36C3 - The Large Hadron Collider Infrastructure Talk

36c3 preroll music

Herald: Our next talk's topic is the Large
Hadron Collider infrastructure talk. You

probably know the Large Hadron Collider
over at CERN. We heard quite a bit of it

in the recent talks. This time we will
have a deep dive into the infrastructure.

You can assume our next speakers are doing
a great job. Basically, it's pretty

obvious because we are not stucked into a
great into a giant supermassive black

hole. So please welcome with a very warm
applause, Severin and Stefan.

Applause

Severin: Yeah. Hello, everyone. Thanks for
coming. So many people here, quite nice.

In the last couple of years we had several
talks about, yeah, basically the physics

perspective of LHC, how physicists analyze
data at LHC, how physicists store all the

data, et cetera. And we would like to give
like more an engineering perspective of

the whole LHC. So three years ago we had a
talk by Axel about how physicists analyze

massive big data and then last year we had
a talk conquering large numbers at LHC by

Carsten and Stefanie. And we would, as
I've mentioned already, we would like to

give like more an engineering perspective.
We are Stefan and Severin. We're both

electrical engineers working at CERN.
Stefan is working in the experimental

physics microelectronics section and he
will give a second talk tomorrow about

designing high reliability digital
electronics together with Szymon tomorrow

morning at 11:30. And I'm, as I mentioned
already, also working at CERN. I'm working

in the electronics for machine protection
section. I will describe briefly later. A

short disclaimer; the LHC is a pretty big
machine and we try to explain it as good

as possible. 45 minutes is not really
enough to talk about everything because I

think you can basically take one of the
topics we are talking here about now and

talk for 45 minutes only about one
specific topic, but we try to give an

overview as good as possible. So imagine
you want to build an accelerator in your

backyard. OK, maybe not in your backyard
because LHC is quite big, so 27 kilometers

in diameter is quite big, but basically we
figured out three main challenges you have

to take. First of all, we have to
accelerate particles because otherwise

it's not a particle accelerator. Second,
we have to keep the particles on a

circular trajectory. And then third, we
have to make sure that the particles which

are inside our beam tube or beam pipe
don't collide with anything which is

there, for example, to beam pipe itself,
air molecules, etc. And the solution we

adopted for LHC there is, that we
accelerate the particles with a high power

radio frequency cavities. Then we have a
beam control system which is quite

sophisticated using superconducting
magnets and then we have the beam pipe

itself, which is evacuated, so it's under
vacuum conditions to avoid any collisions

we have inside with gas molecules, etc. A
brief overview about the location itself.

So probably many of you know already that
CERN is next to Geneva. So it's in the

western-southern part of Switzerland. When
we zoom in a little bit more, then we have

here an artificial like picture of LHC
itself in the red circle there. To put it

a little bit in a perspective, we have a
relatively big airport there. You can see

there, it's a 2200 metre long runway. We
have Geneva Lake next to it. And that's

only one small part of Geneva Lake, but
nevertheless, and what also quite nice, we

see Mont Blanc from LHC, er, from CERN.
When we zoom in a little bit more, then we

basically have the big, circular collider
there. That's LHC itself. And we have pre-

accelerators, I will talk in a few minutes
about. Basically we have two main

campuses: we have Meyrin Site, which is in
Switzerland, and we have to Prevessin

site, which is in France. Then at LHC
itself, we have eight service points. We

also call it just points, to briefly go
through all of them; we have point one

where we have the experiment called ATLAS,
one of the big and major experiments at

LHC. Then at the exactly opposite side of
ATLAS, we have CMS at point five. Then we

have a little bit smaller experiment,
which is ALICE. It was basically

constructed for lead ion runs. We will
talk about this later. And then we have

another relatively small experiment, it's
called LHCb. And that's the only non-

symmetrical experiment at LHC. These are,
I think, the four experiments you already

maybe heard of. Then there are four or
three other smaller experiments. We have

LHCf at point one, it's a forward
scattering experiment at point one. So

basically, they're taking data like
scattered particles from ATLAS itself.

Then we have TOTEM. It's also a forward
scattering experiment and point five, then

we have, sorry, we have MOEDAL, which is
the experiment at point eight. They're

looking for magnetic mono-poles. Then we
have TOTEM, sorry for that, at point five.

And then we have a relatively new
experiment which is called PHASER. It's

actually under construction and it will be
used, starting from 2021 and it's forward

scattering experiment, which, where they
try to detect neutrinos. Then we have

point four, there we have the RF cavities
to accelerate the particle beam itself. We

have the beam dump area. So when there is
like a fault in a machine or we just want

to dump the beam, then we used the mean
dump system at point six. And then we have

two more general service areas. It's point
three and point seven. LHC would not be

possible without the pre-accelerator
complex. So we have a relatively big one

and it's also sometimes relatively old. On
the left hand side of the slide you can

see LINAC2, it's an old linear accelerator
which was used until last year. It's not

now phased out. And now we have LINAC4,
which is also a linear accelerator and it

has a little bit higher acceleration. Then
we have the proton synchrotron booster.

It's the first circular collider. So you
can see two pictures there. What is

relatively special about PSB is that we
have there two, sorry, four beam pipes

instead of just one beam pipe. Then we
have the proton synchrotron accelerator,

which is the next stage for acceleration.
It then has only one one beam pipe. And

then we are going from PS we are going to
SPS, which is the super proton synchrotron,

which is has circumferences of seven
kilometers. There we basically accelerate

the particles the last time and then they
are injected it in the LHC itself. We

mentioned a few accelerators already,
basically all everything we just

highlighted here. But CERN is a little bit
more. So CERN is famous for LHC, I would

say. But there is much more than only the
LHC. So only about 15 percent of the

protons, which are accelerated in the pre-
accelerator complex, are really going to

the LHC itself. So there is much more:
there is material science, there is anti

matter research and all different other
kinds of research going on. Of course,

everything has to be controlled. It's
called a CCC, the CERN control center.

It's located at, the Prevessin site;
looks like that. Basically, we have, four

Cs looking to each other and there the
operators are sitting 24/7 and operate the

whole machine. So basically, the whole
pre-accelerator complex, all the energy

cryogenics and LHC itself. Before you ask,
everything is running on Scientific Linux.

So we have basically our own Linux
distributed distribution, which is used

there and it of course, it's open source.
Talking about the LHC beam itself, we have

two beams: one is running clockwise and
the other one is to running anti-

clockwise because we don't have a fixed
target experiment where we basically let

the accelerated particles colliding with
like a fixed target, like metal or

something like that. We have controlled
collisions at four points, we mentioned

before. Most of the year we have proton
runs. So we have protons and protons

colliding each towards each other. And
then we have at the end of the year,

nearly starting from November to December,
we have lead ion run. The protons itself

is not really like a fixed, straight line
of particles. We have something called

bunches. You can imagine a little bit like
spaghetti. It's basically the same length

of a spaghetti, but it is much thinner
than a spaghetti. And each bunch, when you

have a proton run, then each bunch
consists of approximately 100 billion

protons. And when you have lead ion runs,
then we have approximately 10 million lead

ions in LHC. And last year we operated
with 2565(sic) bunches in the LHC itself.

The LHC tunnel. We already talked about
the tunnel itself. It is 27 kilometers

long and you can see maybe a little bit on
this graph, that we have some, we have

eight straight sections and we have eight
arcs in the tunnel. Basically the straight

sections are always there, where we have
like service cavities or we have areas and

also the experiments. Because it's not so
good visible in this picture, I put the

picture here. Basically, that's a straight
section of LHC. You can basically just see

the beam pipe itself, with aluminum foil
around it, and there are also no magnets.

And when we look in the arc section of
LHC, then you see here the arc itself and

I think it's quite famous picture of LHC
itself because we have blue dipole magnets

there. The tunnel itself is an old tunnel,
used previously by LEP, the large electron

proton collider. It has a diameter of 3.8
metres and the circumference is

approximately 27 kilometers. Inside the
tunnel we have, first of all, cryogenics,

so we have big tubes, stainless steel
tubes to carry all the cryogenic. So

liquid helium and gaseous helium. Then we
have the magnet itself to bend the

particles and then we have electrical
installations to carry like signals from

the magnets to have safety systems,
electricity, etc, etc. Geography is a

little bit complicated in the area because
we have in the western part of LFC we have

the Jura mountain range and this Jura
mountain range has a relatively hard

material. It's not made out of, not made,
but nature. I mean it's a limestone, so

it's relatively complicated to dig into
this material, in comparison to all the

other areas at LHC. So when you would
basically put a straight section of LHC,

then you have to dig much more into the
relatively hard limestone. So that's why

it was decided that the LEP or LHC tunnel
is tilted a little bit. So we have a tilt

angle of one 1.4 percent there. The depth
is approximately between 50 meters at

point one or point eight, up to 170 metres
deep at point four. We already talked a

little bit about magnets, but we would
like to go a little bit more in the

details now. So why do we need magnets?
Um, maybe you learned at school that when

you have a magnetic field and you have
charged particles and you can bend

particles around an arc in a magnetic
field. Depending on the charge of the

particles, you bend them around on the
right side or left side, that's this

famous right hand and left side rule, you
maybe learned during school. And at LHC we

cannot use a normal magnets like typical
magnets. We have to use electromagnets

because normal magnets would not be strong
enough to build in an electromagnetic

field which is feasible to bend the
particles around the whole tunnel. At LHC

we use, in the dipole magnets, a magnetic
field of 8.3 Tesla. And to do this we need

a current of 11850 amps. We have basically
two different types of magnets. We have

bending magnets. So the dipole magnets I
mentioned quite often already. And then we

have injection and extraction magnets.
They are also dipole magnets there, but

they are a little bit differently
constructed, because the injection and

extraction magnets have to be quite fast,
because they have to basically be powered

up at full, the full magnetic field in
several microseconds. Then we have higher

order magnets which are quadrupole pool,
magnet, sextupole magnets and octupole

magnets, et cetera, et cetera. And they
are used for focusing and defocusing the

beam itself. In total, we have 1200 dipole
magnets at LHC. We have around 850

quadrupole magnets and we have 4800 higher
order magnets. But they are normally quite

short or so, shorter than the the other
magnets. The dipole magnets consist of two

apertures. They are used to bend to beam
around, so I already said. In the middle

of the magnet itself we have a cold bore.
So there is basically there are the

particles flying around. Then there is a
metallic structure. You can see this in

the picture. It is just a shiny metallic
sphere you see there. And then we have

next to the cold bore, we have the tool,
the two apertures to bend the particle and

build the magnetic field itself. The
dipole magnets have a length of 50 meters

and the manufacturing precision is plus
minus one fine, one point five percent,

er, one point five millimeters. Then we
have quadrupole magnets. They are used for

focusing and defocusing the beam. The
problem is that we have bunches, were are

basically equally charged particles inside.
And the Coloumb force tells us, that when

we have equally charged particles, then
they are basically want to fade out from

each other and in the end they would just
hit the beam pipe itself and we could

maybe destroy the beam or cannot do any
collisions. So what we do is we use a

quadruple magnets as, yeah, similar to
lenses, because we can focus and defocus

the beam. The quadrupole magnets, the name
already suggested it, that we have

basically four apertures. So we have on
the left and the right side two and then

we have on top and bottom we have also a
few of them. To go a little bit into

detail about the focusing and defocusing
scheme. In the beginning we have a

particle beam which is not focused, but we
want to focus it. Then we go to the first

quadrupole magnet. So we focus the beam.
And this is only done in one axis. That's

a little bit a problem. So, in the second
axis we don't have any focus and we have a

defocusing effect there. And then we have
to use a second quadrupole magnet for the

other axis, in this case the Y-axis to
focus the beam even further. And you can

even see this here in the Z-axis, that's
basically the cut off the beam itself. You

can also see that in the beginning we have
on the left side, we have an unfocused

beam and then we focus it in one axis, so
we have like a little bit more ellipse and

then we focus in the other direction, then
we have a different ellipse. So we have to

use several quadruple magnets in a row to
really focus the beam in the way we want

to have it. In the LHC magnets, we have
quite high currents. We we need these

currents, because otherwise we cannot bend
to the very high energetic particle beam

and to use normal conducting cable, it
would not be possible to basically build a

magnet out of it. So what we do is, we use
materials which are called superconducting

materials, because they're for very good
effect. They go to basically zero

resistance at a specific temperature
point. And after this point or when we

basically go lower, then the current can
flow without any losses inside of it. But

to reach the state, we have to cool the
magnets quite heavily, which is not so

easy, but it can be done. And on the right
side you basically see a very historic

plot. That was 1911 in Denmark, a
researcher called Heike Onnes detected for

the first time superconducting effect in
mercury. And it was detected at 4.19

Kelvin. To show you a little bit the
comparison between a normal conducting

cable and a superconducting cable, as we
put the picture here. So that is basically

the same amount of cable you need to use
to carry, the thirteen thousand amps and

to do the same or to transport the same
amount of energy we also can use a very

small superconducting cable and I think
it's quite obvious why we use here

superconducting cables. At the LHC we use
Niobium tin(sic) as material. And this

material basically goes into a
superconducting state at 10 Kelvin. But to

have a safe operation to LHC, we have to
cool it down at 1.9 Kelvin.

Superconducting magnets have some
benefits, but also some downsides, so

sometimes they change their state because
there are small rigid vibrations and the

magnet or the temperature's not precise
enough or the current is too high, then

they change that state and it's called
"Quench". And, we basically can detect a

Quench when we measure the voltage across
the magnet, because the resistance changes

at this point. So when there is a Quench
then the resistance changes quite rapidly,

in milliseconds and we can detect this
voltage rise with sophisticated

electronics. On the right side, you see a
board I'm working on. So basically here,

we have a measuring system to measure the
voltage across the magnet. And then we

have a detection logic implemented in FPGA
to basically send triggers out and open an

Interlock loop. Interlock loop is a system
at LHC. You can imagine that little bit

like a cable going around the whole tunnel
and there are thousands of switches around

this Interlock loop. And as soon as one of
the detection systems basically opens to

the interlock loop, then basically the
whole machine will be switched off. And

what means switched off is basically, that
we will power down the power converter,

but then the energy is still in the
superconducting magnet and it has to be

taken out of the superconducting magnet.
And therefore, we use dump resistors to

extract the energy. And here you can see a
picture of such a dump resistor. It's

quite big. It's in a stainless steel tube,
oil cooled. It's approximately three or

four meters long. And basically, when
there was a Quench, and the energy was

extracted via these resistors, the whole
resistor is heated up by several hundred

degrees and it needs several hours to cool
it down again. Power converters; the power

converters are used to power the magnet
itself. So they can produce a current of

approximately 13000 amps and a voltage of
plus minus 190 volts. And you can see a

picture how here, how big it is. One
downside with the power converters is that

they have to be, not downside but one
difficulty is, that they have to be very

precise, because every instability in the
current would have or has a direct effect

on the beam stability itself. So basically
the beam would be not diverted in the

right amount of length. So that's why they
have to be very precise and have to have a

very precise stability. So here I just
pointed out, like in 24 hours, the power

converter is only allowed to have a
deviation of 5 ppm. And in comparison, for

13000 amps we have a deviation of 65 milli
amps. So the power converters have to be

very precise. And to do that, we had to
develop our own ADC, because at the time

when LHC was built, there was no ADC on
the market which was able to have this

precision and also the whole ADC is put
into a super-precise temperature

controlled areas and it is calibrated
quite regularly. Okay, cryogenics. We

already talked about that we have
superconducting magnets and they have to

be cooled down quite low. So the
superconducting magnets we have at LHC has

have to be cooled down to 1.9 Kelvin. And
we are doing this when we like start the

LHC then we cool down on the first hand
with liquid nitrogen. So approximately six

thousand tonnes of liquid nitrogen are put
through the magnets to cool them down

to 18 Kelvin and afterwards we cool the
magnets down with liquid helium. And

liquid helium is at 1.9 or 1.8 Kelvin. And
to put it a little bit in a comparison,

outer space, so when we measure like the
temperature of space, we have

approximately 2.7 Kelvin in outer space.
So LHC is much colder than outer space.

The whole cooldown needs approximately one
month and each dipole magnet, which is 15

meters long, shrinks several centimeters
during that. Which also has to be taken

into account, because otherwise pipes
would break. The cryogenic system is that

we have at each of the eight points at LHC
we have compressors to cool down the

liquid helium or the helium itself. And
then we compress the helium and pump it

down. We have one gaseous helium stream,
which is at 15 Kelvin and we have liquid

helium stream at approximately 4.5 Kelvin.
And then we pump it underground and then

we have something called Cold Compression
System. And the Cold Compression System

even further reduces the pressure of the
helium that we have in the end a helium,

which is at 1.8 Kelvin. So it can really
cool down the magnet itself. And helium

has a very interesting effect because at
2.1 Kelvin, it becomes something called

superfluid. So it basically can run around
like holes, for example, or walls. It can

basically flow against gravity, which is
quite interesting. And it has also very

high thermal conductivity and that's also
why we use superfluid helium here. And

that's why we cool down the whole magnets
that low. And one other interesting effect

is also that the LHC tilt angle, which is
1.4 percent, has to be taken into account

because we have very low pressure inside
all the tubes or all the system, at 16

millibars. But we have sometimes to pump
the helium against gravity or going down.

So we also have to take into account the
LHC tilt angle to not have wrong pressure

levels at the whole LHC itself. Okay.
Stefan: All right! So, you probably

already got the idea, that what we've done
in the last 20 minutes, was only solve the

first of the three challenges we had,
which was actually bending the beam around

the circular trajectory. So I'm trying to
go to the other challenges we have lined

up in the beginning. And the first one of
that is the actual acceleration of the

particle beam. And large synchrotrons,
e.g. like the LHC, they use radio

frequency or RF systems to do this
acceleration. And I'm just going to do a

quick recap of the LHC beam and RF and how
they interact. So Severin mentioned

already that the particles in LHC actually
come in bunches. So in like packets that

contain about hundred billion protons and
those bunches are spaced when they are

running around the LHC approximately 25
nanoseconds apart. And starting from that

the tasks of the RF system are basically
twofold. It first has to ensure that these

bunches are kept tightly together in a
process that we call longitudinal

focusing. And the second task is to care
for the actual acceleration of the

particle bunches. So from their injection
energy, when they come from one of the

pre-accelerators up to their final energy,
that they are supposed to collide at

during the physics run. So in general, you
can imagine RF as being a quickly

alternating electric and magnetic field
components. And in the LHC, this RF energy

is basically injected into what is called
a cavity, which is a resonant structure.

And there the particle beams travels
through, while the field quickly

alternates and the RF signal, or the
energy, basically interacts with the

particle beam. So perhaps you know that
the protons are positively charged and

then a negative polarity of the field
would attract these protons, while the

positive field location would basically
move them away. And this has ... well,

after just injecting and with the
frequency of this RF field being the same

as the speed that the particles actually
go round the LHC, solves the first of the

two problems, which was the the focusing
because actually the particles that are

too slow arrive only when the field is
already changed to the opposite polarity

and actually get accelerated a bit, while
the particles that are too fast, they are

actually being decelerated a bit. And this
is a process that we call the longitudinal

focusing, which makes sure that the
bunches stay neatly packed together. And

of course this would be relatively
inefficient if we would only change the

polarity of this field once for each of
the proton bunches that pass by. Which is

why we do it ten times. So the polarity
basically changes ten times or the

frequency is ten times higher than the
bunch crossing frequency. And by doing

that, we make sure that the change of this
field is much faster and therefore the

particle bunches are packed much closer
together and the focusing is better. So

here you can see these cavities that were
shown in the previous picture as a

schematic, how they're actually placed in
the tunnel. So eight of these huge

cavities are used per beam and they are
the actual thing that is used to couple

the RF energy into the beam and transfer
it to the particles. They are also

operating superconductively, so at
cryogenic temperatures, to reduce the

thermal stress and the losses that would
otherwise occur in their materials. And

these are actually – even though they are
so big, similar to the magnets that had to

be very precisely manufactured – these
also have very small manufacturing

tolerances and have to be precisely tuned
to the RF frequency that is used to

inject. So and the second part of this,
that actually produces this high power RF

signal. For that is used what we call
Klystrons. So Klystrons are basically RF

amplifiers. They are built from high power
RF vacuum tubes and they amplify this 400

MHz signal that is used to transfer energy
to the particles. And each of those

Klystrons produces about 300 kW of power
and you can probably imagine how much that

power for an individual unit that is, if
you know that your microwave oven has like

2 or 3 kW. And of course, as we have eight
cavities per beam and one Klystron always

feeds one cavity, we in total have 16 of
those Klystrons and they are in principle

able to deliver a total energy of 4.8 MW
into the LHC beam to accelerate it. But if

we take a small step back for now, we have
only solved the first of the two problems,

which was to keep the bunches neatly
focused. Because currently the particles

have been injected and the frequency is at
some specific frequency and actually they

are only running basically in sync, the
two. So what we do after all the particle

bunches from the pre-accelerators have
been injected into LHC, is that we ever so

slightly increase the frequency, which of
course also means that the particles need

to accelerate together with the RF signal.
And this is the mechanism that we use to

accelerate them actually. And the change,
that is required to do this, is very tiny,

actually. So it is less than a thousandth
of a percent sometimes, that is used to

change the frequency to actually make them
go so much faster. So from their

relatively low injection energy up to the
top energy plateau that they need to have

to produce the actual physics collisions.
And an interesting question to ask here is

where does this signal actually comes from
if it needs to be so precisely tuned to

some specific frequency? Who generates it
or who controls it? And that opens up the

whole complex of the timing of the LHC, of
the machine. So actually this first signal

that I mentioned, this RF signal, it
originates in a Faraday cage. So an

especially shielded area somewhere on the
Prévessin site of CERN. And from there it

is distributed to the low-level RF
subsystem with the Klystrons and the

cavities. But inside this room, there are
also a number of other signals generated.

The first one of that being this Bunch
Crossing Clock, which is the actual clock

that signals one pulse, basically every
time, it changes polarity one time a

proton bunch moves across a specific
location inside the LHC. And another one

is the so-called orbit clock, which always
indicates the start of the first or when

one proton bunch has basically re-arrived
at the same position and has completed one

orbit. And you may ask the question why
this is an important piece of information.

But if you think back to this image that
Severin has already shown, about the

accelerator complex, the big challenge
that all this brings is also the whole

synchronization of all these machines.
Because you have to imagine that while

these proton bunches run around the LHC
and new ones are supposed to be injected

from the outside, from another pre-
accelerator, this has to be very precisely

synchronized. So all these pre-accelerator
systems actually share a common

synchronized timing system that allows
them to precisely inject a new packet of

bunches at the right position, at the
right location into the LHC. And this a

bit how such a timing distribution system
looks like. It is only a very small

excerpt of what it looks like, but it
gives you an idea that somewhere

underground in the LHC there is rooms full
of equipment that is just used to

distribute timing signals between
different parts of the accelerator. And of

course, as CERN is forward-thinking and
realized that future colliders will need

quite a bit more of all this
synchronization and that the requirements

for how precisely everything needs to be
synchronized is ever growing, they

actually developed their own timing
distribution standard which is also

openly available and available for
everybody to use. So if you're interested,

look that up. But of course, not only the
accelerator itself is interested in this

information about what particles are where
and how quickly they interact or how

quickly they go around. But also all the
experiments need this information, because

in the end they want to know "Okay, has a
collision occurred at some specific time

in my experiment?" and actually providing
this timing information about when bunches

have crossed their experiment locations is
also vital for them to really time tag all

their collision data and basically track
which bunches were responsible for what

kind of event or what event throughout
their whole signal storage and processing

chain, let's say. Good. So that is
basically challenge 2 out of the way. So

that was the acceleration of the actual
particles and all the associated issues

with timing. And the third issue we
mentioned was that the particles need to,

let's say, be kept from colliding with
anything but themselves or the other beam.

And that is what we, why we need vacuum
systems for. So, again, it is not as

simple as just putting a vacuum somewhere.
Of course not. Because in fact, there is

not only one vacuum system at LHC, but
there are three. So, the first two of

those are perhaps a bit less interesting
to most of us. They are mainly insulation

vacuum systems that are used for the
cryogenic magnets. So they isolate,

basically thermally isolate the magnets at
those very cool temperatures from the

surrounding air to avoid them getting more
heat load than they need to. And there is

an insulation vacuum also for the helium
distribution lines that are actually

distributing, delivering the helium to
these magnets. And then the third one,

which is perhaps the most interesting one,
is the beam vacuum. So the one where

actually the beam circulates inside the
LHC. And this is a cross section of what

this beam vacuum typically looks like. So
it is approximately this size, so a very

... handful, let's say. And the question
you may ask "OK, if I want to keep all the

like the particles in my particle beam
from colliding with anything they are not

supposed to, for example, rest molecules
of remaining air there, how many molecules

can there still be?" So somebody has to
make up that number. And typically you

express this as a quantity called the
beam lifetime, which basically says if you

were only to keep those particles
circulating in the accelerator, how long

would it take until they have all
dispersed and lost their energy due to

colliding with rest gas molecules? And it
was decided that this should be at a value

of 100 hours, is what the beams should
basically be able to circulate without

collisions, without being lost. And this
gave the requirement for pressures down to

about 100 femtobar, which is a very small,
very, very tiny fraction of the

atmospheric pressure we have here, which
is about 1 bar. And to actually get to

this level of vacuum, it requires multiple
stages and multiple components to actually

get there. So the initial vacuum inside
these beam tubes, which are basically

going throughout the whole LHC tunnel, has
the volume of approximately the Notre-Dame

cathedral. So the first step of getting
all the air out of these beam tubes is

using turbomolecular pumps. And then there
needs to be more mechanisms to reduce the

pressure even further, because these pumps
are not able to reduce the pressure to the

levels required. And they actually use a
relatively clever trick to do that, which

is the use of cryopumping. So the, ... I
cannot show that? Okay. So the outer wall

of this beam pipe cross section that you
see here is actually also where the very

cold helium inside the magnets is outside
of. And what that does is, it leads to an

effect called cryopumping. So actually any
rest gas molecule that hits this wall

actually condenses there. And as the
molecules condense there, they are of

course removed from the atmosphere inside
this beam pipe, which removes them from

the atmosphere and increases the quality
of the vacuum. And with the use of this

and then the warm sections, the use of
getter coatings, which are basically able

to trap gas molecules, you are able to
reach the crazy vacuum levels that are

required to make this happen. But they
realized also during the design that one

big problem – for the first time in an
accelerator – another effect will create a

significant problem for the vacuum, which
is the generation of synchrotron

radiation. So synchrotron radiation is a
byproduct of when you do bend a particle

beam, it results in a phenomenon called
synchrotron radiation. And when this

synchrotron radiation, as it goes straight
on and is not bent, hits the walls of this

vacuum system, or in this case of the beam
pipe, it actually liberates molecules from

there and reintroduces them into the
vacuum, which of course then makes the

vacuum worse again. An additional problem
that gives the synchrotron radiation is,

that it also gives a significant heat
load, and if you need to dissipate all

this heat that is generated through the
very cold helium, this is not a very

efficient process. Because making this
helium so cool, is actually a very energy

intensive process. And just removing a
single watt of thermal power through the

superfluid helium costs about 1 kW of
energy. So that is not the most efficient

part. And this is why the cross-section
you have just seen includes another large

component, which also technically belongs
to the vacuum system, which is called the

beam screen. And this beam screen is
basically another tube running inside the

beam pipe, of which we have, of course,
two, which run inside the magnet cold

bores. And it shields the synchrotron
radiation heat load from the outer walls,

which are at 1.8 Kelvin, while this pipe
itself is actively cooled to only about 20

Kelvin of temperature, which is much more
efficient to dissipate this heat. So it is

basically a steel tube about one
millimeter thick. It has these pumping

holes, where hydrogen gas molecules can go
out of, and on the inside it has a copper

coating, which is used to reduce its
electrical resistance, which is required

because the beam, while it circulates,
also induces current that would otherwise

flow inside this tube, which is really, if
you think about it, only a simple tube and

it would increase the heat load again. So
a lot of engineering already has to go

into a very simple piece of ... a thing
like that. So after having spoken so much

about all the things required to just make
a beam circulate and accelerate and so on,

now it's probably also time to talk a
little bit about the beam itself and how

to control it and how to instrument, how
to measure things about this beam. Even

without going yet about collisions and
doing actual physics experiments. So the

first important bit that is able to
basically control or influence the beam

here is what's called the beam cleaning or
collimation system. So typically such a

particle beam is not very clean. It always
travels associated with what is called

halo of particles around this core area
that is less than a millimeter wide where

most of the intensity is focused. And
these particles outside we want to remove,

because they otherwise would be lost
inside the magnets and for example, would

lead to quenches of the superconducting
magnets. And for collimation, we basically

use small slits that are adjustable and
are located at two main locations of the

LHC. So they have collimation systems
there, with vertical and horizontal slits

that can be adjusted in width, in order to
scrape off all the particles that they do

want to get rid of and extract out of the
beam, while only the core part can

circulate and produce clean collisions
without any background, that otherwise

would need to be accounted for. And then
there is a whole other open topic of beam

instrumentation. So when you run a
particle accelerator, you want to measure

various quantities and performance figures
of such a beam. And that is crucial for a

correct operation and for the highest
performance, getting the highest

performance from an accelerator. And there
are a lot of different types of those, and

I want to go quickly about ... over why we
have them and what we do with them. So the

first and most basic measurement you want
to do, is the beam current measurement. So

the beam current is a basic accelerator
beam intensity measurement. So it gives

you an idea of how strong the beam that is
running inside your accelerator is. And it

is measured using these DCCTs or DC
current transformers. And their basic

principle of operation is, that while the
particles move through this torus, which

is actually a coil or a transformer,
induces a voltage there that you can

measure and then use to quantify the
intensity of this beam. And the big

challenge here is that the dynamic range,
this instrument needs to capture, is

really large, because it has to operate
from the lowest intensity pilot injection

beams up to the full energy, full number
of bunches running inside the LHC. So it

has to cover six orders of magnitude of
measurement dynamic range. Then the second

thing when talking about collisions is the
luminosity measurement. So luminosity is a

quantity basically said to measure the
rate of interaction of the particle beams.

So to give you an idea of how often
interactions happen inside the experiments

or where you want them to happen. And this
measurement is used to first of all,

adjust this interaction rate to a target
value, which is optimal for the

experiments to function and to equalize
the interaction rates in different

experiments. So different experiments also
are specified to have the same interaction

rate so they can get the same let's say
statistical quality of their data. So it's

used to equalize those. And then as a
third thing, this system is also used to

measure the crossing angle of the beam. So
as you may know, at some point, when the

beams are collided, they collide at an
angle, that is very small. And this angle

is actually measured also very precisely
in order to adjust it correctly. And it is

measured to less than a thousandth of a
degree, which is again a very impressive

feat, given that the detection principle
of this measurement is only measurement of

some neutral particles that are a result
of the particle interaction of the beam

... of the collision. Okay, so that is
number two. Then number three that we have

is the beam position monitor. Because
along the LHC, you also always want to

know, where the beam is at any given time.
So you want to measure the position of the

beam inside the beam pipe in order to
optimally adjust it to the position you

want to have it. And for that we use these
beam position monitors of which we have

more than a thousand installed along the
LHC. So they are typically capacitive

probes or electromagnetic strip lines. As
you can see on top and bottom

respectively. And they basically are
distributed along the LHC and provide

position of the particle beam along the
accelerator, which can then be used to

tune, for example, the magnets. All right.
Then we have beam profile. So after the

position, that gives you an idea where the
beam is, you also want to know its

intensity distribution. Basically when you
do a cut through the beam pipe somewhere,

you want to know how the intensity profile
looks like. And for that we have basically

two measurement systems. One measures the
profile in X and Y directions. So if you

really would do a cut and it gives you
something like this and it's, for example,

done with wire scanners, which is literal,
very thin wire that is moved through the

beam. And then the current that the beam
moving through this wire, generates is

used to generate such a profile map, when
scanning with this wire. The other one is

the longitudinal profile, which gives you
an idea about the quality of your RF

system and there you want to know how the
intensity profile of your beam looks like.

If you were looking only at one spot of
the accelerator and the beam would pass

by, and you would basically see over time
how the intensity looks like. And then as

the last bit of beam instrumentation,
there is the beam loss monitors. So they

are these yellow tubes, that are located
on the outside of mostly all the magnets,

of the dipole and quadrupole magnets and
so on. Again, there's more than a thousand

of those. And the idea here is that you
need a lot of detectors that are basically

small ionization chambers, which detect
any showers of secondary particles that

are generated when one of the high energy
protons are lost somewhere in the magnet

materials. So these are really used for
protection of the system, because if a

specific threshold of energy loss is
detected, then the accelerator needs to be

quickly shut down. Which is why they have
to react in a matter of nanoseconds in

order to keep the accelerator safe.
Because any interaction of the particle

beam with for example, the magnets could
just destroy huge amounts of money and of

time that would be needed to rebuild. And
as a last and final thing, we have spoken

one or two times already about shutting
down the LHC. Which sounds also trivial at

first, but really is not. So, the last
thing here is, what we call the Beam Dump.

So the energy content that is contained in
those particular beams, it can be used,

could be used if it were shot on a copper
target, it could just melt 1000 kilograms

or one ton of copper instantly. So during
beam extraction, the process of getting

the particle beam outside out of the LHC,
this energy needs to be dissipated

somehow. And for that, this special Beam
Dump area is constructed. So there are

fast kicker magnets, that are used to ...,
that are able to ramp up in a really,

really short amount of time of
microseconds. And then the beam is

carefully and in a controlled manner
directed into a set of concrete blocks,

that is basically big enough to dissipate
all this energy, when required. And in the

process of doing so, it also heats up to
about 800 degrees Celsius, and then of

course, also needs the associated time to
cool down again. Good. So as you may or

may not know, currently the LHC is not in
operation. So LHC currently is undergoing

its second long shutdown phase, or LS2.
But what we do when the LHC is in

operation, is that we have these status
dashboards, that you can see here, that

are distributed all around CERN, and can
be used by anyone, any passer-by, to

basically monitor what the current
operation mode or the current situation of

the accelerator is. And can be used also
to quickly see if like an operator needs

to go somewhere or is needed, or how the
shift planning for the next shift works

out and so on. And on the right side you
would see what this currently looks like.

So basically black screen saying next beam
expected in spring 2021. And the good

thing about these status pages is that you
can actually see them from your home,

because they're also openly available, as
most of the stuff we do at CERN. So if you

are interested, then perhaps in a year
from now or a bit longer than a year, it

would be quite interesting to follow all
the commissioning process of when they are

trying to start the LHC back up, and
follow that process from your home.

Otherwise, if you now feel the urge to
maybe visit CERN, pay some of the things

we talked about a visit, or are just
generally interested, CERN offers a

variety of tours free of charge. So if
you're interested in that, visit that web

site and we would be happy to welcome you
there. And with that, thank you very much

for your attention.

Applause

Severin: Punktlandung.
Herald: Thank you, Stefan and Severin. If

you have questions, there are six
microphones in the room. Please make a

queue, and we start with the Signal Angel.
Signal Angel, please, first question.

Signal Angel: There is said to be a master
red button for shutting down the whole

system in case of heavy problems. How
often did you push it yet?

Stefan: Master red button?
Severin: Master button ...

Signal Angel: Like a shut down button.
Severin: I cannot really understand you. I

think the question was about how often
basically we used the Beam Dump system to

basically get rid of the beam, is that
correct?

Signal Angel: I guess so.
Stefan: He said there is a master button.

Signal Angel: I guess so.
Stefan: I think there's a master button in

the...
Severin: There is not only one master

button, there are several master buttons.
These are switches, called beam interlock

switch. Basically, at every operator's
screen, there is basically one beam

interlock switch. I don't know. I think
sometimes they get rid of the beam just

because, I mean. When we have LHC at full
operation – Stefan talked about the

luminosity – so what is happening, that in
the beginning we have a very high amount

of luminosity, So many particles collide
on each other. But over time, like after

12 or 15 hours or whatever, basically the
luminosity ..., so the amount of particles

which collide with each other, is going
down and down. So the luminosity

decreases. And then at some point in time,
basically the operators decide, that they

will now get rid of the actual beam, which
is inside LHC and basically will recover

the whole machine and then restart the
machine again. And this is done sometimes,

I don't know, every 12 hours, sometimes
after 24 hours. Something like that, yes.

Herald: Cool. And microphone number four,
I think.

Q: Yes. So where's the energy coming from?
So do you have your own power plant, or

so?
Severin: So, no, not really, not really.

Basically, we get all the power from the
French grid. So we have relatively big

power trails coming from the French grid.
So we get 450 kV of power there. So

basically the voltage is quite high and
then we have our own transformers on site.

And I think only, ... a little bit smaller
fraction of the energy is coming from the

Swiss grid. So basically we use most of
the energy which is coming from the French

grid.
Q: Okay. Thank you.

Herald: Thank you for your question. And
microphone number one, please.

Q: Hi. Thank you for your presentation. If
I'm not wrong, you say the beam can warm a

block of concrete to 800 Celsius. Would it
be possible to use it as a weapon?

Stefan: laughs Very likely not. And CERN
very much condemns these actions in any

form, I guess. So CERN operates in a
purely peaceful mission and would never

think about using their particle beams as
a weapon. And even if they could, it is

probably not the most practical thing to
do, I guess. laughs

Herald: But if your telephone is again
hanging up, you can destroy it, right?

Stefan: laughs
Herald: And microphone number six, I

think.
Q: Yes. So you said, you can stop in

nanoseconds, but just the light would go
just 30 centimeters, you know, a

nanosecond. How will you be able to stop
in this small time?

Stefan: Ah, no, no. So what I was talking
about is that these magnets that are used

to extract the beam out of the LHC, they
have reaction times, or ramp up times that

are in the order of 1, 2, 3 microseconds.
So not nanoseconds, but microseconds. And

really only then basically the particles
still circulate, worst case one full turn,

and only then moving outside of the
accelerator.

Herald: And microphone number one again.
Q: So do you have any photos of the front

of the dump block? It has to look like
it's got hit a lot.

laughter
Severin: No, not really. I think it's one

of the only pictures we could find about,
the Beam Dump system. And these areas, I

think it's not really opened any more. So
since operation of LHC, which was in

basically LHC started in 2008, and since
then, the Beam Dump system was not opened

again because it's completely sealed in
stainless steel. And that's why it wasn't

opened anymore.
Heral: Cool. Question from the interwebs.

Signal: Regarding power supply. How do you
switch or fine-control the currents? Are

you using classic silicone transistors,
off-the-shelf IGBTs?

Sverin: Um, yes. laughs
laughter

Severin: Yes. Uh, the system was developed
at CERN. And I think that's quite common

at CERN that we basically developed all
the technology at CERN or try to develop

nearly everything at CERN. But then
production, for example, is put into

industry. And yes, these are relatively
classical power converters. The

interesting or like challenging part about
the current power converters is really

that the current has to be measured quite
precisely and also controlled quite

precisely so there we use also DCC TS.
Which we have also mentioned before. But

basically all this controlled mechanism
there. That's one of the big challenges

there.
Herald: Cool. Microphone number one again.

Q: You talked about the orbit clock that
detects when the bunch is completed one

round. How is it possible to detect which
is the first bunch?

Stefan: Yeah. So it is it is actually not
detected, but this clock is actually

something that is constructed. So we
basically what we do is, we count these

cycles of the of the RF cycle. Maybe I can
open this slide. So somewhere there is a

counter that basically knows how many 40
MHz clock cycles a full rotation takes.

And then at some point decides this is
number one. And that's also where they

start counting when they inject bunches
into the LHC. So there's no marker, let's

say. But there is a certain structure to
the beam. So you could potentially do

that. So, for example, for these longer
periods where the kicker magnets need to

ramp up, they have something they call the
abort gap. So a number of bunches that are

never filled but are always kept empty. So
the magnets have enough time to deflect

the beam when the next bunch comes around.
So you could probably measure that, but

it's much easier to do it the other way
around.

Herald: Microphone number four, please.
Q: You said you had quite tight needs for

the timing clock. Is it tight enough? That
the speed of light was the limit with the

distances between locations or that was
not a concern?

Stefan: No, it is a concern. So because
just distributing a cable for 27

kilometers produces like just considerable
run times of electrical signals. All the

delays of all the cables need to be
measured precisely for their delay and

then calibrated out so all the experiments
get their clocks at the right time,

shifted, compensated for the delay time
that it just takes to get the signal

there.
Herald: And again, the interwebs.

Signal Angel: Is it too dangerous to stand
near the concrete cooling blocks, like

radioactive wise or, I don't know.
Severin: Yes.

laughter
Stefan: Not recommended.

Severin: Not recommended. We have a very
good interlock system. Also, the doors,

all the doors have switches. So basically
when one door is basically like opened

then basically the whole machine will be
shut down. So we have very critical and

safety related access system at LHC. Maybe
you watch Angels and Demons. This

Hollywood movie that we have, the eye
scanners are shown. It's a little bit. I

mean, it's Hollywood. But, we have eye
scanners. So iris scanners. So every time

like we want to go to the tunnel, for
example, then we have to let also our iris

be scanned because otherwise we will not
be able to go to the tunnel. So there's a

very sophisticated access system to really
go to the tunnel. So when there is

operation, the whole tunnel access is
completely blocked.

Herald: Good, microphone number one,
please.

Q: What is the exact reason to have each
of the experiments, every side. I mean, so

far apart on the LHC. I mean, on
opposite sides.

Severin: Um, basically, you are talking
about Atlas and CMS. The reason for that

is because when, these two experiments
were constructed, there was a little bit

of fear that particles basically interact
at the two experiments. So that they

really are like the most far away. We like
to have a very big distance from each

other. So there is no interaction between
them. That's why we basically put them at

point one and point five. That's the
reason why.

Herald: If I can see it correctly.
Microphone number five.

Q: Yes, hello. I've seen that you're also
using the CAN bus. What are you using the

CAN bus for in CERN?
Stefan: I know of at least one use, but it

is inside an experiment. So there are, as
far as I know, investigations under way to

use the CAN bus to do the actual control
of the detectors of one experiment. I

don't know if there is a use inside the
accelerator itself. So apart from the

experiments. But perhaps if you come by
afterwards we can find one.

Q: Thank you.
Herald: Microphone number one.

Q: Do you have any official data about how
many tons of duct taper used in

daily operations?
laughter

Severin: No. No.
Herald: What about zip ties?

Severin: Many. Yeah. Millions. Billions.
Herald: Okay. As far as I can see... Ah,

the intercepts again with a question.
Signal Angel: Do you know your monthly

power bill?
laughter

Severin: No, not no. No, sorry.
Stefan: No. But it is, I think, in fact

the contribution of France, which is the
main contributor in terms of energy. That

it is part of their contribution to
contribute the electricity bill basically

instead of paying money to CERN. That's as
far as I know.

Severin: Yes. And also we shut down the
LHC and the accelerator complex through

like the wintertime. And one of the
reasons for that is because electricity is

more expensive during wintertime in France
than in summer.

Herald: In this case, I can't see any
other questions. I have a maybe stupid

question. You said earlier you have to
focus and defocus the beam. But as we

know, you accelerated already the
particles. Why do we have to focus the

beam?
Severin: Because every time when we have a

dipole magnet, then basically we bend the
particle around an arc. But then we also

defocus a little bit. And also, the
coulomb force is still a problem because

we have equally charged particles in the
bunch or in the whole beam itself. So they

will by themselves basically go out of
each other. And if you would not focus it

again, then basically we would lose the
whole beam in the end.

Herald: Oh, thank you. I don't see any
questions. Internet? In this case thank

you very, very much, Stefan and Severin.
Please. With a warm applause. The Large

Hadron infrastructure talk.

Applause

subtitles created by c3subtitles.de
in the year 20??. Join, and help us!