Mailport February 2014 Issue

Mailport: February 2014

Hose Clamps

The magnetism test that reader Ed Carter refers to in his Mailport letter was only part of the hose clamp test. A series of bench tests also determined the corrosion resistance (left), clamps thickness, and compression ability.

It is very laudable that you would undertake the testing of hose clamps, but there are several points in your article (PS, February 2013) with which I take issue. As a retired mechanical design engineer with a good deal of material experience, I know that the review of such a simple device is surprisingly complicated.

First is your conclusion that the magnetic properties indicate corrosion resistance. It is true that annealed austenitic stainless (304, 316, and several others) is essentially non-magnetic, that non-austenitic materials are magnetic, and that in general, the austenitic materials have better corrosion resistance. This, however, overlooks that some non-austenitic materials (such as 410 and 17-4) are magnetic and have sufficient corrosion resistance, and that the degree of cold work in the manufacturing process causes most common types of stainless (304 and 316) to become magnetic. This cold work does not reduce the general corrosion resistance of the materials. Both stainless 304 and 316 are available from the mill in varying degrees of cold work (annealed, 1/4 hard, 1/2 hard, and full hard), and any of these could be used for some parts of the clamp (from the stand point of corrosion), even though the higher hardness types would be magnetic.

For a demonstration, just touch your magnet to a piece of rigging wire, which is cold worked to about the full hard condition. Some manufacturers use cold-worked material where the parts must endure high service stress, such as rigging wire. Also cold work, to some degree, unavoidably comes from the fabrication process. For hose clamps, this could show up in the formed screw housing or the rolled thread on the screw. One manufacturer could anneal the parts after forming, which would return the part to the non-magnetic condition, while another manufacturer (using the exact same material) could leave his parts in the cold worked/magnetic condition. There would be no significant difference in general corrosion resistance. So the degree of magnetic attraction is not a valid indication of corrosion resistance.

Another point about corrosion that appears as surface rust shortly after subjecting the item to corrosive conditions is that during the fabrication process, the tools (drills, shears, etc.) that come in contact with the parts cause small amounts of iron particles to be transferred to the part. It is common to have this iron removed at the end of the fabrication using processes called passivization or electro-polishing. Parts that get less of this post processing will rust a little, but this is actually just the iron particles rusting. It looks bad but has no detrimental effect on the general corrosion of the part. Only a detailed examination can tell the difference between surface rust and general corrosion early on.

The bottom line is that in most cases, hose clamps do not fail from general corrosion, but from stress corrosion cracking (SSC) at the inside corners of a slot in the band where the screw engages the band. The T-bolt and ABA types should have a distinct advantage here. We also frequently see massive corrosion of the screw, but this usually doesn’t cause failure of the clamping function, just the inability to release the screw. I have several times seen clamps where the screw was unusable due to corrosion of the head, but the clamp was still holding tight on the hose.

SSC happens when three conditions are present: a susceptible material, a corrosive environment, and high static stress. Since 304 and 316 are susceptible (some other types of more expensive austenitic steels and titanium are not) and boats live in a corrosive environment, then we should reduce the static stress whenever possible. For hose clamps, this means not torquing the screw excessively. Since for most critical hose clamp applications the pressure is less than 2 psi, the screw torque to seal is very low, and, actually the fit of the hose on the nipple will often prevent leaking. I have seen numerous examples of a clamp band completely broken, and there was no leaking at the joint.

The additional important goal of the clamp is to keep the hose from pulling off the nipple. Unlike the presumption of your article, this should be the lowest screw torque that reliably prevents the hose from being pulled off the nipple – thus protecting the clamp from sudden catastrophic failure of the band due to SSC.

My suggestion for installation is: tighten the screw just snuggly; take a healthy pull on the hose; if there is no detectable motion, it is tight enough; if motion is detected, increase the screw torque a little and repeat the pull test.

Ed Carter
Bold Venture, Islander 37
Seabrook, Texas

Excellent letter. We’ve covered most of the stainless topics before, but you’re absolutely right about the pitfalls of the magnet test, and we should have mentioned them in the article. However, we still believe that for many marine products, the magnet test is useful for the average consumer.

The recommended torque for hose-clamp screws that is stated in the article is consistent with what you are saying, but in our experience, this won’t always work with exhaust hose, which is the size that we tested.

Next: Frothy Coffee

Comments (1)

Can the author of the above article on the H2out filter provide me with a little more detail. It looks like he has a poppet check valve, exactly what one? Plus I noticed on the article in the January edition that their is a valve or another connection on the same line as the filter. I have tried to find more information on check valves but there are so many and this me in the letter above is a three way valve. Love to have the detail!
Neville Weir

Posted by: Neville W | April 9, 2015 9:46 PM    Report this comment

New to Practical Sailor?
Register for Free!

Already Registered?
Log In