Dielectric joints in hot water system - please help me understand.

[martin224]

I have a question that I imagine corrosion engineering experts will find trivial - but it's puzzling me.

I'm living in France and have just replaced my "chauffe eau" - a 200 litre hot water tank with built in 2.4kW electrical heater and thermostat.

The instructions say that the tank should be installed with an electrically insulating joint where the hot water outlet pipe is attached. I bought a pair of joints (they are only sold as pairs) and their instructions said that one should be installed on the cold inlet and one on the hot outlet. They are sold as prolonging the life of the tank. I have dutifully installed these joints. They are made of plated steel, with a tough plastic gasket preventing the two ends coming into contact.

I assumed that the function of the "dielectric joints", as they are called, is to prevent the steel tank being corroded by electrolysis resulting from a short-circuited steel-copper cell, the hard water in the system serving as the electrolyte.

However, for electrical safety, the hot and cold pipes, and the tank, are all electrically bonded to the house's ground connection. So there is a low resistance connection across the insulated joints. The steel-copper cell is short-circuited.

Please enlighten me. How can the insulated joints serve a useful function when there is a low resistance connection ensuring that both sides of each joint are at exactly the same voltage?

Martin

 

[cloa]

Sounds like an order of magnitude question. The domestic electrical supply is 100 or 240V and supplies up to 10A and the corrosion potentials 0.3V from each other and some tiny difference from the connection ground and if corroding then supply a few mA. A wire connecting the water tank is too tiny to make a difference to corrosion reactions and unlikely to keep the whole lot at ground potential. Only big solid metal pipes and tanks can electrical bond such tiny currents. A difference matter if it was other metals with bigger potential differences and much larger corrosion currents.

 

[martin224]

Thank you. The tank is bonded to the copper pipes by around 60cm of 2.5mm sq copper wire. This will give a resistance between the tank and the copper pipes less than 0.01 ohm. Even a galvanic current as big as 10A would only result in 0.1V or less voltage difference.

I'm still puzzled why they recommend the use of dielectric joints when the two sides of each joint are connected together via a path whose resistance is less than 0.01ohm. A bit like installing double glazing and then permanently leaving the window wide open.

 

[moltenmetal]

You are quite correct that the dielectric joints do little good if there is an alternative, low impedance connection between the two electrodes. But they will do no harm either- unless they fail due to improper installation.

One thing that might be positively affected is the distribution of the corrosion. Depending on how and where the ground connection is attached to the tank, the localized corrosion that is usually encountered at a galvanic couple may be spread out over a larger area of the steel tank, giving you a longer time to failure.

 

[martin224]

moltenmetal (Chemical)
27 Sep 12 10:44
You are quite correct that the dielectric joints do little good if there is an alternative, low impedance connection between the two electrodes. But they will do no harm either- unless they fail due to improper installation.

Many thanks for confirming this.

They do no harm - but I'm left with feeling of having been taken for a sucker having now twice spent 20 euros and the time to install them (one in 2007, once now).

I think they must be sold on faith in their usefulness - like taping a magnet to the main water pipe to "prevent" calcium carbonate deposits.

If corrosion of the tank due to galvanic currents really is a concern, maybe connecting it to the hot outlet and the cold inlet via a meter of suitably rated polythene pipe would have a useful effect. Any galvanic current would then have to flow through a meter of water, rather than just a centimeter or so. I imagine that this would multiply the resistance in the circuit by around 100 times, with a corresponding reduction in any galvanic currents.

Thanks again.

 

[TomDOT]

Interesting, the dielectric joints on my water heater are more like 6-7cm long, not 1cm.

 

[stevenal]

The answer is to use one of these:

http://www.dairyland.com/insulated-joint-protection.html

It will stop the DC while allowing a path for AC fault current. Might be overkill for a typical hot water system, perhaps a simple marine galvanic isolator (

http://www.yandina.com/galvanicIsolator.htm)

would keep the anode from wearing out too soon. Don't know if the electrical inspector will approve.