otherwise the reflected pulse will arrive back at the input before the source pulse has finished.

In my opinion, if you have access to both ends of the cable, it is easier to send a periodic wave and measure the phase difference between them. The far end of the cable would need to be terminated, of course.

Brian

 

[ThermalRunaway]

If you use this method, what happens if the pulse is reflected multiple times? Then the result would be a distorted pulse, caused by the addition of multiple reflections?

Actually, I'm not sure that's right. It would be correct if you send a periodic signal down the cable, but probably not with just one pulse.

I'm tempted to try this myself now!

Brian

 

[ericgibbs]

If you use this method, what happens if the pulse is reflected multiple times? Then the result would be a distorted pulse, caused by the addition of multiple reflections?

Also, you would need to be careful with the pulse width because otherwise the reflected pulse will arrive back at the input before the source pulse has finished.

In my opinion, if you have access to both ends of the cable, it is easier to send a periodic wave and measure the phase difference between them. The far end of the cable would need to be terminated, of course.

Brian

hi Brian,
A few years ago we made 6 prototypes for a German company using the reflected pulse method, they worked OK.

They used them on underground cable testing to locate the position of the fault.

Another commonly used method for road cables, is a cable frequency pulser connected to the end of the cable and a handheld detector coil/amp.

 

[bigal_scorpio]

Hi Eric,

Nice to hear from you again mate.

I like the cap meter idea but would the result be affected by coiled cable?

I just want something cheap and cheerful, to quote your tag "good enough", it would be used infrequently and mainly to check if there is enough cat5 cable in the box for a particular job so I would only need a rough guide, maybe 5% or even 10% would probably be an accepable error since the boxes have 305 meters to begin with.

So whats your view on the accuracy and coiled to flat question?

Your humble student..........Al

 

[tcmtech]

Being its new Cat5 cable just do a resistance check with a low ohms type ohm meter or home made unit.
If you know the diameter if the wire you can get its cross section and thus get a fair length estimate just by its known resistance.

Basic and simple.

 

[Hero999]

Here's how I find the break in the cable:

1. Make sure nothing is connected to the cable and that it isn't touching an earthed part or large piece of metal casing.
2. Warn people that the cable is going to be live.
3. Connect the cable to the live mains conductor.
4. Use a DVM with a non-contact mains frequency electric field detector to scan the cable. The point in the cable where it stops beeping or becomes less sensitive is where there's a break.
5. Disconnect the cable from the mains.

You could also use a neon lamp and a high frequency high voltage source such as a cold cathode tube driver, just make sure the voltage is below 1kV or it could burn through insulation or arc across connecto contacts.

 

[RCinFLA]

You're not going to find a kinked elbow on a cat5 line with an ohmmeter.

A too sharp a bend on cat5 will cause the four twisted pairs to inter-twine resulting in crosstalk interference. Then there is the poor installs where several inches of the line are un-twisted and inter-twinded to connect to the RJ45 jack.

Cat6 puts a four quadrant plastic divider within the cable to keep the fours twisted pairs separated to reduce crosstalk. Gigabit lines puts tougher requirments on hooking up the RJ45 jack. Keep the twists tight to the jack.

 

[Mikebits]

Could you not accomplish this by measuring the phase difference between an input signal at the source and the output signal at the receiving end?

It would be crude, but you should be able to use some mathematics to calculate the approximate length of the cable based on the phase difference.

You would need to terminate the end of the cable so that you don't get reflections.

Would this work? Not sure - never tried it myself. Try it on a scope with a length of cable and see what happens! There should be a phase difference, and if there is you should be able to measure it and use the difference to calculate a cable length.

Brian.

Depending on the cable length, the phase change may go through several iterations and I see no way to know which iteration your looking at.
The TDR is the best method I know of.

 

[mneary]

For the Original Problem (post #14), (knowing the remaining contents of the box) the easiest way would be to measure the DC resistance of a full (1000 ft) box. I would connect all conductors together at the end, so I can measure at the round-trip resistance without fishing back into the middle again. Be sure to make a tight connection that can't corrode. A twist connector is a good idea. You can measure between any two conductors. Should be about 51 ohms.

You should beware that at the low end of the scale, your probe resistance could introduce errors of several ohms.

 

[ericgibbs]

Hi Eric,

Nice to hear from you again mate.

I like the cap meter idea but would the result be affected by coiled cable?

I just want something cheap and cheerful, to quote your tag "good enough", it would be used infrequently and mainly to check if there is enough cat5 cable in the box for a particular job so I would only need a rough guide, maybe 5% or even 10% would probably be an accepable error since the boxes have 305 meters to begin with.

So whats your view on the accuracy and coiled to flat question?

Your humble student..........Al

hi Al,
The spec for CAT5 is fairly tight, so as suggested a simple resistance check would be fairly accurate.

Category 5 cable - Wikipedia, the free encyclopedia

Of course if you have some decent scales you could consider weighing it.