Author Topic: The CERN crash from July 2008  (Read 29251 times)

Gunther

  • Member
  • *****
  • Posts: 3585
  • Forgive your enemies, but never forget their names
The CERN crash from July 2008
« on: March 29, 2013, 04:25:06 AM »
My work here in Geneva has made a good progress and it seems that we can reach our goals. That's really good news. So I ave a little break now and I'll try to explain the very complex causes of the crash in July 2008.

The central solenoid is located outside the inner ATLAS detector. The 5 tonne coil contains 9 kilometres of superconducting wire, cooled by liquid helium at a temperature of approximately 2 Kelvin. In July 2008 they had an electric current of 8000 Amperes which produced a magnetic field of 2 tesla. 2014 starts the LHC again and at the end it has an electric current of 12000 Amperes which will produce a magnetic field of 9 tesla. That's very much. The magnetic world record is approximately 100 tesla, but only inside a volume of a few mm2. So, the planed values are not so bad, but it'll be a technical challenge.

But back to July 2008.

ATLAS has 10700 such connectors like in the image above. They have to connect the warm with the cold side (the cold side contains the superconduction). To minimize electrical losses, they should have an electrical resistance f 1 NanoOhm. But 1 of that connectors had 100 NanoOhm.
What did follow?
  • The superconduction crashed.
  • The soldered point did evaporate.
  • At the open ends occured a light arc.
  • The light arc melted down the helium tank.
  • 4 tonnes of liquid helium became gaseous.
  • The mechanical shock wave of the explosion riped out the magnets from the foundation.
  • The shock wave destroyed the isolation.
  • The vacuum sucked in that snippets by a length of 3 kilometres.
That was for short the story behind the 2008 crash. More this evening or tomorrow.

Gunther
Get your facts first, and then you can distort them.

dedndave

  • Member
  • *****
  • Posts: 8801
  • Still using Abacus 2.0
    • DednDave
Re: The CERN crash from July 2008
« Reply #1 on: March 29, 2013, 06:45:22 AM »
great pic, Gunther   :t

that must have been an expensive crash   :icon_eek:
hope their insurance covered it - lol

Gunther

  • Member
  • *****
  • Posts: 3585
  • Forgive your enemies, but never forget their names
Re: The CERN crash from July 2008
« Reply #2 on: March 29, 2013, 08:29:59 AM »
Hi Dave,

great pic, Gunther   :t

that must have been an expensive crash   :icon_eek:
hope their insurance covered it - lol

oh yes, it was a real big crash. Yesterday, I've spoken with one engineer which was involved; in other words, he was a part of the first aid team. 3 weeks after the crash did start the cleanup work. At this time they had -40 C temperature inside the tunnel. That's like Sibiria and a real challenge.

Gunther
Get your facts first, and then you can distort them.

KeepingRealBusy

  • Member
  • ***
  • Posts: 426
Re: The CERN crash from July 2008
« Reply #3 on: March 29, 2013, 10:21:40 AM »
Gunther,

Is there any way to test the connector resistance before applying such large amperages? Any back propogation to the power supply?, Do the connectors deterioate and form higher resistances?

Dave.

Gunther

  • Member
  • *****
  • Posts: 3585
  • Forgive your enemies, but never forget their names
Re: The CERN crash from July 2008
« Reply #4 on: March 29, 2013, 10:51:33 AM »
Hi Dave,

Is there any way to test the connector resistance before applying such large amperages? Any back propogation to the power supply?, Do the connectors deterioate and form higher resistances?

Dave.

The ATLAS has 80 km superconducting cable in magnets. The decision is to check every cable and every connection to avoid such sloppy soldered connection. There's no other way and that's is one reason for the current long shutdown. See also that thread.

Gunther
Get your facts first, and then you can distort them.

dedndave

  • Member
  • *****
  • Posts: 8801
  • Still using Abacus 2.0
    • DednDave
Re: The CERN crash from July 2008
« Reply #5 on: March 29, 2013, 11:08:03 AM »
you could put a sensor wire on either side of the connection
if the voltage across the connection increases, it could trigger a remote SCR or similar semiconductor
bang - early shutdown

dedndave

  • Member
  • *****
  • Posts: 8801
  • Still using Abacus 2.0
    • DednDave
Re: The CERN crash from July 2008
« Reply #6 on: March 29, 2013, 02:33:51 PM »


at the electronics end of the wires, use a one differential amp from A to B and another from A to C
feed the outputs of the differential amps into an analog multiplier
(many commercially available analog multipliers perform XY/Z calculation, and Rc is a constant)

assuming you'd have several of these, feed the outputs of the analog multipliers into a "greatest wins" circuit
the greatest wins circuit consists of 1 diode for each input, cathodes tied together to form a common output
(transistors configured as emitter-followers could be used instead of diodes to buffer many inputs)
use a voltage comparator to trigger a Reset-Set Latch, which shuts the system down

for a small fee, i'll design the circuit   :biggrin:

Gunther

  • Member
  • *****
  • Posts: 3585
  • Forgive your enemies, but never forget their names
Re: The CERN crash from July 2008
« Reply #7 on: March 29, 2013, 06:20:39 PM »
Hi Dave,

that's an excellent proposal. I'll forward it to the CERN staff.  :lol: :lol: :lol:

Gunther
Get your facts first, and then you can distort them.

dedndave

  • Member
  • *****
  • Posts: 8801
  • Still using Abacus 2.0
    • DednDave
Re: The CERN crash from July 2008
« Reply #8 on: March 29, 2013, 10:35:00 PM »
ok - my fee is going up - cost of living, you know   :lol:

Gunther

  • Member
  • *****
  • Posts: 3585
  • Forgive your enemies, but never forget their names
Re: The CERN crash from July 2008
« Reply #9 on: March 30, 2013, 03:31:13 AM »
Dave,

that's clear. You're always a first class costumer.  :lol:

I wrote that message from the train, hoping that it'll arrive the destination.

Gunther
Get your facts first, and then you can distort them.

MichaelW

  • Global Moderator
  • Member
  • *****
  • Posts: 1209
Re: The CERN crash from July 2008
« Reply #10 on: March 30, 2013, 08:14:32 AM »
The decision is to check every cable and every connection to avoid such sloppy soldered connection.

Even without the “sloppy”, considering the application using a soldered joint seems seriously low-tech, and if it was done manually, likely to be unreliable.
Well Microsoft, here’s another nice mess you’ve gotten us into.

dedndave

  • Member
  • *****
  • Posts: 8801
  • Still using Abacus 2.0
    • DednDave
Re: The CERN crash from July 2008
« Reply #11 on: March 30, 2013, 10:29:39 AM »
i am sure the nature of the materials and environmental conditions caused them to do it that way
i wasn't going to question a few hundred of the world's best scientists and engineers - lol

MichaelW

  • Global Moderator
  • Member
  • *****
  • Posts: 1209
Re: The CERN crash from July 2008
« Reply #12 on: March 30, 2013, 11:26:42 AM »
i wasn't going to question a few hundred of the world's best scientists and engineers - lol

Considering what happened, why not? If everyone had done their job correctly, the failure would not have occurred. If the problem was the person that made the connection, then this probably indicates that they needed a design/process that was less dependent on the skill of that person, or at least a reliable way to test the integrity of the connection.
Well Microsoft, here’s another nice mess you’ve gotten us into.

dedndave

  • Member
  • *****
  • Posts: 8801
  • Still using Abacus 2.0
    • DednDave
Re: The CERN crash from July 2008
« Reply #13 on: March 30, 2013, 10:49:03 PM »
the entire project depends on the expertise of individuals, at every step
it's like working on the space shuttle, mars rover, or something similar
all tasks have to be performed to the highest level of standards

but, you are right - they should have caught it in inspection/testing
and, i am sure they updated the test procedure

my guess is that the guy that did that one connection is no longer working there - lol

Gunther

  • Member
  • *****
  • Posts: 3585
  • Forgive your enemies, but never forget their names
Re: The CERN crash from July 2008
« Reply #14 on: March 31, 2013, 02:48:05 AM »
Michael and Dave,

the problem is, that a lot of the needed equipment for such experiments isn't designed and produced industrial, but made by hand from the CERN staff.
For example:

The image above shows a connector for such high electrical currents (several thousand Amperes). It's manufactured at CERN and should be impressive enough. The CERN staff acts as a knit community: they win together and they lose together. Interesting enough: they didn't search the person that made the sloppy soldered connection. The damage occured, it was repaired and they are willing to make it better for the future. That is my strong impression.

Gunther
« Last Edit: March 31, 2013, 10:37:06 AM by Gunther »
Get your facts first, and then you can distort them.