Page 3 of 4

Re: 2018 Operation

Posted: Wed May 09, 2018 3:05 pm
by dukwon
The water leak at point 8 last week was a bit more spectacular than I thought:
Fuite_SF8.png

Re: 2018 Operation

Posted: Thu May 10, 2018 5:20 am
by andrewp
Thanks so much for the exquisite detail. As a rank amateur I noticed that the peak initial lumi had been dialled back after 3 abortive starts at 2556b. Seems like starting with >20,000 lumi is a no-no, even though achievable.

Re: 2018 Operation

Posted: Sat May 12, 2018 10:09 pm
by andrewp
Looks like the current run (#6675) will smash all previous LHC records for sustainable integrated lumi per 24 hour period - at least 1.0 1/fb. :dance:

Notionally this consists of 2 switchovers (time of dump to time of next stable beams) and 2 runs. The switchovers can be 2 hours each and the runs can be 10 hours each. Each run can deliver 0.5 1/fb.

Re: 2018 Operation

Posted: Mon May 14, 2018 9:05 am
by dukwon
Indeed, machine availability was >95% over the weekend, with 70% in stable beams.

They are starting to gradually increase the bunch intensity. Fills 6675 and 6677 peaked at over 20 Hz/nb.

The turnaround from 6674 to 6675 was 1:59:47; a new record!

At the average rate over the last 2 weeks, we might well have 45% of the conservative lumi goal (see start of thread) by MD1, which is about 1/3 of the way into the year in terms of regular proton physics running. I wonder if we can hit 90/fb (3/fb for LHCb) this year.

Re: 2018 Operation

Posted: Mon May 14, 2018 7:06 pm
by andrewp
Yes, the machine is purring along on song now. Clearly there's an optimal run time for average delivered lumi per time period, which depends on the switchover duration and on the decay exponent coefficient. I suspect that, with 2 hour switchovers, 13+ hours is somewhat too long.

Re: 2018 Operation

Posted: Mon May 14, 2018 10:04 pm
by RocketManKSC
Yes I have posted on this subject a few time. The factor that some people miss is that it also depend on how long you can expect to stay in stable beam if you restart. The simple formula that models Lumi vs time, Stable Bean time, and turn around time is not the best you can becauve of early un-planned Beam Dumps. This being true running a little longer will improve performance in the long run .

Re: 2018 Operation

Posted: Tue May 15, 2018 1:08 am
by jmc2000
I think the aim is to achieve 150\fb in total for 2015-2018, working out at approx 60\fb for 2018, and making Run 2 a great success even if nothing else has been discovered since the Higgs.

Currently averaging 4\fb/week, I can see this being achieved by the start of MD3; so perhaps the remaining time will be given over to MD since little will be gained for the additional 10-20\fb.

Re: 2018 Operation

Posted: Tue May 15, 2018 7:50 am
by dukwon
even if nothing else has been discovered since the Higgs
Discovering fundamental particles shouldn't be a main criterion for success. The LHC experiments have done a lot of physics in a variety of areas. Last time I looked they have published about 2,200 papers between them.
so perhaps the remaining time will be given over to MD since little will be gained for the additional 10-20\fb.
I think that would upset quite a lot of physicists...

Re: 2018 Operation

Posted: Tue May 15, 2018 10:11 am
by andrewp
What's the highest energy particle that can be discovered currently? (assuming such a beast exists). Is it in fact 13 TeV?

Re: 2018 Operation

Posted: Tue May 15, 2018 6:52 pm
by dukwon
16L2 has returned with vengeance over the last 24 hours:
  • 21:52 Programmed dump in stable beams to end fill 6681
  • 22:01 to 00:09 SPS cavity trip
  • 02:22 Back to stable beams
  • 02:57 Dump (stable beams) 16L2
  • 03:29 to 05:19 MKI2 fault
  • 07:03 Dump (flat top) no reason given in logbook
  • 10:46 Dump (ramp) 16L2
  • 13:14 Dump (ramp) 16L2
  • 16:23 Dump (injection) 16L2
  • 19:13 Stable beams with 987b

Re: 2018 Operation

Posted: Tue May 15, 2018 9:22 pm
by andrewp
I thought they'd fixed 16L2 by removing a bunch of frozen air during the shutdown. What's the theory about the nature of the 16L2 UFO?

Re: 2018 Operation

Posted: Wed May 16, 2018 5:34 am
by dukwon
It was seen during scrubbing that the intervention to remove frozen air did not eliminate the problem. The best tactic was found to be adding bumps and restricting the current through the appropriate magnet.

The working theory is still that is frozen air or similar. If there is water ice in there it won't have been affected by the warm-up and extraction

Re: 2018 Operation

Posted: Wed May 16, 2018 7:37 pm
by tomey36
Does anyone know why they are redoing the intensity ramp up?

Re: 2018 Operation

Posted: Wed May 16, 2018 8:16 pm
by dukwon
Look at the last few posts. 16L2 "woke up" and it took reduced beam intensity to get back to reliable stable beams. It was all going so well, too

Re: 2018 Operation

Posted: Thu May 17, 2018 1:28 am
by neocortex
Yes, there's been quite a few dumps from 16L2 lately. I'm not worried about it though. The operators have made some changes to the cycle since those dumps with 2556 bunches.

First thing is they've improved the beam injection through transfer-line steering, and perhaps small changes were made to the tune feedback; the tune shift (a good measure of the instability of the beam) of beam 1 is now in line with that of beam 2, and so both beams suffer just about the same amount of loss throughout the cycle. Beam 1 is still slightly worse than beam 2, and if you take a look at the bunch trains after a while in stable beams, you can clearly see the electron cloud doing its dirty work.

Whatever the changes were, they've significantly improved the health of beam 1, and also mitigated 16L2. That's probably not a coincidence, and we should be back to smooth running with 2556 bunches within the next 24 hours.