Outline:

1
XCES Input Path Status
Chris Neu, Daniel Whiteson, Rick Van Berg L2
Upgrade Meeting 11 June 2004
Outline

• Current status
• Active splitter
• Plan

2
XCES Input Path Status
Recall from last time

• 48 XCES input channels for L2
• Tx downstairs use Taxi
• Optical Tx strength varies component-to-component
• Weak optical Tx strength rampant half too weak
  to split others marginal or sufficient
• 16 split channels
• 16 NetOptics splitters
• Housed near bottom of rack 2RR31C (near
  entrypoint for XCES fibers from CH)
• 2U chassis, Siemon patch panels on front and rear
  (thanks to Dervins guys for fabrication)
• Inputs fed in through back
• Half of outputs sent out the back, half out the
  front (particularly convenient)
• Inputs to splitters coupled directly to XCES
  fibers at original patch panels
• (to avoid moving input fibers)
• Have each output marked with ID and loss value
  to facilitate channel swapping

3
XCES Input Path Status
Pulsar channel XCES channel Taxi Tx loss (-dBm)
0 0W 15.2
1 0E 15.91
2 12E 15.56
3 15W 16
4 14E 15.73
5 3W 14.67
6 15E 15.47
7 16W 15.12
8 5W 13.22
9 19W 15.69
10 7W 15.33
11 19E 15.68
12 20E 15.5
13 9W 15.08
14 21W 14.82
15 21E 14.97

• Staged 16 new 20 long fibers
• feed pulsar board
• Left 4 fibers from stage 0 in as spares
• Installed 16 splitters on 4/15
• Chose 16 strongest Tx channels for splitting
• Tested with L2 torture
• Tev studies time my favorite time
• Script from Masa allows user to adjust by hand
  the XCES thresholds to tune occupancy
• During studies one can mimic real data
• All looked well with Run 2A system
• Beam tests

4
Results of Beam Tests Run 2A system
ADC counts and turn-on curve for high and low
XCES thresholds.
5
Results of Beam Tests Run 2A System
split channel
6
Results of Beam Tests Run 2A system
Same run, restricting attention to events away
from threshold to remove turn-on effects
7
Results of Beam Tests Run 2B system
Two references for Pulsar path, each with their
own quirks
TL2D as reference
CESD simulation as reference
Rare differences
Compare Pulsar data to threshold trigger bits in
alpha
Compare Pulsar data to simulation of trigger
threshold
8
Results of Beam Tests Run 2B system
Using absolute reference L2Torture test with
patterns in the SMXR thresholds so that the bits
were predictable. In 1 million bits
Run 2B TP2D Zero errors!
Plots like these to be available in PulsarMon!
Run 2A TL2D Rare errors
9
Commissioning Plan Timeline Made Before Tx
Power Problems IDd
3/15 Stage0 splitter box complete DONE 3/16
Install Stage0 splitter box DONE 3/15-3/22
Perform optical measurements on all 48 XCES
signals coming from downstairs DONE 3/22 CDF
power back 3/22-4/1 L2 Torture testing
debug/use/refine monitoring code DONE 3/23-4/10
Beam tests after understanding L2 Torture
DONE 4/10 Stage0 complete DONE 4/10-5/1
Develop DAQ code for XCES readout DONE 5/1
Stage1 splitter boxes completeDONE 5/7 FW for
driving out over SLink complete
DONE 5/1-5/7 Install Stage1 splitter boxes in
systemDONE 5/7-6/1 L2 Torture and beam tests of
split system DONE L2T of Slink
output Progress Made BR, SP, Frans et
al. 5/7-6/1 Install and test XCES Merger link
Progress Made 6/1 Stage1 complete 6/2004
Build and install Stage2 splitter boxes .
complicated by active splitter End of 6/2004
Have N Taxi Rx boards in hand ready for system (N
2 spares) READY thanks to others 7/2004
Install 2 more Taxi Rx boards in system test
with L2 Torture, beam 8/2004 Burn-in time in
beam readout by default monitor for
errors 8/24/2004 Stage2 complete 8/25/2004
10-week shutdown begins
10
The Virtues of Splitting All 48 Channels

• Commissioning of system possible with only 16
  split channels
• But painful
• Turn-on of XCES path cannot be parasitic-gradual
  
• would need to be threshold moment
• this is not optimal
• Future would like to have a hot spare Pulsar
  system ready
• being fed parasitically with upgrade in place
• Will make things clean
• each channel treated equivalently

11
Design of Active Splitter
RVB and I came up with this design at Penn a
couple weeks back
Taxi Rx HFBR 2416
AD96687
MC10H115
Taxi Tx HFBR 1414
-5V
-5V
Variable Resistor To set Comparison Voltage
Taxi Tx HFBR 1414
-5V
Prototype has been built at Penn will be here at
FNAL for testing next week.
12
Two Possible Paths and Timelines

• Delay getting prototype here from Penn
• Atlas test beam work required significant
  attention
• Things calmer now
• Once the hand-made prototype is here we can test
  it using Taxi Tx-Rx setup in teststand crate
• Those tests will determine if basic circuit logic
  is working and whether we can proceed
• Would like to solicit help getting testing off
  the ground and ultimately testing all boards once
  they arrive
• Need 48 channels split spares so lets say we
  need 60 split channels
• Choices to consider 2 channels-per-board x30, or
  1 channel-per-board x60
• Two possible paths
• Use PCB
• Make all splitters by hand
• Timeline estimates are different

13
Two Possible Paths and Timelines
PCB
Assemble by hand
6/14-6/18 test prototype here at FNAL work on
layout (4 layer, at Penn) 6/19-6/25 finalize
layout and determine mechanical issues
6/28 submit design for fabrication
7/12 delivery of boards 7/12-7/18 stuffing of
parts at Penn boards start to arrive at FNAL
on daily basis en-masse testing
begins 7/14-8/1 testing in situ 8/1
all 48 channels split
6/14-6/18 test prototype here at FNAL
finalize circuit determine mechanical
issues 6/19-7/1 assemble at Penn using vector
boards (2 students x several per day)
boards start to arrive at FNAL on daily
basis testing as they arrive 7/1-7/18 assemble
boxes, install test in situ 7/18 all 48
channels split
Schedule complicated by honeymoon in Spain and
Italy leaving in 7 days. But I will have time
to test the prototype and look into mechanical
issues before leaving.
14
Logistical Issues Splitting 48 Channels

• Input fibers from CH come into the rear of the
  rack
• Prefer to leave them in place do not want to
  disturb these
• So input to splitters should be at rear of rack
• Output fibers (for convenience) should come out
  the back as well
• Have 18 inches vertical space for splitter
  chassis
• Thinking about 2 rear-facing chassis on shelves
  looking like this
• Power comes from a 5V supply plugged into power
  strip boards get power across a simple backplane
  or jumpered togetherneed to talk to Bob DeMaat.

7in
17in
15
Summary

• Stage 0 and Stage 1 splitting completed ahead of
  schedule
• Stage 2 splitting complicated by need for active
  splitters
• Have an active splitter design and prototype
  built
• Needs testing - will be done next week
• Decision on path needs to be discussed
• Cost issues?
• Will need help testing boards
• Lots of things need to be done.