Before changing the prescale column, remember to talk to the Shift Leader.

Prescale Column Choice instructions: PLEASE REFRESH AND READ AT START OF SHIFT

Collisions (1900 Bunches)

Preamble: If you (after consulting with the shift leader) find these instructions are unclear, ambiguous, or contradictory, please ring the HLT doc to clarify. The HLT doc writes them and is responsible for ensuring that these instructions are correct and clear.

The Level-1 rate before deadtime ("FINOR" rate) should never exceed 100 kHz. Before switching into another column, check in the "Rate Preview" panel that its FINOR rate will not be too high. During Lumi Levelling, rates may vary considerably, so check the Rate Preview over a few lumisections. By default, the Rate Preview panel shows the next column (n+1) compared to the one you are running in (n). You can select another Preview column but remember that this will affect what everybody is seeing (this switching occurs in the hardware).

When in flat top, squeeze, and adjust, stay in column 0, which will give a ~few Hz each of stream physics, express and DQM. When in stable beams, move to the appropriate column based on the table below. Currently we expect this column to be 2. Note, at the start of the fill the luminosity can be higher than expected for a lumi section or two as the LHC finalises the leveling. So if the L1 rates are very high, dont panic, wait two lumi sections and see if the LHC has brought us down to the expected lumi of ~1.5E34. If after two lumi sections (~1min) the rate is still high, take appropriate action as below in "Chosing a column while leveling". If you have concerns over which column to use, please ring the HLT doc who will clarify.

During emittance scan only: it is not necessary to change columns to try to match the luminosity. Stay in the same column (appropriate for the lumi at the beginning of the scan) until the end of the scan.

Express is expected to be 60-100Hz during collisions.

How to check if a prescale is really the right one? (after switching based on lumi)

  • L1 rates of 95KHz are fine, in fact we do not care about the L1 rate, we care about deadtime. High L1 rates can cause high deadtime though, thats why we are so concerned about L1 rate.
  • What is high deadtime? Rule of thumb, ~<5% normal, ~>5% above normal, >10% problem.
    • If you see high deadtime and the pre-deadtime L1 rate <90KHz, there is an unrelated problem and the shift crew should understand its origin (and hopefully get it fixed).
  • As the fill progresses, please resist the urge to change to a higher column because you think the L1 and/or HLT rate is too low. The luminosity-based switch points of the prescale columns have been designed to target a certain average HLT output rate over the course of the fill, so that high rates at the beginning of the fill are balanced out by low rates at the end of the fill. Please stick to the table below unless there is a problem.
  • We expect 0.7 to 1.3KHz for the HLT. We aim to average roughly 1KHz over a fill.
  • If the rate has significantly jumped, check if its all due to one trigger or not. There should be no trigger with rate >300Hz which starts with "HLT_" (so not DST_ or Alca_)

Important Note: If the deadtime spikes to >50% for multiple (2+) lumisections this should be investigated immediately. First check if any individual L1 triggers have a pre-deadtime rate >50 kHz, next determine if these high rate triggers are associated with a specific sub-system (e.g. L1_ETM* / L1_SingleEG* --> ECAL, L1_SingleJet* --> HCAL). If the high rates are clearly related to a specific sub-system you should immediately contact the corresponding sub-system expert and notify them of the high deadtime as well as which triggers have increased rates. [In case of ECAL subsystem (> high EG rates), if the DOC is not reachable, please RED RECYCLE immediately ECAL] If there are high L1 / HLT rates and it is still unclear what is causing the high deadtime, please call either the L1 or HLT Docs.

Note: when we are running in columns 11 - 20, an additional 1kHz of rate will be added to the normal physics rate. So a streamPhysics rate of ~2kHz (0.7kHz-1.3kHz goes to 1.7kHz-2.3kHz)

Choosing a Column While Leveling:
We expect that you will be able to start in column 2, the 1.5E34 at the lumi leveling period and can select the columns as normal. If there is high deadtime (>5% and L1 pre-deadtime rate < 95kHz) its probably not trigger but you should identify and have the source fixed. If there is high deadtime (>5%) and L1 pre-time 95kHz - 120kHz, move to column 1 and assess why there is increased L1 rates (hot calo cell, muon issues etc), moving back to 2 once it is fixed. If you can not see a reason why there is increased L1 rates, please ring the HLT doc. If there is high deadtime (>5%) and > 120kHz L1 rate, its probably a hot tower or similar hardware issue, identify it, if you are unclear what is causing it, ring the L1 doc.

Currently there is another dimension to choosing the column while leveling. The MET PD is very large due to ECAL laser corrections introduced in Fill6012. This is slowly returning to normal as we continue to take data. See this plot : http://cern.ch/go/gKk7 which is a MET trigger rate for the first runs of Fills 6011,6012,6013 and 6014, it shows the rate decreasing with each fill. If the MET PD is greater > ~350 Hz, this presents a problem. As such after switching prescale columns please make sure that the MET PD rate remains < 350Hz. If the MET PD rate is above 350Hz, please ring the HLT doc immediately (day or night). You can see check the MET PD date from WBM with this link: https://cmswbm.cern.ch/cmsdb/servlet/DatasetSummary?RUN=305341&NAME=MET&OPTION=EVENTRATE, replace 305341 in the url with the current runnr.

In general the shift leader should verify with the L1 shifter that the preview pre-dead time trigger rate for the target column is below 95 kHz when switching columns.

Column number Column name When to use Min Lumi Max Lumi Notes
0 Emergency FLAT TOP*, SQUEEZE and *ADJUST     All data in this column is bad for physics Call the RFM if you feel the need to go to this column during stable beams
           
1 1.6e34 STABLE BEAMS 1.55e34 - 1.6e34 physics column.
2 1.5e34 STABLE BEAMS 1.45e34 1.55e34 1.5e34 physics column. we expect fills will start in this column
3 1.4e34 STABLE BEAMS 1.35e34 1.45e34 1.4e34 physics column.
4 1.3e34 STABLE BEAMS 1.25e34 1.35e34 1.35e34 physics column
5 1.2e34 STABLE BEAMS 1.15e34 1.25e34 1.2e34 physics column
6 1.1e34 STABLE BEAMS 1.05e34 1.15e34 1.1e34 physics column
7 1.0e34 STABLE BEAMS 0.8e34 1.05e34 1.0e34 physics column
8 7.5e33 STABLE BEAMS 0.6e34 0.8e34 7.5e33 physics column
9 5.0e33 STABLE BEAMS 0.4e34 0.6e34 5.0e33 physics column
10 2.5e33 STABLE BEAMS - 0.4e34 2.5e33 physics column
           
11 1.6e34 STABLE BEAMS 1.55e34   Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
12 1.5e34 STABLE BEAMS 1.45e34 1.55e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
13 1.4e34 STABLE BEAMS 1.35e34 1.45e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
14 1.3e34 STABLE BEAMS 1.25e34 1.35e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
15 1.2e34 STABLE BEAMS 1.15e34 1.25e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
16 1.1e34 STABLE BEAMS 1.05e34 1.15e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
17 1.0e34 STABLE BEAMS 0.8e34 1.05e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
18 7.5e33 STABLE BEAMS 0.6e34 0.8e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
19 5.0e33 STABLE BEAMS 0.4e34 0.6e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required
20 2.5e33 STABLE BEAMS - 0.4e34 Adds 0.5kHz HLT_Physics + 0.5kHz ZeroBias, explicit HLT DOC/RFM permission required



Cosmics

The cosmics mode has prescale columns to fit your needs in case you need to do a test (e.g. with large random rate), or a problem crops up (e.g. noisy triggers). The rate of the Express stream should not exceed ~100-120 Hz during cosmics. Note the only triggers in express for cosmics are HLT_L1SingleMuOpen, HLT_L1SingleMuOpen_DT, HLT_L1SingleMuCosmics, HLT_Random. So any high express rate has to be either due to muons or randoms. In the case of high express rate, the first thing you should do is ask the DAQ shifter what rate of random triggers they have selected. If it is >600Hz, this is likely the reason for your high express rate and you should immediately switch column 1 or 2 depending on the random rate.

we have a known issue with the L1_SingleEG5, it is currently 2kHz often during cosmics due to some noise. Given this is no issue, it will not be fixed in the near future. So if in cosmics/circulating the L1_SingleEG5 rate is 2kHz, no need to call the ECAL Doc

The following information describes the use for each column:

column 0: Nominal Running

column 1: High Random Rate 100KHz (Randoms prescaled by 10000)

column 2: High Random Rate 20KHz (Randoms prescaled by 2000)

column 3: High ECAL triggers (L1_EG triggers prescaled by 100)

column 4: High HCAL triggers (L1_Jet triggers prescaled by 100)

column 5: High BPTX rate (BptxOR /AND prescaled by 10000)

Circulating

The circulating menu is essentially the same as the Cosmics menu (with the addition of the BptxOR path). The prescale columns are the same as described above for Cosmics. For nominal running use column 0.

XeXe Test

The prescale columns are the same as described above for Cosmics. For nominal running use column 0. Please note that the rate of HLT_Random is expected to be high in this menu, which is a few hundred Hz.

High Rate Test

The high rate test key will not write output to deliver to Tier 0, however as an extra safety mechanism it's recommended to ask DAQ shifter to turn off transfer to Tier0. Even though there are no events written, the HLT will still produce files to monitor rates, hence you will get information on HLT rates in the control room. This also means that our rate monitoring script will notice the high rate and start to complain. Currently this warning can be ignored only in the high rates mode.

1300 Bunch Menu (Ramp-up after TS2)

Column numberSorted ascending Column name When to use Notes
0 Emergency At flat top and squeeze or Lumi>2.1e34 To be used in case of emergency Call the HLT DOC if you feel the need to go to this column during stable beams
1 2.0E34 Lumi >1.9e34, Lumi<2.1e34 2.0e34 physics column Call the HLT DOC if you feel the need to go to this column during stable beams
2 1.8e34 Lumi >1.7e34, Lumi<1.9e34 1.8e34 physics column
3 1.6e34 Lumi >1.5e34, Lumi<1.7e34 1.6e34 physics column
4 1.4e34 Lumi >1.25e34, Lumi<1.5e34 1.4e34 physics column
5 1.2e34 Lumi >1.1e34, Lumi<1.25e34 1.2e34 physics column
6 1.0e34 Lumi >0.8e34, Lumi<1.1e34 1.0e34 physics column
7 7.5e33 Lumi >0.6e34, Lumi<0.8e34 7.5e33 physics column
8 5.0e33 Lumi >0.4e34, Lumi<0.6e34 5.0e33 physics column
9 2.5e33 Lumi <0.4e34 2.5e33 physics column
10 1.8e34 +ZB Lumi >1.7e34, Lumi<1.9e34 Adds 1kHz ZeroBias
11 1.6e34 +ZB Lumi >1.5e34, Lumi<1.7e34 Adds 1kHz ZeroBias
12 1.4e34 +ZB Lumi >1.25e34, Lumi<1.5e34 Adds 1kHz ZeroBias
13 1.2e34 +ZB Lumi >1.1e34, Lumi<1.25e34 Adds 1kHz ZeroBias
14 1.0e34 +ZB Lumi >0.8e34, Lumi<1.1e34 Adds 1kHz ZeroBias
15 7.5e33 +ZB Lumi >0.6e34, Lumi<0.8e34 Adds 1kHz ZeroBias
16 5.0e33 +ZB Lumi >0.4e34, Lumi<0.6e34 Adds 1kHz ZeroBias
17 2.5e33 +ZB Lumi <0.4e34 Adds 1kHz ZeroBias

600 Bunch Fill (Ramp-up after TS2)

The plan is to first have ~1 hr of head-on collisions during the beginning of this fill, during which time you should be using the normal collisions run mode (collisions2017). The prescale columns for this are the following:

Column number Column name When to use Notes
0 Emergency At flat top and squeeze or Lumi>2.1e34 To be used in case of emergency Call the HLT DOC if you feel the need to go to this column during stable beams
1 2.0E34 Lumi >1.9e34, Lumi<2.1e34 2.0e34 physics column Call the HLT DOC if you feel the need to go to this column during stable beams
2 1.8e34 Lumi >1.7e34, Lumi<1.9e34 1.8e34 physics column
3 1.6e34 Lumi >1.5e34, Lumi<1.7e34 1.6e34 physics column
4 1.4e34 Lumi >1.25e34, Lumi<1.5e34 1.4e34 physics column
5 1.2e34 Lumi >1.1e34, Lumi<1.25e34 1.2e34 physics column
6 1.0e34 Lumi >0.8e34, Lumi<1.1e34 1.0e34 physics column
7 7.5e33 Lumi >0.6e34, Lumi<0.8e34 7.5e33 physics column
8 5.0e33 Lumi >0.4e34, Lumi<0.6e34 5.0e33 physics column
9 2.5e33 Lumi <0.4e34 2.5e33 physics column
10 1.8e34 +ZB Lumi >1.7e34, Lumi<1.9e34 Adds 3kHz ZeroBias
11 1.6e34 +ZB Lumi >1.5e34, Lumi<1.7e34 Adds 3kHz ZeroBias
12 1.4e34 +ZB Lumi >1.25e34, Lumi<1.5e34 Adds 3kHz ZeroBias
13 1.2e34 +ZB Lumi >1.1e34, Lumi<1.25e34 Adds 3kHz ZeroBias
14 1.0e34 +ZB Lumi >0.8e34, Lumi<1.1e34 Adds 3kHz ZeroBias
15 7.5e33 +ZB Lumi >0.6e34, Lumi<0.8e34 Adds 3kHz ZeroBias
16 5.0e33 +ZB Lumi >0.4e34, Lumi<0.6e34 Adds 3kHz ZeroBias
17 2.5e33 +ZB Lumi <0.4e34 Adds 3kHz ZeroBias

Then, during the low-PU part of the same fill (after the beams have been separated), please switch to the "lowPU2017" run mode. The inst lumi for this part of the fill is supposed to be kept at/around 6e31, but may fluctuate. Please start in the 6e31 column, and move to a different one if the lumi is significantly different from 6e31. The prescale columns are:

(Use column 4 when there's 20kHz random trigger input)

Column number Column name When to use Notes
0 Emergency The rate is too high To be used in case of emergency
1 11e31 Lumi > 9.5e34 11e31 physics column
2 8.5e31 Lumi > 7e31, Lumi < 9.5e31 8.5e31 physics column
3 6e31 Lumi > 5e31, Lumi < 7e31 6e31 physics column (target lumi)
4 6e31 + 20kHzRandom Lumi > 5e31, Lumi < 7e31 with 20kHz Random Trigger  
5 4.5e31 Lumi < 5e31 4.5e31 physics column

50 Bunches HLT Menu (Ramp-up after TS2)

Column number Column name When to use Notes
0 Emergency The rate is too high To be used in case of emergency
1 DataTaking Use this column to take ZeroBias This column gives 3kHz ZeroBias

Quiet Beam for CTPPS and 1~3 bunches filling scheme (Ramp-up after TS2)



Column number Column name When to use Notes
0 Emergency The rate is too high To be used in case of emergency
1 halfZeorBias Use this column with 3 bunches fill This column gives half rate of ZeroBias
2 oneforthZeroBias If the rate is too high for 3 bunches, use this column This column gives 1/4 of ZeroBias
3 fullZeroBias Use this column for 1 and 2 bunches This column passes all the ZeroBias

low PU fill 26/07/2017 during VdM scan

Very Important: instruction for lowPU runs on 26/07/2017

This fill uses a special run mode: HI_lowPU2017 and HLT menu: /cdaq/physics/Run2017/LowPU25ns/v2.0/HLT/V1.

There WILL BE 24b during ALICE/LHCb scan AND 32b during ATLAS scan. The menu targets total L1 rate < 80kHz and total DAQ bandwidth <~3.3GB/s.

  • for 24b (ALICE/LHCb scan) Start with column 0 and switch to column 4 -> 5 -> 6 based on lumi
  • for 32b (ATLAS scan) Start with column 0 and switch to column 1 -> 2 -> 3 based on lumi

Column number Column name When to use/notes
0 Emergency At flat top and squeeze, To be used in case of emergency Call the HLT DOC if you feel the need to go to this column during stable beams
1 32b_3.2E30 default. PU~0.7, 32b, 0.25M collision rate, expect ~55kHz of L1 and ~25 kHz of HLT rate. 0.117 Mb for event size
2 32b_2.3e30 PU~0.5, 32b, 0.18M collision rate, expect ~50kHz of L1 and ~25 kHz of HLT rate. ~0.1 Mb for event size
3 32b_1.4e30 PU~0.3, 32b, 0.11M collision rate, expect ~65kHz of L1 and ~25 kHz of HLT rate. ~0.1 Mb for event size
4 24b_2.4e30 PU~0.7, 24b, 0.19M collision rate, inst lumi 2.4E30, expect ~45kHz of L1 and ~25 kHz of HLT rate. ~0.1 Mb for event size
5 24b_1.7e30 PU~0.5, 24b, 0.14M collision rate, inst lumi 1.7E30, expect ~45kHz of L1 and ~25 kHz of HLT rate. ~0.1 Mb for event size
6 24b_1e30 PU~0.3, 24b, 0.08M collision rate, inst lumi 1E30, expect ~55kHz of L1 and ~25 kHz of HLT rate. ~0.1 Mb for event size

VdM Scan, 26/07/2017

The current VdM is /cdaq/special/VdMScan/v5.2/HLT/V5, We should use the run mode VDM2017.

The menu used ZeroBias or alternatively L1AlwaysTrue.

Expected L1 rate is ~20kHz, Expected HLT rate is ~20 kHz

Some features of the menu are:

  • Accept 5 BCIDs with very high rate (18.7 kHz total!)
  • Different trigger columns for L1_AlwaysTrue and L1_ZeroBias (both should produce thesame results)
  • 2.5 kHz of ungated ZeroBias for lumi measurements
  • ~300 Hz of triggers for beamspot assessments
  • ~500 Hz of randoms at HLT

Start with column 2 and switch to column 0 (or to 1 in case ZeroBias triggers do not work, call the L1DOC in case L1_ZeroBias rate in column 0 is not the expected (5 * 11.2 / 3 = 18.7 kHz))

Column number Column name When to use/notes
0 Index 0 ZeroBias Default
1 Index 1 AlwaysTrue use in case for any reason the gating of ZeroBias does not work
2 Index 2 Test emergency column

How to Change the Prescale Column

Change of prescale column index during a run

  1. Go to the uGT, commands panel ControlPanels -> uGT Prescales
  2. Click on "Refresh Table". Check the Current Prescale index at the top of the panel.
  3. Click on the column to be chosen.
  4. If the trigger is reconfigured, this needs to be repeated after the reconfiguration.

-- AlessandroThea - 2017-10-25

Edit | Attach | Watch | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions...
Topic revision: r1 - 2017-10-25 - AlessandroThea
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback