New luminosity record for LHC + Injector Chain

I know this is getting a bit repetitive but the Large Hadron Collider has a new luminosity record of 1270/μb/s. This was achieved this morning using 1092 bunches.

Notice that this is a change from the plan previously reported. The new plan is to run with 1092 bunches at least once with 108 bunch trains. This is the maximum number of bunches that can be fitted in with 108 bunch injection. They will then switch to 144 bunch injection with the same number of bunches in total. It will then be possible to fit in another train of 144 bunches to make 1236 in total, then the final train goes in to reach the limit of 1380 bunches. All this will take two or three weeks depending on how smoothly it goes. The final peak luminosity will be around 1600/μb/s but there is potential to increase it a little further by pushing up bunch intensity.

Update 29-May-2011: More records! The current fill finally collected over 37/pb which is a record for one fill. CMS have over 500/pb delivered in 2011 and if you count the data from last year as well they will soon pass 500/pb recorded too. The data collected in the last seven days is 169/pb despite the days lost to cryogenic problems. If they keep the same rate for the 130 physics days they have left this year they would add another 3000/pb = 3/fb, but we expect them to do much better.

Update 30-May-2011: With the second run at 1092 bunches well underway, a record for total luminosity over 24 hours has been added to the list. 69.2/pb has been delivered in one day. If this could be maintained for 130 days they would collect 9000/pb = 9/fb to add to the 0.5/pb so far collected bringing the total for 2011 very close to the 10/fb I have predicted. Of course they can’t run this efficiently every day, on the other hand the peak luminosity will rise by another 30% in the next few weeks and possibly further later on if they decide to increase the bunch intensity.

The Injector Chain

Unless you have been following the developments at the LHC in detail,  all this talk of bunches and luminosity probably sounds a bit confusing. Since its a lazy Sunday afternoon I’ll dig around and put together some info about how the process of injecting proton bunches into the collider works.


The protons that collide inside the experiments of the LHC begin their life as ordinary hydrogen gas inside a red bottle. From there they are piped into a duoplasmatron. Hydrogen molecules are made of two hydrogen atoms each of which contains one proton and one electron. The duoplasmatron uses some strong electric fields to tear the molecules and the atoms apart so that the protons move freely in a not plasma. More electric fields then draw the protons out from the duoplasmatron in pulses and accelerate them into a beam. They then shoot through a radio frequency quadrapole that focuses and accelerates the protons towards the next stage in their journey, the LINAC


LINAC 2 is a 30 meter long linear accelerator that has been in use since 1978. Pulses of protons from the duoplasmatron are accelerated in a straight line using electric fields. The fields alternate from positive to negative in a wave that travels up the LINAC at just the right speed so that the protons are always being forced forwards, riding the wave like a surfer on a breaker that somehow gets faster and faster as it approaches the beach. Finally they shoot out the other end with an energy of 50 MeV. The proton mass is equivalent to about 20 times this energy so they are only travelling at about half the speed of light. The pulses of protons can be sent down the LINAC at a rate of about one every second. It is these pulses that will eventually form the bunches of protons inside the LHC ring and already they need to be well collimated so that they remain in a tight bunch all the way.

The PS Booster

From the LINAC the protons pass into the proton synchrotron booster (PSB), a circular accelerator with a diameter of 50 meters that dates back to 1972. The PSB is used to accelerate the protons by an energy factor of 280 up to 1.4 GeV. It is actually made of four separate rings stacked one on top of another each of which takes in a bunch of protons from LINAC 2 to accelerate.

The Proton Synchrotron

The four bunches of protons in the PS Booster, followed by another injection of two bunches are fed into the Proton Synchrotron (PS), another circular accelerator 4 times as big as the PSB.  The PS has been in use since 1959 making it by the oldest part of the injector chain. When it was first built it was itself the most powerful accelerator in the world and it remained CERNs principle collider until 1970. It has of course been upgraded since then to serve as the feeder to later generations of accelerator. The PS has a diameter of 200 meters and uses 277 ordinary electromagnets to bend the protons around its circular ring. Between the magnets the protons are accelerated using the same principles as the linear accelerator.

In its role as a feeder to the LHC, the PS now has a new party trick. With 6 bunches inside to start with it can split the bunches to make more. Initially each bunch circulates with a separation of 300 ns and a bigger gap at one point to allow them to be kicked out. The LHC may run with bunch spacing of 150ns, 75ns, 50ns or 25ns. To form these spacing the PS can split each bunch into either three smaller bunches or two bunches. To get this years spacing of 50ns it must be split once into three, each of which is then split again into another two. The PS then has 36 bunches circulating in a nice train with a much bigger gap at the ends. Next year the bunches may be split one further time to produce trains of 72 bunches with 25ns spacing. these bunches can be accelerated by a further energy factor of 18 to provide a proton energy of 26 GeV before they are fed into the next stage, the SPS.

The Super Proton Synchrotron

In 1976 CERN commissioned a new accelerator ring that dwarfed the PS. The Super Proton Synchrotron has a diameter of 2.2km and can now accelerate protons fed from the PS by a factor of 17 up to 450 GeV using 960 superconducting magnets to bend the proton bunches around the ring . The SPS never held the title of the worlds highest energy accelerator because it was beaten to that title by the Tevatron at Fermilab. However, it initially had some design advantages that enabled it to find the W and Z bosons first.

Now the SPS is the final stage that injects the protons into the Large Hadron Collider. Since it is more than ten times as big as the PS it can easily build up much larger trains of protons by collecting together the trains of bunches fed from the PS. For the fill running now it takes three trains of 36 bunches from the PS one at a time to form trains of 108 bunches to feed into the LHC.

The Large Hadron Collider

The main LHC accelerator is an underground circular tunnel over 8km in diameter, nearly four times as big as the SPS. It has 5000 superconducting magnets and accelerates the protons up by the last factor of over 7 to 3.5 TeV. In 2014 this will be increased to 7 TeV. It actually consists of two rings that cross over at intersection points inside large particle detectors where the protons finally collide. Maximizing the rate of collision means packing in as many bunches as possible into these rings

The trains from the SPS can be directed along tunnels into either ring of the LHC to circulate in opposite directions. Once the larger trains are in, smaller trains can be injected to squeeze into any remaining available gaps but there will always be some bigger spaces left that cannot be filled. This is because the process of moving the trains from the PS to the SPS and then to the LHC requires powerful electromagnets to be fired up that kick the protons out of one ring into another. It is as if the trains are being transferred from one railtrack to another by moving points. Because the points take time to switch position there must be gaps between the trains to allow time for them to move, otherwise the trains will be derailed when they run into the moving points! There are gaps of different sizes because as the trains get faster at each stage, heavier points are needed (meaning more powerful magnets) which take longer to switch.

While the bunches themselves are separated by gaps of 50 ns, the trains of 36 bunches that come from the PS are separated by 225 ns. The trains of 108 bunches must then be separated by bigger 975 ns gaps because bigger magnets are needed to kick the trains at higher energy into the LHC. Finally, an even bigger abort gap of 3000 ns must be left as they circulate in the LHC so that the protons can be safely dumped out of the LHC at its full energy of 7 TeV. To pack the maximum number of bunches into the LHC the number of larger gaps must be kept down. That is why the number of bunches injected in each train must be increased. Trains of 144 bunches will be needed to get the maximum of 1380 bunches into the LHC in the next few weeks.

With more bunches there are more proton collisions which means more data for discovering new physics.

By the way, if you want to watch the injection process in detail the page to look at is SPS-page 1.  The yellow line in the graph shows the build up of protons in the PS in the steps needed to build up the three or four trains that go into the SPS. The white line shows the rise of the magnets in the PS and finally the SPS as the protons are accelerated. A moving vertical white line indicates the progress through this build-up. When it reaches the end after about 40 seconds,  the trains are then finally ready to be injected into the LHC. Use the LHC announcer page to follow these injections and the subsequent ramping up of energy inside the LHC.

12 Responses to New luminosity record for LHC + Injector Chain

  1. carla says:

    I come to this blog for the LHC updates, so keep them coming!

    The beam size instabilites are slight this time compared to going to 912b, so hopefully things will run smoother this week with this intensity and the smashing of previous integrated luminosity records.

  2. algernon says:

    Hi Philip,

    thanks for the update on the roadmap. Apparently you’re the only source of information that routinely discusses “what’s next”, and I always come back for that.

    A question: I see you’re reporting 540/pb of integrated lumi in the blog’s header. Where do you get this figure?

    I mean, right now ATLAS claims 484/pb on their site (total updated just a few hours ago). Fill #1815 is now undergoing and recording ~25/pb as I’m writing, so that’ll be 509/pb total.
    Where does that 540 come from?

    • Philip Gibbs says:

      I have been adding up the luminosities for each fill as I go along and cross checking them with the official data at I noticed they have missed out some fills in their data such as fill 1745 which collected 26/pb. That is why I have bigger numbers. There may be reasons why they left out those fills but I prefer to show the full data delivered. The amount recorded as usable data will a less anyway.

      • algernon says:

        Holy cow, I didn’t even know that particular statistics subsite existed. Now I’ll end up wasting countless hours checking all of that stuff 🙂


  3. felix says:

    With fill 1815 and 1816 they achieved 68/pb within a duration of 24 hours.

    • Philip Gibbs says:

      I made it 69.2/pb and have added an update to the post. This is a good result because I am hoping they will be able to log 70/pb average per day for the scheduled physics days. They need to stop those power converters from spoiling the day though!

  4. carla says:

    I don’t think you’ve taken into account the days scheduled for machine development, technical stops and the extra week needed to get the machine to where it was. I make it 100 days left after 7th June. With a conservative 30/pb per day this gives an integrated 3/fb. Maybe 5/fb if they’re lucky which would be seen as a magnificent achievement given 1/fb was the target at the start of the year.

  5. Bill K says:

    From today’s Latest News: “There is a chance to reach 1/fb end of this week or beginning next week.”

    • Philip Gibbs says:

      A few good runs and they will be there

      • carla says:

        They still have that weird problem with emittance blow up when they try to fill after a long run. They were doing an hour monitoring of the rf cavity so maybe things will be better for the next fill.

  6. […] rings to put together the bunches in trains (see my earlier post about the injection chain). The early stages have lower energy so the bunches can be slotted closer, but the rings are […]

  7. […] título de la entrada). Recomiendo leerla a los interesados y también su entrada anterior “New luminosity record for LHC + Injector Chain,” viXra log, May 29, […]

%d bloggers like this: