Improvised plumbing 101

[unixronin]

For today's additional entertainment, we got to vacuum three gallons of water out from under the kitchen sink with the steam vac.  Turned out the flare nut on the hot water line to the kitchen faucet had fractured, and there was a bubbling stream of water coming out the side of the nut.  Turn off water, remove offending section of pipe, and find a crack extending around four sides of the flare nut.  Trim the end of the pipe, remove the flare nut, spend an hour hunting for a replacement flare not of the correct thread.  No joy.

"Hmm," says I, looking at the nut.  "Maybe I can weld this."  Five minutes with a file cleans up the nut and runs a V-groove around it along the crack.  "Hmmm," says I, "that filed too easily."  Check with magnet.  "Fuck!"  It's not galvanized iron, it's a zinc die-casting alloy (probably why it fractured in the first place).  Can't weld it.

OK, on to plan B, blanking everything off for tonight so we can turn the water back on and just not use that faucet.  Predictably, although we can find two extra flare nuts that match the fitting on the supply end, at the T down by the dishwasher, we can't find a blanking cap to fit it (though we find several that would fit the faucet, as if that did us any good).

"Hmmm," says I, eyeing one of those flare nuts.  "If I just had a piece of copper sheet, I could make a blanking disc for that flare nut."  Of course, there's no copper sheet to be had.  Except what's that?  A spare end of nice bright shiny copper pipe.  And what's a pipe but a rolled-up tube?

Ten minutes later, I've sawn open the side of the piece of pipe, unrolled it, hammered it out flat, cut out a rough half-inch octagon from the resulting inch-square sheet with tinsnips, filed it into a half-inch circle that fits perfectly into the flare nut, filed it smooth on both sides, and annealed it with a propane torch and a bucket of water^[1] (twice, for luck).  Come back in, install the flare nut with its brand new blanking plate, tighten it up, turn the water back on, and it doesn't leak a drop.

I still need to go out and buy a new flare nut for the faucet end of the pipe tomorrow.  But we have water again for tonight.

---

[1]  Copper behaves, metallurgically, much the opposite way to steel.  If you want to harden steel, you heat it and quench it; to anneal it, you heat it and let it air-cool slowly.^[2]  The reverse applies to copper: heating and air-cooling it hardens it, while quenching it rapidly anneals it and leaves it nice and soft.  Amateur plumbers should also remember that copper work-hardens rapidly, so when you're working with flare fittings, for best results anneal the end of the pipe after you flare it.  You'll get a tighter, more leak-free seal.  (If working with compression fittings, and it's a copper olive, anneal the olive, but if it's a brass olive, don't push your luck.)

[2]  This is true of MOST steels.  Some special alloy steels have very different behaviors.  (If you're really curious, ask me about unorthodox applications of maraging steels sometime.)

Comments

[ilcylic.livejournal.com]

I'd like to hear about unorthodox applications of maraging steels.

-Ogre

[unixronin.livejournal.com]

Well, the key thing about maraging steels is they are designed to work-harden. Extremely rapidly. Like, one forging operation work-hardens the steel to the point that you can't work it any more. There was a really cheap little anti-armor missile called Atlas that made use of this property.

Basically, the missile had a one-piece maraging-steel body into which what amounted to a crosswise smoothbore gunbarrel was forged in a single operation. Into the base went a solid rocket motor, into the nose went an infrared detector, and into the barrel went a shaped charge with a tungsten liner. It was totally unguided, and was designed to be fired across the battlefield a couple hundred feet up. It was spun for stability via offset rocket nozzles, and the IR detector looked at an angle off the nose, which caused it to sweep a helical path and scan the battlefield in successive stripes as it flew. When it detected the IR signature of a tank, it waited a calculated number of milliseconds, then fired its shaped charge, which created a self-forging projectile from its tungsten liner and blasted it through the lightly-armored top of the tank at (iirc) something on the order of 8,000 feet per second.

What you ended up with was a very cheap, very simple-to-manufacture missile that could could afford to fire across a battlefield covered with advancing Russian tanks in swarms of a hundred or so at a time. It didn't matter if half of them missed their targets or fired on tanks that had already been hit, because when you could afford to saturate the battlefield with these things, the odds were any motorized vehicle on the field was going to get hit at least once. What's more, they didn't need line-of-sight for targeting -- they were indirect-fire weapons, so you could lob'em over from behind cover, protected from direct return fire. And you could afford to saturate the battlefield with them, because the use of maraging steel made such a simple design workable.