316L Cone Fitting Leaks

[goldenspark]

We have a cooling system problem where 60 degree cone type fittings are leaking. We have systems that do not leak, and the common factor appears to be Ethylene Glycol in the systems that leak. The coolant is de-ionised water at about 0.5uS/cm.
I have heard that both de-ionised water and glycol are "searching" liquids that will find their way out. The leaky systems are showing carbon deposits on the conductivity probe (carbon electrodes), to the extent that the probe gap was bridged competely. The pump seals are leaking and are a standard carbon loaded type, which I have also since found out is a no-no for de-ionised water as it leaches the carbon (antimony loaded seals are preferred).
We are looking at two options. Either remove the Glycol, or change the steel to steel cone fittings.
Does anyone have some words of wisdom on this?

 

[moltenmetal]

More details please:

1) "cone-type fittings": I presume tubing here. Do you mean compression fittings, or "cone and thread" fittings? A brand name would also be helpful.

2) What's the material of construction of the leaking fittings?

3) What are the temperatures and pressures involved?

There's nothing magical about deionized water or glycol as far as leakage is concerned. For the pressures I imagine when I hear "cooling system", there should be no problem getting the right materials and a good fitting design to seal reliably if installed properly. We use compression fittings AND cone and thread fittings successfully for hot hydrogen, which makes sealing against DI water or glycol look very easy in comparison.

As to your pump seal problems, I trust there are far more knowledgeable people to deal with these than myself. But I've never heard of graphite dissolving in DI water. A pump seal has numerous components in it: graphite is one seal surface in this case- what's the other, and is it compatible with DI water and/or glycol solutions?

 

[goldenspark]

hello moltenmetal,thanks for your reply.

The fittings we are using are 316L stainless. We have a stainless manifold with welded bosses, each with internal thread and a flat face.
The first part of the fitting is screwed into this with a Dowty washer to seal. This joint is not leaking.
The other end has an outside thread with a male cone in the middle. The mating half of the fitting has a crimped hose (reinforced silicone rubber) on one end, the other end has a free-rotating nut and female cone. The nut pulls the female cone onto the male cone attached to the manifold. It is the cone-to-cone joint that is leaking.

The temperatures are approximately 40-60C with gauge pressures of about 4 bar maximum. Static leak tests had been performed to about 7 Bar to ensure the system had no leaks.

The pump seals are graphite loaded. I think the graphite is being freed into the system by dissolving of the bonding material, not dissolving the graphite.

We are looking at two possible mechanisms for the leaks.

1) The Glycol is breaking down and the acidic properties are corroding the gap between the cones.

2)The carbon has definitely disrupted the electrodes of the conductivity probe, possibly causing it to over-read, meaning the de-ionised water is scrubbed clean to maybe 0.1 uS/cm, making the water/Glycol more agressive as a solvent and / or more searching, which is allowing it to find a way out between the steel/steel joint.

Both of these seem unlikely. Perhaps there is something we have missed?

We do have equipments that are identical in every respect except they do not have Glycol, and these do not leak. There is a possibility that the "worst" equipment for leaks also had higher than specified Glycol levels.

Any thoughts?

 

[goldenspark]

Another piece of information has come to light. It appears we have copper in the cooling system. The black deposits may be cupric oxide.
It is possible we have release of Cu2+ ions into the water which is causing a reduction reaction at the SS interface, possibly there is Fe contamination on the surfaces to allow this?

The plot thickens!