Dan Maloney
hilarious
-1 = 50meters or more underground
160 meters where at point 4 of LHC
And here I was thinking I was being overly enthusiastic with my descriptions of working at CERN
@Dan Maloney Lol.
my ears popped when going down there
they installed a counter in the new elevators :-)
SPS is different. "only" tens of meters, up to 60 I think. But 24 storeys
ever thought about making it faster so you could experience zero-g?
Awesome.
they are very very fast
At least there's a bathroom right off the elevator
Hehe
@Thomas Shaddack funny enough, exactly the opposite. The elevators are not used very often, when the machine runs, there is no access for months. In this particular one, I got stuck about 15 seconds down every time I was called
first time panic. Second time not again... Third time. ok let's wait 20 seconds, it will unstuck itself
and have to be triple redundant in their operation because huffing up the stairs in case of an emergency isn't really an option
what are the emergencies? helium/nitrogen leak from the magnets?
actually CERN is the only place in the world, where you must use the elevator in case of fire. And you must not use the stairs. 160 meters of staircase is good 50 storey skyscraper
can you run it up in one go?
isnt there a dedicated fire crew for the complex?
What kind of radiation protection do you use?
I'd guess "couple dozen meters of solid rock shielding"...?
@Thomas Shaddack yes, sometimes there is a helium release from the cryogenic system. It is usually collected via the recovery line, valves unfreeze and we carry on. But sometimes there is bigger loss of helium into atmosphere
there's a fob everyone I saw wearing around their neck that measures exposure
@caladan yes, everything is 100 meters underground. You need to shield the machine from the cosmic rays and the human activity too
And the protection for people who have to access the machine?
https://photos.app.goo.gl/92vYVaMHXYeyVYS46
yes, dosimetry is everywhere. But the total absorbed dose for all people working there in a year is a small franction of an annual limit for one person. It is very well organized
that is an image of the observation deck in the CMS cavern, the beam tube is within the concrete blocks below
there are multiple layers of protection. Biometry, mechanical barriers etc. Even if a door would be forced, the LHC will stop and by the time the offender would have a chance to get to the machine, energy from the magnets will be extracted
it is not an easy task if the stored energy in one arc is about a GJ. Hude resistors
Yeah, for those who saw "angels and demons" movie - there are indeed retina/iris scanners at the LHC/SPS access doors. Everything else in this movie is BS, but the scanners are true.
the door interlocks are something... there's an rf fob, iris scanners, etc.
and switches of size of a truck. When you stand next to that switch, and it operates when you don't expect it, you may need to change you brown emergency underwear
do you have a picture of one of the switches?
+
let me find one
those are only the switch elements, it is built into an enclosure of size of a shipping container
and few of the resistors. You dissipate all magnet energy there in 1 minute, then you need to cool it down for more than an hours. If there is a stormy weather in the summer, and there are multiple dumps, it takes time to restart
Ah, I've seen those during open days in UA43
Me too! :-)
the red phenolic boxes are the actuators?
Amazing stuff...
yes, that is typical location, In the power converter gallery parallel to the tunnel in all even points
@dallas those are arc extinguishing chambers. You are opening a circuit with 300 superconducting magnets at 12kA current
not a particularly easy task to do :-)
is that chamber within an additional vacuum chamber or immersed in dielectric oil or some such thing?
for some applications, they managed to replace them by IGBT switches. But mechanical is better, very low resistance, so multiplied by 12 kiloamps squared, it is not producing a lot of heat
no, that is all in air
I have seen it to open, it is impressive
quite a pop when that thing does it's job...
yup... when you think about all the energy you must dispose very quickly in a very controlled way, not an easy job. And you do not want any false positives, it must work only when there is a problem
when energized? or actuation not while electrified?
my experience was a test stand, not energized. The real arc is difficult to see when energized, there is no access
everything is designed to be inherently safe.
yeah. one sector of the LHC has a few henries (H) of inductance. And a 14kA nominal current. I leave the calculation of the energy stored in the magnets as the homework :D
the power converters for dipoles deliver 190V/13kA. But resistive voltage on those coils is exactly zero
fascinating :-)
Agreed.
seriously amazing.
the quench protection system must detect millivolt of resistive voltage in a presence of volts of L*di/dt, and in a very radioactive environment. Collegues from the magnet protection gorup have no simple job
it is mostly digital now
How very ironic that I am this late to the hack chat - I was trying out the CERN Choir!
So, I've got to drop off now, but feel free to keep chatting. Daniel, thanks for the time today, this was great. And thanks to Thomas too -- great insights into the workings of CERN, it was a lot of fun. And thanks to everyone for the great questions. Transcript coming up soon.
https://home.cern/news/news/engineering/engineers-refine-protection-system-lhc-magnets
Engineers refine protection system for LHC magnets
This week, the Large Hadron Collider (LHC) was switched off for its second scheduled technical stop since starting to run at the new high energy of 6.5 teraelectronvolts (TeV) per beam. These regular stops allow engineers and technicians to maintain the machine and ensure that all its components are working well.
I'm a Technical Student, just started this month, so I would've loved to soak up some details here ^^
How is it with these kinds of chats, is the transcript available later?
Yes, I'll be posting a transcript in a few minutes
@dallas guess how many square millimeters of super conductring wire carries those 20kA of current in your photo
hello xasin, welcome to cern :-)
oooh, my host told me that and im now failing the pop quiz. 5?
yes, it is maybe 3-5 mm^2. All the rest is just a resistive filler, like copper, or aluminium. Superconductors are impressive
the situation however dramatically changes, when the superconductor loses its properties and becomes resistive. Then 5kA/mm^2 is not your friend :-)
I remember it being miniscule, especially next to the standard coils used for some of the dipoles.
I heard a story...
Oh, I actually know a little about that! I started in the Quench Detection / Protection systems.
IIRC, when a quench is detected, one of the things you can do is to intentionally make the whole rest of the superconductor ALSO resistive, so that the energy is more evenly dissipated. Was that about right?
Our colleagues are working on high temperature superconducting links to bring current from the converters at the surface down to the tunnel. In a cross section corresponding to your palm they transfer 150 kA in maybe 10 different circuits
@xasin say hello to Jens and company
by high temp, you still mean quite chilly though correct? not high temp as in 30c
@Xasin yes, the coil is warmed up, so the energy is dissipated into a large volume, while rest of the current bypasses the quenched coil by a humungous diode
@dallas just above the critical temperature. 5 Kelvins can be sufficient
5K is hot when you operate at 1.8 :-) everything is relative
ah yes, balmy.
Haha, I will do that, he's my supervisor @daniel valuch
If in the future the uQDS system works better/worse, it's my fault. I'm learning how to change the VHDL code to add some things to it.
working at so low temperatures can be very dangerous. Everything is perfectly thermally insulated. So you can melt a coaxial cable sending 20 Watts of Rf power through it
@xasin LHC had made its first big bang already. Your FPGA will not be the first. Sorry :-)
I would not want to be the one to cause another event like that, don't worry :P
the dielectric changes properties with temperature. can it be used to diagnose the cable state while running?
yes Thomas. We actually do this quite a lot.
you need to use any observables available. Time domain reflectometry, fibers in the cold bores to measure temperature etc.
there may also be quite some overlap with the field of software-defined radio.
everything is software defined radio these days. You try to digitize as soon as you can and then do things in digital. Regardless if RF, or DC metrology. The software does not drift
I think someone in my dept. is working on a very interesting AC analysis of the superconducting magnets too. Wonder how he's doing it...
Definitely digital though, the frequencies are very low there.
In many machines, we already directly sample RF up to hundreds of MHz and directly generate RF too. A nano-second long pulses from pickups are digitized at 12 Gsps and all the rest os done digitally
12 Gsps. WHOA. at how many bits?
yes
What's the data rate of some of the measurements at CERN? I think it goes into the TB/s...
I heard something like 2Pb for a total event data set in ATLAS
I have had a great privilege to work in the electrical metrology section for the past 14 months. That is where the famous HPM7177 was born https://hackaday.com/2021/02/26/homebrew-metrology-the-cern-way/
I have used the "software defined radio" to measure 20kA with a sub-ppm accuracy :-)
Yeah, but that's probably just the filtered data, events that were deemed interesting enough...
Tangentially related, the public CERN Grafana panels are really fun to look at :3
@Thomas Shaddack 12 Gsps, you get 8-10 bits ENOB
that's... VERY impressive.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.