AFP To F LowNoise Michael Rijssenbeek LowNoise To
AFP To. F – Low-Noise Michael Rijssenbeek Low-Noise To. F Frontend for Run 3 Ø Planning for LS 2 2016 To. F is good to have but NOT essential To. F is needed 2017 SD DPEjj CEPjj γγ→ W, Z, γ XRP Far – 218 m TCL 6 06 DEC 2018 Patch. Panel – 212 m AFP-To. F@Praha XRP Near – 206 m Q 5 1
To. F Ultra-Low Efficiency in 2017 (Review 6 June) Reasons: 1. PMT life time (≲ 1 C/cm 2) was exceeded: ~3 -6 C/cm 2 gain deterioration 2. PMTs gain goal was 5× 104: actual gain @2 k. V was 1 -2× 104 CFD threshold inefficiency … – Glue transparency deteriorated by <20% (measurements) … Cure: – ALD coating = long-life – Gain goal: ~104 high rate capability to ~8 MHz (12 MHz in Run 3? ) Ømust measure the PMT gain vs HV (To. T) !! In situ !! Note: Log(Gain) is not linear with HV. Also, at low HV, special HV divider ratios apply to maintain p. e. collection eff. – reduce noise/pick-up to ≲ 10 m. V for MIP=~30 m. V at PAb out: 08. 05. 18 Report by Plzen : 5 m. Vpp – add 3 rd stage amplifier PAc + inverter (need ~10× gain): 08. 05. 18 Report by Giulio: OK 06 DEC 2018 AFP-To. F@Praha 2
Shielding, Filtering, etc. • Shielding of all components INSIDE the pot from sources INSIDE the pot (Si. T, …) • Filtering of all electrical lines ENTERING the pot from the outside – LV supply cables: Ferrite choke (note Si. T LV goes on FFC; To. F LV goes on coax) – HV supply cables: RCR filters – Sensor cables: RCR filters – Si. T signal cables: Ferrite choke on FFC • Avoid ground loops and long antenna action • Noise is most effectively reduced at the earliest stage, i. e. PMT + PAa: – shield PMT – Note: possible conflict with PMT HV operation! – shield PAa – lower-noise PAa design? • Reduce Ringing; while this is not noise, it worsens the occupancy and deadtime … – Improve PMT Anode PCB ? – Improve PAa-feedthrough-PAb impedance matching 06 DEC 2018 AFP-To. F@Praha 3
AFP PMT HV & GND Connections Shielding 40 cm Shielding HV PC MCPin MCPout MCP 1 MCP 2 4× PA GND 16× 06 DEC 2018 AFP-To. F@Praha 4
Anode-Side Connections All others #2097, #2096 06 DEC 2018 AFP-To. F@Praha 5
Planacon Modifications (T. Sykora/Alice-FIS Coll. ) First of all, we use 75 Ohm resistors (R 6 -R 21 in fig. 2) at the output of each anode, before the signal from 16 anodes is collected to the total “quadrant” signal. These are needed to damp the “ringing” of the output signals’ falling edge when short pulses are detected. This originates from the resonant (LC) circuit due to the anodes capacitance and the inductance of their leads, causing the falling edge oscillations during the anode potential recovery after short pulses of current are generated. Offthe-self Planacons are not supplied with these 75 -Ohm resistors: we installed them to the voltage divider board when used the standard version (board 3 in fig. 1), and used a dedicated board in the modified version (board 2 in fig. 2). This board featured equalized connection length and actually was a part of the PMT housing - boards 1 and 2 were manufactured at INR RAS and sent to Photonis to be installed to the PMTs. Moreover, the standard version featured a ground plane on the inner side of board 1, which increased the anodes capacitance – this plane was moved to the outer side of the board in the FIT version of Planacons to reduce LC. The other modifications of the Planacon circuit used in FIT are mainly addressed to abolish the positive part of the bipolar cross-talk between anodes. As it was written in the paper mentioned by you, this positive part is directly connected with the common signal formation, since common signal is positive. After deleting the common output load resistance and the common output by itself, any positive part of the cross-talk is no longer presented. Although, there still is the direct negative cross-talk between anodes, whose amplitude reaches ~5% for adjacent anodes and ~1% for the diagonal ones (here I do not include the influence of possible optical cross-talk originating from the 2 mm-thick PMT window when it’s traversed by a charged particle). Comparing our circuit with the one presented in your mail, I may mention that we use also a 100 Ohm resistor (R 5 in fig. 2) to minimize the Q-factor of the circuit of capacitors and their leads. Moreover, the capacitors (C 4 -C 7) had excessively long leads in the standard version – we’ve replaced them with C 4 -C 11 capacitors grounded directly to the large-area ground plane. This could also make certain influence to the pulse shape, including its falling edge. 06 DEC 2018 AFP-To. F@Praha 6
Summary – Noise Suppression Noise suppression is critical to good To. F performance Actions/Responsibilities: • shielding PMT (from in-pot sources) • shielding PAa • Filtering: PMT HV, LV, Temp sensors • PMT Anode PCB design (reduce ringing) • Improve PAa-Feedthrough-PAb impedance matching (new Feedthroughs? ) • Lower-noise PAa? Task Responsibility FTE LS 2/Run 3 Estimated cost MA-MCP-PMT shielding 0. 1/0. 1 ? Filtering 0. 2/0. 1 5 K 0. 1/0 ? 0. 2/0. 1 ? Impedance improvement Lower-noise PAa 06 DEC 2018 AFP-To. F@Praha 7
- Slides: 7