Peter Barnes, LLNL
April 15,
2004
v1
MIPP (E907) TPC
Internal Safety Review
Abstract
The MIPP Experiment (Fermilab E907) uses a 2 m3 time projection chamber (TPC). This note lists questions raised during the MIPP Internal Safety Review, and the responses from the TPC experts.
General Issues
Electrical Issues
High Voltage Issues
Gas Issues
Following the fire in the MIPP (E907) RICH, the MIPP collaboration is executing internal safety reviews of all systems, with special attention to the kinds of issues exposed by the RICH fire. This note documents the various questions that have come up during the review discussions, and the responses from the TPC experts.
1.1. Reviewers (email all):
Peter Barnes pdbarnes@fnal.gov
Craig Dukes dukes@fnal.gov
Pierrick Hanlet hanlet@fnal.gov
Michael Heffner mheffner@fnal.gov
Holger Meyer hmeyer@fnal.gov
Andrew Norman anorman@fnal.gov
1.2. Is documentation sufficient and accessible?
See the sections below for references to the documentation.
1.3. Need to compare what is documented to what is actually there.
Volunteers?
2.1. Documentation:
MIPP (E907) TPC Electrical Description
2.2. Circuit diagram should show cable gauges, fuses, PS limits, PS requirements, PS monitoring, and interlocks.
2.3. Is the cooling system on the sticks truly fail-safe?
Please explain what is meant by fail-safe.
2.4. Could a short in the low voltage system create enough heat to start a fire anywhere?
How much heat can start a fire? The +5 V supply can deliver 15 W; ±15 V can each deliver 12 W; the ±5 V supplies can each deliver 7.5 W.
2.5. Low Voltage monitoring?
Is this a safety issue or data quality issue?
I believe the safety perspective is adequately addressed in the interlock design.
As for data quality, each stick measures its voltages and temperature; the data is part of each event. In the worst case when a power supply trips the stick fails to send data to the VME. The TPC code explicitly checks for this each event.
3.1. Documentation:
MIPP (E907) TPC Electrical Description
This document needs more detailed description of the cathode resistor chains.
3.2. Does all HV have trip limits set?
Both anode and cathode supplies have voltage and current limits set by software. Over-current trips have been tested. The anode control software will refuse to set a voltage over the programmed limit. The cathode supply will similarly refuse to set the voltage over the programmed limit.
3.3. Does ramping require different limits and (if so) should this cause concern?
Neither the anode or cathode supplies implement voltage
ramps. We synthesize ramps in
software by increasing the voltage in steps. The ÒrampÓ is not adiabatic; each step results in an inrush
current, initially 
, where 
is the voltage
step, and 
is the input
impedance of the load. This inrush
current is in addition to the quiescent current at the operating voltage.
In the case of the anodes, the input impedance is just the resistance of the cable, and the quiescent current is essentially zero. Therefore the inrush current is much larger than the quiescent current. Ramping the anodes requires much higher current limits (or even current limits disabled) than the quiescent limit. The LeCroy 1444 supply is capable of delivering 1 mA. Typical operating voltage is 1300 V, so the available power is 1.3 W.
The cathode is at the other extreme: the input resistance is 125 MW, the quiescent current at 10 kV is 80 mA. A 125 V step results in an inrush current of 1 mA. When ramping by hand, we typically increase the trip limit to 120 mA, to allow for the sensitivity of the control knob. We expect that we can use a much smaller current limit during the programmed ramp. (Software control of the cathode is not implemented yet.)
3.4. Is all HV monitored?
The 1444 monitor program monitors the anodes. Software control and monitoring of the cathode is not implemented yet.
4.1. Documentation:
Gas System Diagram (descriptive documentation not yet available).
4.2. Could the P10 supply get contaminated with air/oxygen? (If someone changes bottles, but does not purge the line? If gas runs out for a day? For a week?)
Failure to purge the high pressure line when changing gas bottles (methane, argon dewar, P10 backup cylinders) will introduce oxygen and water (from air) into the P10 supply to the chamber. Likewise, if the gas supply runs out, residual leaks in the TPC will vent the overpressure normally maintained by the oil bubbler. The leaks then become sources for diffusion of water and oxygen into the chamber.
Significant quantities of air can be introduced without leading to a flammable mixture. Haggerty, et al., [1] have measured the flammability range of P10 in air to be 44Ð54% air. Given the 2 m3 volume of the TPC, approximately 1 m3 of air would have to be introduced to achieve a flammable mixture. This does not seem possible to accomplish during a poorly executed bottle change.
Of course, we want to keep the chamber dry, so we would be very remiss to let the gas run out for long enough that diffusion through leaks would become a safety problem.
A sufficient quantity of either water or oxygen (greater than
about 100 ppm) will affect the 
particle
identification measurement, since they are both electro-negative and eat up the
primary ionization electrons. We have
not implemented O2 or H2O monitoring yet. This will be monitoring to ensure data
quality, of course, and doesnÕt directly address possible safety issues.
5.1. What materials is the TPC constructed from? Are they flammable?
The gas envelope is aluminum and stainless steel. The pad plane is (of necessity) copper-clad G4. The wire planes terminate on additional G4 and copper-clad Kapton boards. The lid guard ring is also copper-clad Kapton. The field cage is copper-clad Kapton on each side of a foam core, probably Rohacell. The field cage is supported by an insulating post, material unknown (Delrin?, ceramic?) at each corner.
The front-end sticks are made from a pair of conventional circuit boards mounted back-to-back on an aluminum wedge. These reside below the pad plane, outside of the gas envelope.
[1] H. Haggerty, J. L. Priest, and T. Marshall, ÒFlammability Tests on Run II Muon PDT Gas and P-10 Gas,Ó unpublished, (Fermilab, 2001).