HEC Parameters from As-Built knowledge HEC is a homogeneous detector and any mechanical tolerances are well below any significance for physics..... this.

HEC Parameters from As-Built knowledge

HEC is a homogeneous detector and any mechanical tolerances are well below any significance for physics..... this holds also for the cabling (match for timing, impedance) and calibration.

So in first order I have nothing

further of significance to say

HEC Simplified

• 12.5, 25 or 50 mm copper plates

• 8.5 mm gaps

• SS Tie Rods

• 2 mm inter-module gaps

• 2 readout depths per wheel

Outline

Module Production Data (e.g. Copper thickness)Wheel Geometry (e.g. Alignment of readout elements, absorber)Electrical parameters (e.g. Cold Amplifier Properties)Corrections/parameters for cluster reconstruction from past test beam experience HV peculiarities Others

High current running effects: Pile-up & Ion-Buildup X-talk

I will make tentative

“recommendations” for each of

these aspects in a red font

Module Production Data…Variation in Production

• Anticipated RMS Signal Variation from Manufacturing Variations: ~ 0.3%

Recommendation: make no initial corrections based on variations observed in the production data.

See: http://atlas.web.cern.ch/Atlas/GROUPS/LIQARGSTORE/HEC_Calo/Database/Documentation/HEC_as_builtFINAL.pdf

Module Production Data … Average Values

See: http://atlas.web.cern.ch/Atlas/GROUPS/LIQARGSTORE/HEC_Calo/Database/Documentation/HEC_as_builtFINAL.pdf

and http://atlas.web.cern.ch/Atlas/GROUPS/LIQARGEXT/TDR/html.1812/LARG-TDR-690.html

We recommend the following values be used in the Monte-Carlo (cold):

Densities (at 88K):

Copper: 9.01 g/cm3 Argon: 1.3895 g/cm3

Kapton & Glue in PAD and EST Boards: 1.45 g/cm3

Volume of Argon Displaced by Honeycomb: 3.1%

Thickness (mm):

Copper: 12.500, 25.010, 50.001 (warm)

12.464, 24.938, 49.8575 (88K)

PAD (incl copper): 0.335 (warm), 0.333 (88K)

EST: 0.1513 (warm), 0.1504 (88K)

8.5 mm Cu-Cu Gap: Front 8.495 mm, Rear 8.468 mm (warm)

Front 8.457 mm, Rear 8.440 mm (88K)

8.5mm Spacers are SS: 3161.584 x 10-5

m/m

Wheel Geometry

The anticipated contributions to the variation from nominal within a HEC wheel:

RMS in x and y : 1 mm (Cu radius variation 0.5mm) RMS in z : 0.5 mm

It is recommended that we start initially with the assumption that each HEC wheel has only 6 alignment parameters: Δx, Δy, Δz, Δθx, Δθy, Δθz, or equivalent. These will be determined using muons when cold. Otherwise the alignment is as in the MC. A minor concern is that depending on the exact handling of wheels and cryostats, the readout boards (which define the readout geometry) might be displaced downwards with respect to the copper. Hence it might be advisable to have 2 separate alignment parameter sets for each wheel (one for copper, one for electrodes). This relative drop (copper to PAD boards) would be of order 0.5 mm.

Electrical parameters

Perhaps the biggest concern is the variation of the parameters of the cold amplifiers. These are detailed in our Wheel Assembly Database (See: http://atlas.web.cern.ch/Atlas/GROUPS/LIQARGSTORE/HEC_Calo/Database/Documentation/HEC_Wheel_DB_7.pdf )

These are only an issue if the electronic calibration does not work for these channels

Recommendation: Assume electronic calibration is robust, and when not available reliable estimates will be provided using the parameters in the database

Electrical parameters

We have established in the test beam that we can reliably move from the electronic calibration (with an exponentially falling input pulse) to a physics signal calibration (with a triangle pulse). All signal shape parameters will be measured in ECC/ECA cold tests, running full calibration + delay curves (also at the same time checking the channel timing). Please note that while the copper is twice as thick in the rear HEC wheels (being 50mm). This is, to first order, compensated for in the electronics in the preshapers, they have twice higher gain in LS3, LS4. We recommend using the electronic calibration corrected pulse height as our first best estimate of the energy deposited in a cell. This will later be corrected either by e/h or energy density corrections… once the cell is part of a cluster.

Corrections/parameters for cluster reconstruction from past test beam experience

For the EndCap we must combine HEC with EMEC to answer this question. We anticipate using a energy density in the shower corrections to get the energy.

We recommend no corrections due to the HEC being less absolutely uniform in its response due to such things as: tie rods, cabling notches, etc. We can only detect these with muons (and electrons in the test beam). Their effect is very

minor. They are in the Monte-Carlo.

HV peculiarities … Description of HV system

Each HEC gap has 4 sub-gaps. Loss of one of these gaps causes loss of, to very good approximation, ¼ of the signal of that gap. Only if all 4 HV fail do we lose a readout volume. To good approximation the only loss when HV is lost is in signal/noise. So each readout volume has a correction factor from HV failure: 4/3 , 4/2 , 4/1, dead. However….

HV peculiarities …Details

In addition to HV failures in readout volumes we can have single gaps without HV or with a tile not connected. These non-conformances will be documented initially in the wheel assembly database. Recommendation: A single cell correction factor for all effects not covered by electronic calibration shall be assigned to each readout volume. These initially will come from the known non-conformance and MC estimates of the effects of these failures.

Once identified with a cluster the effects of the non-conformance will be fine tuned (missing gaps at the observed shower max will have a bigger correction than those away from the dense part of the shower).

In time these cell correction factors could also contain any small corrections identified in the data. From the test beam we anticipate these small corrections to be less than 0.5%.

Other Effects:High current running effects

Once the LHC attains full luminosity a number of effects will come into play:

Varying HV during data taking Varying ion-buildup during data taking Varying Pile-up

These are slow time varying effects that will depend on the current draw in the HV. Missing pulses in the LHC chain will cause pile-up to change the pedestals of our readout volumes. Recommendation: A common LArg strategy should be developed to provide a correction function for each readout volume that is (to first order) solely dependent on the known HV current in the power supplies feeding that volume and the time of the event relative to the LHC pulse chain.

Other Effects:High current running effects

HEC is the last EndCap detector to be effected by ion build-up.

Other Effects:High current running effects

X-talkX-talk (~1% in HEC) is not a significant issue during low

intensity running, because clusters can be enlarged to include the cells with significant X-talk in them.

At high intensity, to enhance S/N clusters will have to be as small as possible.

Recommendation: For high intensity studies resistive X-talk should be added to the MC. For the HEC X-talk is to nearest neighbours in the module. Adjacent volumes in different modules have essentially no X-talk.

Other Effects: High current running

effects… dominant X-talk Path