Megger testing control cables

[Srcarroll]

Hello all,

I have multiple bundles of 190, #14 control cables I need to megger test.

How would I go about testing these cables? I’ve never had to megger control cables.

Thank you in advance.

 

[ptonsparky]

That could be time consuming.
Maybe speed things up by shorting all conductors of a cable together on one end or both. Then check from cable bundle to ground/earth. Open the connections, then check conductor to conductor. Testing does not guarantee that conductors do not have missing pieces of insulation.

 

[Srcarroll]

That could be time consuming.
Maybe speed things up by shorting all conductors of a cable together on one end or both. Then check from cable bundle to ground/earth. Open the connections, then check conductor to conductor. Testing does not guarantee that conductors do not have missing pieces of insulation.

Yes, very time consuming. I’m not familiar with this and have never seen #14 control cables megger tested before. There are 8 conduits with 190 cables in each. can You please elaborate on your previous response? Best way to short 190 cables?

Thank you in advance

 

[cadpoint]

Not so lite reading this weekend for you would be to find and download the PDF on "A Stitch in Time"
A good overview of megging.

 

[tom baker]

Agree on Stitch in Time. Excellent on insulation testing

 

[hbiss]

I wouldn't short them all together except as a preliminary test to see if there is a short to the conduit. You would want to individually test each conductor to every other conductor one at a time. That's 190*190= 36,100 tests?

-Hal

 

[ptonsparky]

I wouldn't short them all together except as a preliminary test to see if there is a short to the conduit.

-Hal

That would be the only reason.

Testing plus documentation. Thirty minutes each?

 

[hbiss]

That would be the only reason.

What about conductor to conductor?

-Hal

 

[Ken_S]

Do you want to test them or is there a contact requirement that you need to test them? If it's a contractual requirement then an RFI is in order to clarify what is expected.

 

[ptonsparky]

What about conductor to conductor?

-Hal

Conductor to conductor and conductor to ground are the only tests I can think of.

 

[hbiss]

My point is that conductor to conductor testing is the killer with that many conductors. I have no idea if anything is available, but the only real way to do it is with some kind of automated system.

-Hal

 

[ptonsparky]

My point is that conductor to conductor testing is the killer with that many conductors. I have no idea if anything is available, but the only real way to do it is with some kind of automated system.

-Hal

I agree. A real PIA.

 

[cadpoint]

Were just poking at the problem, the OP needs to come back and give some more exposure to this situation.

Anyone that bundled that many wires did a disservice to the vocation.

 

[JoeStillman]

You could short the cables into two bundles of 85 each and test bundle to bundle. Then do two more tests of 42/43 each, then four tests of 21/22 and on down til you have pairs of wires left. You would only have to do 255 tests (assuming there are no shorts.)

 

[JoeStillman]

Another way would be to bundle 189 wires and test the one left against the bundle. Then keep taking one wire out and testing it against all the rest.

 

[W@ttson]

On the far away end (from the user) keep all the cables separated and disconnected).
Now at the side where the user will have the megger get jumper cables and connect all of the cables to a long piece of bare copper EXCEPT disconnect one from that shorted bundle. That is the conductor under test. Connect the bundled set to "ground" by use of another jumper. Connect the "black" lead of the megger to the bundled set. Connect the red - test lead to the conductor under test. Do the 1 min test. Record the value.

You just tested that conductor to ground AND to every single other conductor.

Leave that conductor disconnected, and peel off the next conductor under test. REPEAT.


This way you will only be doing the necessary COMBINATIONs (order doesn't matter) and not PERMUTATIONS (order matters). Since you don't care about the test value of cable #1 to cable #79 and then cable #79 to cable #1, you just care about the Cable 1 to the rest, since you already did it the first test, when you get to cable 79 you just need to test that one compared to the remainder of the cables.

 

[W@ttson]

---

If you did it brute force (and not the bundled set) it would be:

190 conductors + GND

---

C(n,r)=?
C(n,r)=C(191,2)
=191!/(2!(191−2)!)
=191!/2!×189!
= 18145

 

[W@ttson]

With the bundled set you only do (190conductors + 1gnd) - 1 = 190 times.

Can see this easier with a smaller sample size:

A B C Gnd

Brute force:
A : B
A : C
A : Gnd
B : C
B : Gnd
C : Gnd
Total 🟰 6
=========

With bundled set:
A : B C Gnd
B : C Gnd
C : Gnd
Total 🟰 3

The down side of the bundled set approach is that when a conductor fails, you don’t know which conductor in the bundle set caused the conductor under test to fail. Was it to one of the other conductors (nick in two adjacent conductors in the same place) or was it to ground (weak point in the conductor under test and leaking to ground).