Send mail to Ken Read
S. Held
University of Tennessee
(phenix-muon-96-12; submitted: September 16, 1996)
This is a modest attempt to collect information in one place concerning past experience by selected experiments with Iarocci tubes/ plastic proportional tubes/ streamer tubes. The information provided about other experiments' applications is incomplete and is mainly selected to provide a comparison with current PHENIX muon identifier applications and concerns. This is an internal PHENIX document and not meant to be an authoritative reference concerning other experiments. We apologize in advance for any inaccuracies concerning other experiments. Corrections and additions are WELCOME! This document will be revised accordingly.
Streamer mode.
Proportional mode. See "A muon identification detector for B-physics near e+e- ==> BBbar threshold," D. Bortoletto et al., NIM A320 (1992) 114-127.
See G. Alexeev et al. "Studies of stability and systematics of operation of the DELPHI plastic tubes," NIM A292 (1990) 551.
Proportional mode. See "The D0 Muon System Upgrade," by R. Jayanti et al., D0 Note Number 2780. This paper was the proposal for the forward muon detectors consisting of Iarocci tubes (1 cm x 1 cm x 8 channels plastic extrusion) with 50 micron anode wires up to 5 m long. The surface resistivity of the cathode paint is 1 kOhm/square. A fast gas is used Ar(80%) + CF4(10%) + CH4(10%) prividing a maximum electron drift time interval of 60 ns.
Also, see:
Proportional mode with 1.5 m long plastic tubes and 50 micron anode wires. Contact is Lou Remsberg (remsberg@nc6.chm.bnl.gov), 516-344-4335, BNL Chemistry Department. Also, Steve Gushue (gushue@nc7.chm.bnl.gov), 516-344-4336.
Chose proportional mode partly because at the time WA80 had experienced some difficulties operating in streamer mode. E802 was concerned about possible dead time. Thus, proportional mode was chosen out of caution and not because streamer mode was known to be unworkable at the AGS. E802 used pad readout with preamps for the pads. The tubes operated at 1900 V. The tubes were extruded conductive plastic tubes manufactured in Japan. The gas was Ar(60%)+Iso(40%). The tube cross section was 1 cm x 1.5 cm. The pad sizes were 7 mm x 2 to 3 cm, 3 cm x 4 cm, and 4 cm x 6 cm. The coverage was from 5 degrees to 45 degrees from the beam line (around a pseodorapidity of 3). The occupancy was 5% per pad. A particle going down the length of a tube could create a current of 1 uA in a wire.
They used a LeCroy streamer tube readout system (from Tunnel). There was a preamp, discriminator, and one-shot for each pad. There were two chanels per chip. The preamp was a grounded-base basically charge-sensitive design. The input signal was about 2 mV and the output was 70 to 80 mV. The discriminator was leading edge. [As an aside, during assembly the elctronics all appeared to behaving fine up to a point. Once the full detector was loaded with electronics, they experienced oscillations due to inductive coupling. The problem was subsequently solved.]
[Note that E871, a rare kaon decay experiment, (Kane, Molson, Walecka, Wojicki, Ritchie, ...) in AGS B5 line uses long aluminum proportional tubes with a trapezoidal cross section.]
Streamer mode. Extruded Noryl tubes (not PVC). See "The OPAL Muon Barrel Detector," R.J. Akers et al., NIM A357, 253-273; "Production and Testing of Limited Streamer Tubes for the End-Cap Muon Subdetector of OPAL," G.T.J. Arnison et al., NIM A294 (1990) 431-438.
Streamer mode. An ORNL/Univ. Tennessee/Georgia State test-beam experiment that ran in the AGS B2 line at BNL in summer 1994. Used 2 m long streamer tubes manufactured by Hodotector, Inc. (Texas).
Streamer mode. Tubes manufactured in Italy. Electronics by MIT. Wire electronics from Italy. Present tube operation overseen by U. Wisconsin. Contact = Jim Johnson (jkj@slac.stanford.edu) 415-926-2928.
Their longest tubes are 8.1 m long. The tubes are connected serially for gas flow in daisy-chains up to 43 tubes long. The inlet pressure is about 1.4" H20. The outlet pressure is about 0.15" H20.
See "Experience with Iarocci Tubes Produced on a Large Scale," Wit Busza, NIM A265, 210-217; "The Limited Streamer Tubes of the SLD," A.C. Benvenuti et al., NIM A290, 353-369; "The Front End Electronics and the Fastbus Readout Module for the SLD Limited Streamer Tubes," F. Beconcini et al., NIM A277, 222-229.
Streamer mode. LeCroy electronics. See their circa
1985-6 reference article.
WA80
Streamer mode. Pad readout. See 1988 NIM paper for details. (This section does not address new WA98 tubes from Dubna.)
In 1985 and 1986, E802 and WA80 had concerns about hit-rate effects with Iarocci tubes. E802 chose proportional mode and WA80 chose streamer mode.
The WA80 pads were as small as 1.5 cm x 2 cm. The electronics was yes/no and consisted of a discriminator, latch, and a one-shot. WA80 experienced currents of 2 to 3 uA per tube during a spill. This lead to current draws of 60 uA or more for groups of 24 tubes sharing a single CAEN HV channel. The WA80 tubes ran at a lower rate than the E802 tubes (in 1992 at least).
Streamer mode with invididual wire readout. The operating voltage is about 4.6 kV. A typical signal is 30 pC. The gas is Ar(3%)+Isobutane(8%)+C02(89%). They were not allowed to use Iso(75%)+Ar(25%) which is explosive. The maximum drift time interval is 100 ns. (The HERA bunch crossing time is 96 ns. The ZEUS level-1 trigger latency is 5 us.) They electrically OR tubes staggered by one-half of a cell. The tubes are extruded Noryl (not PVC) because of safety regulations. (Note that PVC is cheaper and easier to extrude.) The tubes are arranged in supertubes which are glued to a honeycomb with Aerodyte. This allows a streamer tube to slide in or out of the supertube for replacement. All wires are oriented parallel to the beam.
The ZEUS inner barrel gets background from the uranium. The outer barrel (which has 11 m long tubes) has a very low "dark current." To reduce the gas flow impedance of a tube, they maximized the inlet hole in the endcap plug. (Requiring an external injection-molded plastic threaded mating gas connector.) They typically connect twenty 11 m tubes serially with two gas volume exchanges per day.
They run at 4.6 kV (from a modified CAEN supply), follow atmospheric pressure, and have active gain monitoring. The HV distribution connects 30 tubes to one HV channel; however, they can readout the current for and turn on/off individual tubes. They use eight HV decoupling capacitors per tube (this is a "weak point") and automatically invoke current limiting by dropping the HV if the current on a tube exceeds 1 uA (for over 2 hours). The HV is reduced to 2 kV during injection. A "good" dark current is considered to be 50 nA or less per wire. They note that longevity could be an issue for a high rate experiment.
They use very low resistivity tubes and took care to keep the profiles very clean during assembly. The electronic connector cards were made in Padua. A card-edge connector plugs in to the LST endcap. A 30 cm cable connects that card to the electronic board. Differing cable lengths are compensated for by choice of capacitor values.
The overall tube mortality has been 9% for the inner, outer, and rear tubes since 1989. Most of the failure was experienced in the early years with a settling down of the rate. Note that the automatic current-limiting system was not installed until 1993 and is attributed with playing a significant role in safeguarding the tube longevity.
See "The Zeus Barrel and Rear Muon Detector," NIM A 333 (1993) 342 and the "Blue Book" which ZEUS colleagues at BNL have. Contact is Umberto Dosselli (dosselli@padova.infn.it) 011 39 49 82 77076 or 011 39 49 82 75111 (operator) or 011 39 49 82 92419 (Linario Labs).
We thank Umberto Dosselli (ZEUS), Max Meoni (Pol.Hi.Tech.), Jim Johnson (SLD), Lou Remsberg (E802), Mike Tannenbaum, and Glenn Young (WA80) for discussions concerning their respective experiments.