Hi Potting a Cable Length


I have an Industrial Installation of a string of about 20 SPST Emergency Stop Switches in series, They are wired Normally Closed. 24VDC is connected to the string and the switches will interrupt the 24V when pressed in and Emergency.

The wire is 18 GA stranded and the total length of the string is about 800 feet.

The wire runs in conduit.

I believe I have an intermittent short to the conduit, usually the short occurs just momentarily so troubleshooting is difficult.

I was thinking of Hi Potting the Switch String. I was going to disconnect each end of the string and Hi Pot the cable run along with the installed E Stop Switches. To verify if some part of the conductor is very close to the conduit.

I believe the wire is rated at around 600V.

The question is what Voltage should I set the Hi Pot tester to and why. The other end of the Hi Pot tester will connect to the conduit.


Reply to
Tube Audio
Loading thread data ...

Crank it up until something breaks. The objective here is, if possible, to turn the intermittent into a hard failure, so you can find it and fix it. I'd use a neon sign transformer, but you may not be allowed to do that.

Have you inspected all the switches? A wire short inside a conduit is much less likely than a problen at or in a junction box.


Reply to
John Larkin

Don't do that. If the fault is intermittant in nature but does not occur during the test, you'll just burn out what was perfectly good cable. Or other devices in the circuit.

Also, use a tester designed for the application. Rent one if necessary. If you damage the e-stop system in such a way that it fails during a subsequent emergency, the sh*t will come back on you.

Crank up the voltage to a level compatible with the wire and device insulation ratings. If that doesn't reveal the failure, start shaking components and operating each switch (with insulating gloves!). Heck, try the shake test with a continuity tester first before fiddling with the hi-pot.

Here's one link:

formatting link

Paul Hovnanian     mailto:Paul@Hovnanian.com
 Click to see the full signature
Reply to
Paul Hovnanian P.E.

Typical control cable is 600V. When we used to test the cabling for new control circuits (I was with a large electrical utility), we always tested them at 1000VDC. Usually resistance testers ('Meggers') are DC and 600VAC is pretty close to 1000V peak, so that was our logic. Certainly a 600V control cable should accept 500VDC and still show several Megohms, even if it is very old and in in a poor environment. New cable would be 200+.

Reply to

What makes you think it is a short ?

800M is a long run at 24V. If the circuit is holding a contactor in, small voltage fluctuations will drop it out.


Reply to
George Vest

If normal operation is causing a short either by an arc of not, then a "hi pot" tester is CERTAINLY going to fire at the short location. You STILL have the problem of determining WHERE to break in insulation is at.

In an 800 ft run, I cannot imagine that it would be a single piece, so there has to be break points.

Find one. Preferrably near the center of the run. Open ALL conductors at that break point. Test both segments. Test each conductor separately. That will narrow things even better. You should now know which end of the cable it is on (which segment fails). Go to that segment, and make another break point disconnection and test. Leave the good segment OUT of your test runs, and when you find another segment good, leave it out of all further testing.

After these two breaks and two tests, you should have the fault narrowed down to a 25% long segment of the whole run.

Reply to
Archimedes' Lever

Even with pure silver wire, 800ft pushed by only 24 V will have significant loss at the other end. Proportionally speaking. That is why we use HV across long spans. Less loss after conversion.

Could he not use DC-to-DC converters at each load node (or specific number of feet or loads)that converts 15-14VDC to 24 VDC? Then, he would have the exact same voltage at each "load node".

Then, we could bluetooth each node and he could operate them all independently via remote console. tee hee hee :-)

Reply to

800m of 18 awg is 17 or 18 ohms. Marginal for a moderately large contactor maybe but if it was so marginal that voltage fluctuations drop it out he would never pull it in in the first place.
Reply to


After reading the posts I would like to know the current and voltage ratings of the contactor coil.

18 guage is small for 800 feet with the usual current requirements for a contactor. With #18 wire and 24VDC you can sink only 1.33 Amps. And for practical purposes your contactor current is limited to .13 amps. More than this the voltage across ther contactor coil will be too low for reliable operation. Not enough current for the contactor.

If you can't replace the wire with 16 or 14 use a transistor as a driver/switch at the contactor to make the 18 wire a signal conductor instead of a switch leg. Use the 24VDC for Vcc on the transistor. There are also Solid State Switch packages for this.

Most technicians should be able to wire this up with no problems with the switches, connectors and wire. Your switches do need to be selected for this. Are your switches rated for DC?

Your momentary "short" in this case may be a result of the component activity consisting of the L of the coil and the RC of the wire and the conduit. Possibly caused by spiking from an external source. A diode across the contactor coil will minimize this. 800 feet is a long way for #18 wire.

My best philosophy at 12:23 AM!

Bob AZ

Reply to
Bob AZ

^^^ ^^^^

800 FEET, not meters...

5.21 ohms total loop resistance. A drop of about 1/2 volts if the coil drew 100 MA, which it may. Not significant.

Reply to

Get one of those frickin-hot pulse generators from John Larkin and scope the result. Where it arcs over (from the high voltage impulse) will reflect voltage back. ;-)


Reply to
Tim Williams

Before i would start chasing down a possible short, I would verify that my coil is getting enough current to stay in, and that my 24VDC supply is not overloaded. Do you have other loads on the DC supply? If you have other loads on this supply, especially intermittent ones, that could be enough to cause the coil to drop out.

  1. I would force all other loads on the DC supply to be on, and then see how the coil in question behaves.
  2. If the supply is ruled out, then I would pull the cable off of the coil and measure the resistance of the cable, including all switches (bypass the supply). If the supply voltage divided by the system resistance is marginal, that would be a good place for a problem.
  3. If this isn't the problem... 800 feet of conduit will expand in the sun. And if you have a problem which shows up or goes away with temperature, the expanding conduit might be pulling on a loose connection.
  4. Then I would check for the short.

Good luck, Andy

Reply to

Having read all the replies and good suggestions, what ever the fault is, intermittent ones are the hardest to find.

Assuming all the cable is the same age, the cable is in conduit, why not simply buy a reel of cable and replace all of it, not more than a couple of hours work to pull in a new cable with the old one as a draw wire at the lengths you describe and then you'd have solved the problem, made your insurance company happy that you have serviced and renewed the safety system and saved a lot of time and effort. Not to mention that in this condition you really shouldn't be using the equipment the safety system is there to isolate, H&S will be all over you if they knew - they wont care that the fault actually is rendering the system safe, they will only care that you have a fault with it.

Reply to

I am a firm believer in cause and effect. When is the fault happening? day night? machines on? machines off? cold? hot?

If the switches are closed, who's to say you don't have a intermittent switch contact from temp extremes or vibration? Or a loose wire to switch contact? Corrosion from water damage or humidity?

I would ohm all the switches for closed contract resistance. and check and re torque all switch to wire connections.



Reply to

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.