Project

General

Profile

Electronics used for MicroBooNE DAQ (still to be reviewed !!!)

Introduction

The propose of this page is to list and give basic information regarding all electronics used for the DAQ system.
Most of the listed parts are not under the responsibility of the DAQ team, and those who are will be detailed in the last part of this page.
The electronic readout system is receiving the analog signals from the LArTPC, amplifying, digitizing and passing them on to the DAQ system according to the data flow chart given in The DAQ system .
Some of the electronics receiving the signals from the TPC part of the detector had to be be customized low-noise electronic, capable of operating in cold cryogenic liquid Argon environment and is referred to as "cold electronics".

Cold electronics

Cold ASICs

ASIC is an Application specific integrated circuit. The ASICs includes a preamplifier, shaper, and signal driver. Its output are still analog signals. The ASICs are located inside the cryostat (to minimize the electronic noise by minimizing the wires length to the preamplifier).

Each ASIC handles 16 channels (1 channel = 1 wire)
It has various settings for:
  • Amplification: 4.7, 7.8, 14, and 25 mV/fC, using settings g0, g1, g2, g3, respectively
  • Shapes: 0.5, 1, 2, and 3 us using settings p0, p1, p2, p3, respectively
  • Baseline - the ASIC has programmable baseline settings (200 or 900 mV)

We typically run in g2p2, but some at different configurations due to issues

Cold motherboards

The MicroBooNE motherboards are hosting the ASICs, providing signal interconnections both between the detector wires and
preamplifier inputs as well as between the driver outputs and cold cables to the signal feed-through. It also provides a calibration network
and bias voltage distribution for the wire planes.

The hardware is based on the Rogers 4000 series

There are two kinds of motherboards, to account for the different positions of the wire attachments along the top and sides of the
LArTPC:

  1. Horizontal boards / top version:
    • It is using feedthroughs 2-10 (4 MBs per FT)
    • 12 ASICs per board out of which:
      6 on front - handle induction channels (alternating U and V).
      6 on back - handle collection channels (all Y plane).
      The 12 ASICs handle in total 192 readout channels per board: 96 Y channels, 48 U channels, and 48 V channels.
  1. Vertical boards / side version:
    • It is using feedthroughs 1 and 11 (7 MBs per FT*):
    • 6 ASICs per board, one-sided - Only handle induction channels: U on FT1, V on FT11.
      The 6 ASICs handle 96 readout channels.
      Note: difference in number of channels for these FTs

Warm/readout electronics

  • Intermediate amplifiers, Handle 32 channels each: no mixing of induction or collection
  • Just line drivers: getting analog signal to front-end boards (FEBs) on platform
  • Front-end boards
  • Front: digitizers from BNL
  • Back: Front-end module from Nevis with FPGA
  • The TPC FEM
    1. 2 MHz samples (500 ns per tick)
    2. Basic unit of time = Frame: 3200 ticks = 1.6 ms =expected drift time for 500 V/cm
    3. NU (triggered) stream for TPC data
    4. Receives trigger info (frame and sample number)
    5. Makes three new 1.6 ms-“windows” out of those frames, with trigger sample at beginning of second window. Full waveform readout.
    6. Option to Huffman compress data for those windows, We do this by defaultàcompression factor ~5
    7. There is an SN (continuous) stream too which does not readout the full waveform
      • Handle 64 channels total (most have 32 collection and 32 induction, but some have 64 induction)
      • TPC FEMs are arrayed in 9 readout crates: most have 15 FEMs, but crate 1 has 12, and crate 9 has 14
      • Each crate has a controller card, which configures crate, provides monitoring readback, and receives trigger information
      • Connected to DAQ PC (SEB) via optical cable link
      • Each crate has an XMIT module, which attaches header and trailer info per event, and sends data via optical link from crate to PCIe cards on our DAQ PCs (the Sub-event-buffers, or SEBs)
      • Two cables per stream, one stream per PCIe card → 3 PCIe cards per TPC SEB

Optical and trigger readout

See DocDB 6534 for a good review of the optical readout from top to bottom
Both readout applications run on SEB10:
PMTs assigned fragment id “10”
Trigger assigned fragment id “11”

For optical readout, 3 FEMs:
  • FEM1: High gain, beam “discriminator” only
  • FEM2: Low gain, beam+cosmic discriminators
  • FEM3: High gain, cosmic discriminator only
  • FEM1+2 receive some beam pulses on channels, but nothing on FEM3 à when running with PMT HV off, need to exclude FEM3 (no data!)

64 MHz clock à 15.625 ns per sample

DAQ dedicated electronics

As mentioned most of the electronics are not under the direct DAQ team responsibility. Most of it is on the platform and off limit for the DAQ team.
The DAQ room host the DAQ servers. note that the maintenance of these servers are under the supervision of the SLAM team lead by Bonnie King. You can learn more about the online machines running on these servers here.
There are, however a few parts that are falling under our responsibility in there are herby listed:

  1. PCIe Card w/ fuse. These card were made by NAVIS and they are also responsible for them. contact person: Georgia Karagiorgi,
  2. 10 Gigabit Network Interface Card, Intel.
  3. GPS card
  4. usb ranger

Spare parts

Below is the most up-to-date list of spare DAQ machines and parts, which should be trusted over the google doc above.

last update: 02/14/2018
last update with the help of SLAM: 04/16/2018

  • PCIe Card w/ fuse
    • Manufacturer: Nevis Labs
    • used in uboone: 5
    • spares: 8 (02/16/2018)
      • 1 in DAQ room cabinet ("Nevis Spares Box”). This is the preferred spare if one of the PCIe cards in production fail.
      • 4 in ubdaq-prod-smc2 . These should be the second choice if more PCIe cards in production fail.
      • 3 in uboonedaq-seb01 (the test stand seb01). We should avoid taking these PCIe cards out of uboonedaq-seb01 unless we have already used all the other spares. The test stand seb01 is still connected to a test crate, and can be used to test DAQ features before deploying them into production.
  • 2TB disk
    • Manufacturer: Seagate
    • Used in uboone as RAIDs on SEBs, EVB, NEAR1
    • Spares: 6
  • 500 GB disk (for SEB10)
    • Manufacturer: Seagate
    • used in uboone as system mirrors on SEBs, EVB, NEAR1, disks in gateways
    • spares: 1
      • 1 on LArTF computing room shelf
  • WD 2.0 TB
    • Used in can be used in uboone for RAIDs on SEBs, EVB, NEAR1
    • Spares: 2
  • 4TB disks
    • used in uboone in ubdaq-prod-smc
    • NO SPARES!!!
  • EVB machine
    • used in uboone: 1 (ubdaq-prod-evb)
    • spares: 2
      • ubdaq-prod-near1: also currently in use for uboone, but intended as first-option hot spare for ubdaq-prod-evb. See the plans for what to do in case of machine failure.
      • uboonedaq-evb ("test stand machine") in LArTF computing room rack 01.
  • SEB machines
  • 10 Gigabit Network Interface Card
    • Manufacturer: Intel
    • used in uboone: 2
    • spares: 3
      • 2 in LArTF computing room shelf ( on is to be installed in evb test stand )
      • 1 in ubdaq-prod-near2
  • GPS card
    • used in uboone: 1
    • spares: 1
      • 1 in ubdaq-prod-smc2
  • USB ranger
    • used in uboone: 2
    • spares: 0
  • Power supply
    • used in uboone for each one of the machines
    • spares: 2 at SLAM

We are waiting for KOI list of spare part and to understand which one of the disk still has a warranty