sPHENIX tracking discussion

["Frawley, Anthony" <afrawley AT fsu.edu>]

The minutes of yesterday's sPHENIX tracking meeting are below.

Cheers

Tony

sPHENIX tracking meeting June 17, 2016:
-------------------------------------------------------
TPC status report - Tom, Carlos, Sourav, Veronica:
-----------------------------------------------------------------
Tom reported that they are close to finishing the purchasing cycle for prototype parts.

Carlos made an initial report on TPC performance using the current sPHENIX framework (see posted slides).
Mike McCumber visited SB last Friday, gave a good tutorial on use of the sPHENIX framework to simulate TPC tracking to Tom, Carlos, Sourav, Veronica.
Since then they have been looking at the large TPC memory use. Mike had suggested clusterizer memory use may be the culprit.
They investigated that, found that it does not seem to be the problem.
Mike: We need to go to a memory profiler (Massive (sp?)).
Where to next? Carlos said they want to understand the implementation of the algorithms, so they can be useful contributors to developing the code.

Status report on MAPS, and TPC code merge - Mike McCumber:
-------------------------------------------------------------------------------------
MAPS has (internal to sPHENIX project only) project review on 30th June.
LANL LDRD decision day is July 7th.

Mike has forked the code in the master so that the TPC and silicon tracking do not interfere.
He chose Alan's tune that has high efficiency, but poorer momentum resolution - will work more on the momentum resolution.

Now he has turned to thinking about pileup (see slides 5-8 for nice visualizations of the problem).
The drift speed quoted is 6 cm/microsec, but this depends on the gas. Tom said that Neon gas is slower (4 cm/microsec)

Slide 8:
Mike had to modify the cell creation code to use Geant time and the drift time, it did not previously do this.
Found that tracks from the previous time bucket (106 ns earlier) fail to connect between the TPC and inner barrel.
This is due to the fast drift time in the TPC - they move 6 mm between crossings, which is enough so that they are killed by the chisq cut.
Still have to look at the effects of high occupancy though.

Mike commented that if the standalone TPC+maps tracking does not suffice to resolve pileup issues, we could use in addition:
  The calorimeters
  A fast BBC to get the vertex position accurately
  A fast intermediate tracker to disambiguate time and position

Tracking capability of intermediate tracker + thinner outer tracker - Gaku Mitsuka:
--------------------------------------------------------------------------------------------------------------
Gaku presented results of simulations for a tracker consisting of:
  1 layer of VTX pixels at 2.48 cm.
  4 layer intermediate silicon strip tracker at 6-12 cm
  6 plane drift chamber at about 77.5-85 cm (mocked up by silicon strips in the simulation).

The proposed outside tracker is a drift chamber (a jet cell chamber ala PHENIX)
A 2-track separation of 1 mm is possible.
The integration time of the DC would be ~1 microsec.
It would have no z resolution, the z position is provided by the pixels, intermediate tracker and EMCal. However the intermediate tracker here is a strip tracker with 12 mm long strips.  

Slide 4: The wiggle in momentum resolution at 10 GeV/c may indicate that the Eigen bug was present when these simulations were done.

There was a long discussion, which I will not try to reproduce in detail here.
The consensus was that issues with jet chambers in this application need to be understood.
The viability of the proposal for a DC at this track density needs to be understood. Need Garfield simulations.

Ed requested that the concept for the outer tracker be written up in greater detail and presented in 2 weeks.

We meet again next week:
-------------------------------------
We agreed to meet again in one week (June 24) to discuss TPC and MAPS issues, and preparation for the tracking review.