Copper pipe contraction

[GMM]

I've been puzzling about this one today so I thought I'd post here and see if anyone had a view.

At my Dad's place yesterday and there was a panic call from a neighbour. A pipe was leaking on the outside (so not that much of a panic after all ) of her conservatory.

Basically, it was about 10 feet of 15mm copper running along the wall to an outside tap with a compression isolating valve in the middle of the run. From the look of it, the valve had simply pulled off the pipe on one side. I turned the water off, took it all apart and, of course, had to get it stable without there being any parts (a new olive wouldn't have gone amiss!).

Issue number one was that I couldn't get the olive back onto the pipe: I could see a mark where it had sat before (no deformation in the copper as I would normally expect if the olive had 'bitten') but there was no way it was going to slide over the end. A bit of filing and a few whacks drove it onto the pipe, so I could get the service valve back on and tighten it, so then the whole thing was stable enough to turn the water on. Then I realised the pipe it was connected to was a good half inch from the other end of the service valve (with the compression nut on IFYSWIM), so about an inch short of the right length. So I shut the service valve and could turn the water back on to the house and scarper without leaving them with a flood. I doubt they'll need to water the lawn for a while anyway(!) Undoubtedly it'll need some proper fixing now.....not to mention that it clearly should have been lagged in some way by whoever fitted it in the first place.

So a couple of questions: First, why wouldn't the olive go back on? I assume that not being tight before had allowed it to slide off but is it possible that it was (eg) a half-inch olive on a 15mm pipe or somesuch combination? (The pipe didn't look particularly flared or anything) Or was it differential contraction in the cold weather? Second, how could it ever have fitted together in the first place? I know it was cold and a pipe (anchored well at both ends) may have contracted a little in those conditions, but surely it wouldn't have lost an inch over about 10 feet, would it? There was no way I could have breached the gap using the service valve even with the olives just hanging on the very ends of the pipe.

Any explanations, chaps?

[Dave Baker]

The coefficient of thermal expansion of copper is about 17 times ten to the minus 6 per degree C. That means a change in length of about 1mm for a 10' length of pipe for a 20 degree C change in temperature. Even if the pipe had been fitted at the height of summer it couldn't have contracted by an inch in winter.

[andy]

According to my calculations (if I gottem right) if the pie was installed when the temp was 20deg C and the temp when you got there was 0 deg C that run of pipe would have contracted by about 1.6 mm so I doubt that would be the cause. Much more likely the pipes weren't the right length in the first place. No idea about the olive save that brass would have contracted a little more that the copper.

A.

[Dave Baker]

You haven't.

[A.Lee]

They never do go back on easily, or, generally, come off easily. It was pushed off by the expansion.

Yes, you are missing the most obvious thing - the water in the pipe had frozen. Water expands when it freezes, so it pushed toward the path of least resistance, which would have been the end fittings. They do sometimes split the pipe, but the ones I have attended usually push off one of the fittings. Alan.

[Harry Bloomfield]

That is the result of the water freezing in the pipe and the fitting simply being pushed off along with the olive. The pipe and the ridge made by the olive, will have been crushed slightly as the olive was forced off - then once the olive came off the ice probably refroze expanding the pipe back out again. Hence you now have an olive which will not go back on.

[andy]

OK very helpful ;o) Would you like to correct me.

A.

[andy]

Dammit I thought he said 15ft of pipe so actually my calcs were correct but using the wrong data ...... doh. End conclusion still correct however.

A.

[GMM]

I'm sure the freezing explanation is right and that flaring the pipe slightly may account for not being able to get the olive back without a file, but I'm still puzzled about the overall length: The olive made a mark on the 'upstream' joint that was in the right place and the valve was put back almost exactly in the same place, but the gap that then left to the 'downstream' pipe was just way too big to even think about connecting (I'm glad it was a valve rather than a joint, as I could re-fit in , turn it off, and scarper). At the coefficients quoted (ca 20 x10-6 per degree), it seems there's no way it could have contracted that much, even if it was first fitted on a tropical day. Indeed, if it had, it would mean copper would be pretty much useless for all plumbing applications outside of a very stable temperature environment, which it pretty evidently is not.

P'raps it will remain a mystery, known only yo the cowboy who fitted it originally.....

[george (dicegeorge)]

i guess that what was once a straight pipe is slightly curved and therefore shorter in a straight line (imagine tieing a rope to each end like a bowstring then tightening the rope, curving the pipe...)

[g]

GMM wrote:

[Roger]

The message from GMM contains these words:

The reason you had difficulty getting the olive back on was because it was a used olive and had been slightly distorted the first time around. For an olive to grip the pipe it needs to be compressed and it will not return all the way to its original shape when the pressure is released.

One possible explanation for the apparent reduction in length of the pipe might be found in how the other joints in the run were put together in the first place. If they were all made up with a significant gap between the end of the pipe and the end of the bore within the fitting the force that led to the original break being forced open might also have led to the other joints being closed up.

[Dave Liquorice]

How rigidly fixed is the rest of the pipe work? The ice core that forced apart the joint could just as easyly moved somethimg else once the joint was apart. The pipework springing back to just touching when the core melted.