Water in transformer oil

Its turbidity will increase. How's that?

Full-length sightglass.

I looked it up and I believe it was a Dielectric tester, but I was afraid it was by color of the oil, which would you say?

The first indication would be turbidity (a slight milkiness). Another test which takes a little longer is to take a few drops and place them on a flat surface and leave for about an hour. Any water in the oil will usually seperate out and show quite clearly as a drop of water on the side of the oil.

Following small chck will serve the purpose.

1) Rube a drop between thumb and fore finger and feel the stickiness. If less can have water.

2) Put adrop on glass plate and check and view. IF water contenet more will try to expand on the surfce.

3)Di- electric strength any way will tell you the water content.

?! I guess if you have enough water in transformer oil to see it in such a small sample, it will already have short-circuited

The oil should be heated and if there is crackling sound then it sure to contain water

Try a crackle test. Pour a sample of the transformer oil on a steel plate that is heated to a very high temp. If is sizzles and crackles then there is a good chance of water being present.

Terrible test question!

What concentration of water? 50 PPM can cause serious damage to the winding paper over time, but you won't see any significant turbidity or discoloration in a sample. A transformer can operate for months with up to several hundred PPM of water, but eventually the water will deteriorate the cellulose and corrode the core steel, tank welds & assembly hardware.

What's easy depends on what you have available. For me, it's easiest to ship a sample off to a lab & let them do all the work. Any of the following field tests will also work, depending on the equipment you can access:

• For relatively high concentrations, leave the sample untouched in a glass container for 12-24 hours & allow the water & oil to separate.
• For moderate concentrations, place the sample in a glass vessel, with a tube connection to a second vessel. Heat the sample vessel to about 80°C and look for condensation in the second vessel. (If you can't heat it, use a vacuum pump to drop the pressure to <200 mBar & achieve the same result.)
• Place the sample in your freezer for 24 hours, then inspect for ice crystals. The oil will be viscous, but not gelid. Inspection may require a magnifying glass or microscope.

----------Terrible test question!---------------

You may be right--But it was a test question--What is the easiest way to detect water in transformer oil--EASIEST. It was multiple choice and two of the four choices were COLOR and the other was DIELECTRIC TESTER. I don't remember the other two.

I don't work with high voltage/transformers very often and felt deficient guessing at this question and thought that it may be an obvious answer to some members here.

It is supposed to be the easiest, not best, not most conclusive, but easiest. I guessed dielectric tester because I have never seen a sight glass on a transformer, but I haven't seen many either. I had never heard of a dielectric tester--what is it a Megohmmeter that is designed for liquids?

If "easy" means "least amount of equipment and setup", color would be correct. Place the sample in a clear glass tube, then visually compare it to a reference color wheel. Compare that to the dielectric tester, which requires AC power and weighs 75-100 lbs.

The dielectric tester is a just a pair of electrodes in a cup, conforming to a specific ASTM test protocol. The electrodes are usually heavy duty brass to handle repeating arcing, and the gap between them is calibrated. Place the oil in the cup, then raise the voltage until you get an arc. Water reduces the dielectric strength of the oil, so the arc will occur at a lower voltage. This is the unit we use.

Thank you, PWR.

foggy side glass?

You simply take a sample of the transformer oil and do a Karl Fischer moisture content tests.

Simple. Many KF equipment out in the market.

Start first with a dielectric test. the breakdown voltage will give a pretty good indication !!!!!!