Bunch trains have arrived

Big news on Thursday 23 of the first physics fill with bunch trains...
ATLAS is now taking data with 47 colliding bunches at IP1, organised in 6 trains with 8 bunches/train (the 3rd train has only 7 bunches colliding at IP1, the 8th has no pair in B2). The peak luminosity is 2.1 10^-31 s-1cm-2 (new record), ie, ~75 nb-1 per hour (decaying of course).


Update on Friday morning...
Since a few people asked: the luminosity per bunch is higher in this last run (165632, fill 1366) than that of Wed (165591, fill 1364), mostly (not entirely) because the transverse beam emittance was narrower (which goes along with somewhat "longer" bunches in z - 61mm instead of 57mm at stable-beam start). The beam intensities per bunch were similar (~10^11 protons) in both runs. The transverse beam spot size is ~35um in 165632, while it was ~40um in run 165591 (at the beginning of the collisions, both increasing by ~5um during the fill), yielding a factor of (40/35)^2=1.3 in the per-bunch luminosity.


And from Rolf Heuer on Friday... "A game-changing fill for the LHC":


A long period of machine development paid dividends last night with a game-changing fill in the LHC. As I write this, the fill, which started colliding at 19:00 yesterday evening, has just wound down. Both ATLAS and CMS have posted integrated luminosities of over 680 inverse nanobarns, and the initial luminosity for the fill doubles the previous record at 2x10^31cm-2s-1.


But it’s not the records that are important this time – it’s normal that in the start-up phase of a new machine, records will fall like autumn leaves – what’s significant here is that the LHC’s performance this fill significantly exceeded some crucial design parameters, opening up the path to much better still to come.


Last night’s fill was the first with 56 bunches arranged in trains of eight bunches per train. The significance of bunch train running is that we can configure the orbits such that more bunches collide in the experiments, so even though the number of bunches may not be much higher, the collision rate is. For example, last night’s 56-bunch fill had 47 bunches colliding at ATLAS, CMS and LHCb, with 16 colliding in ALICE, whose needs are lower. This compares to a maximum of 36 colliding bunches out of 48 total before we introduced bunch trains.


A big jump in luminosity was clearly expected in moving to bunch trains and colliding more bunches. What came as a pleasant surprise is that it was accompanied by an exceptional beam lifetime of 40 hours, and less disruption to the beams caused by packing more protons into a smaller space (in technical terms, the beam-beam tune shift was much less destructive to the beams than anticipated). This result means that the LHC operators have more leeway in operational parameters in the quest for higher luminosity.


The plan for today and the weekend is to run for one more fill with 56 bunches in trains of eight before moving on to 104 bunches in 13 trains of eight, with 93 bunches colliding in ATLAS and CMS. Ultimately, the LHC will run with 2808 bunches in each beam, so there’s still a long way to go. We’ll get there slowly but surely by adding bunches to each train until the trains meet in a single machine-filling train. That will take time, but for the moment, last night’s fill puts us well on the way to achieving the main objective for 2010: a luminosity of 10^32cm-2s-1.