Tevatron Gets the Hiccups

[beamjockey]

Apparently last night's earthquake near Gilberts, Illinois, about 25 miles from Fermilab, quenched the Tevatron. Its magnitude was 3.8.

("Quench" means "go from a superconducting state to not-superconducting." The magnets heat up and boil a lot of helium, which is discharged safely through relief valves.)

There are tilt meters mounted at a few spots in the tunnel to measure ground movement; sometimes distant major earthquakes can be detected. This is certainly the closest earthquake we've ever had. (I slept through it.) The Tevatron can withstand a little jiggling, and earthquake waves tend to be long, so the relative motion of different parts of the machine is generally small. Not this time.

Beam is back, and we expect to resume collisions shortly.

Comments

[wcg.livejournal.com]

Having been present when one superconducting magnet quenched, I can only imagine what that must have sounded like. Dozens of steam whistles all going off at once, I would guess.

[beamjockey.livejournal.com]

Yes. While the Tevatron was a-building, I used to work testing hundreds of magnets. We deliberately quenched each one four or five times to "train" them. It made a glorious noise.

[marsgov.livejournal.com]

I thought the beam quenched when it hit a collimator, not that the magnets quenched — why would they?

From the context of the announcement, the beam quenched, not the magnets.

By the way, I prefer to say, "loss of anti-matter containment field."

[beamjockey.livejournal.com]

I thought the beam quenched when it hit a collimator, not that the magnets quenched — why would they?

Because each proton is carrying 1.6 ergs of kinetic energy. Count a smidgen more for each antiproton that hits. And there are trillions of them. That's a lot of heat. Over a megajoule.

When they hit something, they make a shower of Greek letters, sharply peaked in the forward direction. You can't confine the spray to a collimator in the Tevatron pipe-- it's smeared out in a long line for a considerable distance downstream. That's one reason we lock everybody out of the tunnel before we start up the machine.

Even a little bit of this, hitting a superconducting coil, produces local heating. Local heating, in a coil that's carrying 4400 amps, soon becomes global heating.

Clever circuits may detect the sudden rise in resistance and fire heaters, dumping a cap bank to warm up the whole magnet above superconducting temperature and create one big resistor. This is better than having one resistive hot spot with runaway rising temperature. This is a shock to the magnets, but they're built to take it. You can kiss your liquid helium goodbye. It boils and flashes into "steam" instantly. Cue the relief valves.

The refrigerators will cool down the magnets and you'll be running again in a few hours.

(As you probably know, the quench protection at the Large Hadron Collider did not work properly on 19 September 2008, and CERN was off the air for more than a year.)

By the way, I prefer to say, "loss of anti-matter containment field."

This is not incorrect.

[marsgov.livejournal.com]

Ah, thanks, I knew about the heat dump of a beam defocus but my brain forgot that a collimator would be a magnet in this case.

Just a few hours to cool down? That's pretty impressive.

[marsgov.livejournal.com]

If the magnets had quenched I believe that there would be no assurance that Fermi would suffer no additional downtime.

[vicarage.livejournal.com]

Was it designed to cope with something like the New Madrid quake?

[marsgov.livejournal.com]

Yes, the bean refocuses to create a small black hole, which is then vibrated by the so-called "wiggler" magnets to produce gravitational waves. These are coordinated by a network of sensors to cancel out the effects of the earthquake.

[vicarage.livejournal.com]

I thought you'd tune it to focus the energy at unfriendly nations.

[marsgov.livejournal.com]

There are friendly nations?

[beamjockey.livejournal.com]

Was it designed to cope with something like the New Madrid quake?

No.

If a bigger quake hit and there was a lot of movement of the magnets (fractions of a millimeter or more?), first, we would probably lose the beam and quench, see above.

Second, if things got jostled enough to disturb electrical, mechanical, cryogenic, or vacuum connections, we would need to fix those.

Third, we might need to check alignment of a thousand magnets, and tweak them back into place.

SLAC has quakes. They endured the 1989 Loma Prieta quake and got themselves running again (PDF).

We are not as prepared as they are for a quake, but then Chicago is pretty far from New Madrid, and I don't expect even a big quake there to be very big when it reaches Fermilab.

[isherempress.livejournal.com]

Chicago is pretty far from New Madrid, and I don't expect even a big quake there to be very big when it reaches Fermilab.
Are you serious? I thought the New Madrid quake of 1812 rang church bells in Philadelphia and Boston. Wasn't it like "the Big One" of that age? It is difficult to imagine that something which could change the course of the Mississippi river would not impact your lab.

[brian-s.livejournal.com]

I just wanted to say that "Quenched the Tevatron" would be a great name for a band.

[gomeza.livejournal.com]

Spooky, but good that no permanent harm done.

Question: when a quench happens, is the liquid He recovered or vented to atmosphere?

[beamjockey.livejournal.com]

Question: when a quench happens, is the liquid He recovered or vented to atmosphere?

It's recovered, unless there is too much of it, such as when a large number of magnets quench at once; in that case, it is vented to atmosphere. A cryo tech I once knew called that "giving the helium back to God."

[purpleranger.livejournal.com]

Assuming that God wants it back, of course.

[purpleranger.livejournal.com]

I'm guessing that this was not a particularly good situation. Something along the lines of Dr. Venkman's comment of "I thought you said crossing the beams was bad."