resistance of water (measuring IR in water)

Thread Starter

strantor

Joined Oct 3, 2010
6,782
I work at a manufacturing facility that makes cables. We will make a reel of single conductor insulated wire maybe 40,000ft long, then submerge it in a water tank and megger it to the water. This is to check for pinholes in the insulation (faults) and get an insulation resistance reading. The (-) of the megger is connected to earth ground, and into the water at one end of a 20 ft pit; the (+) is connected to both ends of the conductor, elevated above the surface of the water. I have beef with this. I was sent out there to troubleshoot their megger not working. I tried several different things with my megger (working) and noticed that I was getting high resistance readings on the water alone. I had always assumed that water had very low resistance, but seemingly not so. I also believe that the resistance of water is going to vary depending on how far of distance between measuring points. My tests with the megger seemed to prove this at the time. So, my theory is that if you were to megger the reel at one end of the tank, near the ground wire, you would read Insulation resistance of the wire, plus resistance of ~1ft of water; then if you were to move that same reel to the far end of the pit, you would read Insulation resistance of the wire, plus resistance of ~20ft of water, which might be double the first reading. This could cause product which should have failed, to pass with flying colors due to being on the far end of the tank (the operator is only looking for a resistance value). I brought my concerns up to an electrical engineer and he defended the principle, saying that resistance of water is negligible and that it is constant throughout any body of water, no matter how far between measuring points. He can't be right, can he? If he were right, then we would be using water pipes to transmit electricity!
 

Wendy

Joined Mar 24, 2008
23,415
Actually your friend is way off base. Water is only conductive with impurities. If you had specified sea water I wouldn't quibble. Pure water (AKA, DI water) is an insulator, but this is not found in nature. However, the exact conductivity is a major variable.

It would make showers an electrifying experience, and kinda unpopular.

Welcome aboard! Katy huh? I think I may have passed by there now and again.

Nope, looked it up, way far south of my stomping grounds. Heard of it though.
 

Kermit2

Joined Feb 5, 2010
4,162
If you are basing the results on current flow then yes, you are correct.

Think of it as a high voltage electric field causing a leakage through an insulator. The water can have equal potential through out since voltage is so high and current flow is so small. There is an "excess" of electrons ready to go and a very high potential electric field just fractions of a millimeter away that would love to have some electrons join it.

In your reasoning the measured body would not be tightly wrapped with insulation, but would be conducting and therefor the body of water would be less many of its "excess" electrons. It could only supply more from the point of supply( a ground connection), so distance from this current source would be a big factor in the reading.

Does that help? I don't know if I'm explaining it right.
 

jpanhalt

Joined Jan 18, 2008
11,087
Maybe the company really doesn't want to find defects, but needs to do something to "qualify" their product for sale. Forty-thousand feet of cable with a defect is a lot of lost revenue. It might be cheaper to fire you, or at least reassign you to jobs more suitable for your talents until you learn what the real motives are.

I am not sure immersion in water, even if it were conductive, is a good way to find pin-hole defects. Ask the engineer if s/he ever heard of bubble point/pressure? How hydrophobic is the insulation? In other words, there is a limiting hole diameter through which water will not pass at atmospheric pressure.

John
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
Nope, not sea water. It's just city water right out of the main, nothing added, ever. In fact I know that salt water is more conductive than freshwater (didn't know DI water is an insulator, though) and I reccommended adding a couple hundred lbs of salt, but he didn't even entertain that idea. I gather that it's a "We've done it that way for 40 years" thing where they just assume that they and those before them haven't been wrong all along.

So, theoretically, if it were salt water, would the resistance increase as distance increased, or would it stay the same no matter which 2 points you measure in the tank?
 

Wendy

Joined Mar 24, 2008
23,415
Yes, but the calculation is quite complex. Even in 2D network resistance is extremely complex, but it boils down to ohm's per square (if you break the shape down to squares, every square, no matter what the dimensions, is the same). I assume (a bad word that) that in a 3D matrix it is ohms per cube.

So upon further consideration, the length isn't as important as I thought. With seawater the resistance would drop way down.

Old fashioned high power lasers use DI water washing over 300VDC terminals about one foot apart on a xenon tube (flash tube) to pump the laser. The tube is fired continuously, which is why the DI water is used for cooling.

I tend to agree that someone is justifying their jobs with this test, if the conductance of the water is not firmly established then it is a waste of time.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
I doubt that my company intentionally/apathetically produces a substandard product for 1 reason: they use all of this in-house. These conductors, once they go through the spark test, water pit, et. al., go into large cables, which are not sold to customers, but used within the operation of the company.

So I plan to do a study and present it the chain about just how useless this is. My proposition will be to
1. install a water resistivity meter such as :http://www.horiba.com/application/material-property-characterization/water-analysis/water-quality-electrochemistry-instrumentation/benchtop/details/ds-51-ds-52-benchtop-conductivity-meter-360/
2. Create cylindrical wire mesh cages that will encase the reels and be attached to ground, so that the maximum distance that current passed through the water is known ("distance constant" - measured from center of reel to inside of cage).
3. Have the operator perform a calculation for each reel measured: Measured Mohms - (distance constant * resistivity) - I think this should allow us to air on the side of caution and give conservative final measurements

-if they seem at all interested, then I may recommend:

4. add a non-corrosive electrolyte to the water to further reduce the effect of water resistivity effecting our measurements.

-or-

1. stop wasting time dunking reels in water
 

DonQ

Joined May 6, 2009
321
How about dunking a known defective length of cable in the water? Doesn't need to be 40000 ft, just a loop off the end of a reel with a hole poked in the outer jacket. I suspect it would test as 'good'.

This 'water is a conductor' nonsense is reinforced by TV shows and movies that show that anybody touching water anywhere near electricity are instantly electrocuted to death in a shower of sparks. People need to realize that these are stories, told by people with absolutely no idea of what they're talking about. CSI, and such shows are the worst culprits. No one should base their science by what they see on TV.

Yes, still stay away from water and electricity, you don't know what else is in the water. But depending on a tank of water to find pinholes in a cable's jacket is pure fantasy, for a number of reasons besides the most obvious ones.
 

jpanhalt

Joined Jan 18, 2008
11,087
So I plan to do a study and present it the chain about just how useless this is.
Good idea. Be sure your supervisor knows about it ahead of time and buys in.

My proposition will be to
1. install a water resistivity meter such as :http://www.horiba.com/application/material-property-characterization/water-analysis/water-quality-electrochemistry-instrumentation/benchtop/details/ds-51-ds-52-benchtop-conductivity-meter-360/
2. Create cylindrical wire mesh cages that will encase the reels and be attached to ground, so that the maximum distance that current passed through the water is known ("distance constant" - measured from center of reel to inside of cage).
3. Have the operator perform a calculation for each reel measured: Measured Mohms - (distance constant * resistivity) - I think this should allow us to air on the side of caution and give conservative final measurements
Some good operational ideas there, but first define what is a defect? That is, what is the largest hole/defect that would still be considered OK. Run positive (wire with defects) and negative (wire with no defects) controls. Since your product is assumed to be mostly "good," you may want to look at many more negative controls than positive. False positives, i.e., finding a defect when none really exists can be extraordinarily expensive.

-if they seem at all interested, then I may recommend:

4. add a non-corrosive electrolyte to the water to further reduce the effect of water resistivity effecting our measurements.
Don't know how much conductivity you really need. Carbonic acid in water is a poor conductor, but probably better than water alone, and it totally evaporates. Ammonium carbonate is also totally volatile and would give better conduction. You probably want to avoid chlorides, like NaCl, but note that ammonium chloride is completely volatile. There are also some volatile weak acids to consider, such as formic acid and acetic acid. If you don't need to remove the electrolyte from the wire after the test, then there are lots of options.

I am a little pessimistic that water immersion is a very good
test method. Bubble adherence to the wire is going to be a problem. Ultrasonic might help. Have you considered something like HF electrical discharge (e.g., Tesla coil). A defect in the insulation would show as an increase in current. Or, is that what you mean by a spark test?

John
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
Good idea. Be sure your supervisor knows about it ahead of time and buys in.



Some good operational ideas there, but first define what is a defect? That is, what is the largest hole/defect that would still be considered OK. Run positive (wire with defects) and negative (wire with no defects) controls. Since your product is assumed to be mostly "good," you may want to look at many more negative controls than positive. False positives, i.e., finding a defect when none really exists can be extraordinarily expensive.

A "Defect" in this case would be a low Insulation Resistance reading, possibly caused by thinner than specified insulation or conductor not centered in insulation. We have threshold numbers which we look for, but these numbers are likely generated from water pit testing also. Not really looking for pinholes in this phase of the testing, as by now it has already gone through "Spark Testing"....

Don't know how much conductivity you really need. Carbonic acid in water is a poor conductor, but probably better than water alone, and it totally evaporates. Ammonium carbonate is also totally volatile and would give better conduction. You probably want to avoid chlorides, like NaCl, but note that ammonium chloride is completely volatile. There are also some volatile weak acids to consider, such as formic acid and acetic acid. If you don't need to remove the electrolyte from the wire after the test, then there are lots of options. Thanks for the recommendations. I am no chemist, so I really didn't know where to start looking for an electrolyte. Words like volatile, acid, and chloride scare me. I will consult with one of our engineers here on this. The electrolyte will need to be removed from the wire; it will need to be noncorrosive. Also, having a docile-sounding name would be a plus since my supervisors don't know anything more about chemistry than I do, and if I go up saying that I want to add something like "morbic acid sulfide" it might slam the door on the whole thing.

I am a little pessimistic that water immersion is a very good
test method. Bubble adherence to the wire is going to be a problem. I never thought about bubbles being a problem, but you are right. The wire is completely covered in bubbles when it goes down, and it stays that way. Ultrasonic might help. Ultrasonic, as in a vibration induced into the water to knock all the bubbles loose? Have you considered something like HF electrical discharge (e.g., Tesla coil). A defect in the insulation would show as an increase in current. Or, is that what you mean by a spark test? ...Spark testing: The entire reel is spooled off onto another reel, running through a Spark Tester between. The spark tester is an electrode (big wad of metal bead curtains) with 15kv 3khz applied to it. The (grounded) conductor runs through the electrode, and any current going to ground indicates a fault and the machine will stop. Not sure if it uses a tesla coil for that.

John
Thanks for all your input!
Charlie
 

jpanhalt

Joined Jan 18, 2008
11,087
OK. If I get what you say, spark testing finds pinholes. The resistance testing is for finding non-uniformity of insulation, but still insulated.

Ultrasonics could help get rid of the bubbles, but a surfactant that would facilitate wetting the surface might be better. Volatile simply means it evaporates readily. Water is relatively volatile. If words such as chloride scare you, then be sure you never eat at McDonald's.

Now one question you need to consider is whether anything you add to the immersion bath will need to be removed by a second or third wash bath. I suggested using something volatile as that would avoid those wash steps. If the wash steps are not an issue, that is, the surfactant/electrolyte can just be left on the insulation, then the choices are much greater.

John
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
I did some testing, measuring the resistance of the water from 1 corner to all the other corners and at random spots in between and at random depths. I got readings between 3KΩ and 750KΩ (20ft). The spec for the IR is 15GΩ/ft. We usually run 30Kft reels (450GΩ). so, the maximum possible error caused by water is 0.15%. Even if it were DI water (18.2MΩ * 609cm (20ft, longest possible distance) = 11GΩ) the max possible error would be 2.46%.

I'm glad I crunched the numbers before I blew the whistle or I would look like an idiot right now. I still believe that all the things we discussed before would be necessary if we were testing wire with a smaller IR spec.

BTW does it make sense for an IR spec to be "per foot?" Seems to me that (assuming the wire is uniform from end to end) the IR along any 1 ft of wire is going to be the same along any other 1ft of wire, and does not need to be multiplied. It can't hurt anything, only makes for a higher spec, but wouldn't it be better to say that the spec is 450GΩ for any length of wire? Or maybe my understanding of this is a little off.

Thanks for the discussion, I have learned alot.
 
You are a living organism and if you want to keep on living, treat water as a conductor and not an insulator. Only about .007 amperes of current is sufficient to stop your heart. The most affordable conductor of electricity found on this planet is copper, which is about 8000 times more conductive than water. This doesn't mean that water does not conduct electricity. A little bit knowledge can be a very dangerous thing.

Regards, DPW [ Everything has limitations...and I hate limitations.]
 
The first thing that comes to mind is that your latest statement is correct. If the water tank stays the same (water not changed every day) then it is possible to do a reference testing setup with your meter. Meaning that you take the clean water reading and deduct it from your cable-in-water reading and you should have a fairly accurate idea of the resistance. The clean water reading should be re-done if water is added or every 4to5 days due to evaporation. (Which increases the concentration of dissolved salts and therefore the conductivity)
Regarding a previous posting, it is absolutely true that pure water is not conductive at all. Pure water ,obtained through high pressure reverse osmoses is used in high pressure pumps to wash down the sea water (that is conductive) from high tension power lines next to the coast. The sea water damp settles on the wires and more importantly the insulators on the towers and can cause shorting at voltages in the 130Kvolt range.
Bottom line, be very careful with electricity , particularly AC, and "normal" water.
 
Top