July 2002
In this document we quickly
summarize the risks of the EMTiming project and our plans. The primary design
consideration (other than cost and physics motivation) for this system is risk.
For this reason we have made the entire system connectorized so it can be
disassembled quickly (approx 1 day) such that the detector is easily returned to
its pre-modified state. This effectively minimizes the risk. Thus, we are
covered for the unlikely case where the design doesn't work, or somehow
compromises the existing system, or the case if part of the project is
uncompleted. We have also chosen to only use off the shelf parts and simple
construction techniques, and re-used parts or designs. This minimizes the risk
of large cost overruns, schedule delays and technical and scope risks. Here we
have tried to identify the risks, both in terms of the things which easily fit
into the resource loaded schedule, and those which transcend boundaries and are
not easily put into that scheme.
Specific questions:
Q: What if,
after installation, the system interferes with the operation of the CEM?
A: If the system is determined to be interfering it can be disconnected
quickly (~1 day) and the detector returned to pre-installation running
conditions until the problem is corrected. It can then easily be re-connected.
Q: What if the project is late in installation?
A: Currently the
project is not on the critical path for installation. If the project is late, we
will not slow down the rest of the Run IIb installation as we do not need the
detector to be rolled out for installation (other than 40 cables). Thus, there
should be no need to wait for us to install. It will be incumbent upon us to
finish during other shutdown times. Our system is completely connectorized so
that the entire system may be decoupled from the rest of the detector with a
day's worth of work and we will install when available and not interfere in the
meantime.
Q: What is the risk of injuring the detector during
installation?
A: Our primary installation on the detector is laying
cables (the splitters) from the EM phototubes to the readout crates and we plan
to have professionals (Dervin Allen etc. has agreed to help) helping us with the
installation and the dressing. As they have done this many times before this
should minimize any potential risks. The other cable installation is from the
racks upstairs. Again this is done under the supervision of professional FNAL
technicians. The installation of the Transition Boards and ASD is simply
plugging in the boards downstairs as is standard operating procedure for the
current HADTDC system. The remaining work is upstairs where installation issues
are smaller.
Q: What if the splitters affect the readout of the CEM?
A: Every attempt will be made to minimize the impact of the splitter
readout on the CEM. If it is found that they do not work they will simply be
disconnected and the anode line to the ADMEM will be reconnected (it is a LEMO
connector) directly to the PMT bases. In this way, there is no impact on the
system as it the same as before any installation. If there is enough time before
the Run IIb shutdown, we will investigate the possibility of returning to the
base modification scenario. In the PEM, there are no splitters so this is not a
risk. The PEM harnesses are simple rg174 cables and 50 ohm terminators which are
identical to those currently in the system except that they have LEMO's on the
ends and not soldered directly to the boards (the LEMO connection is the
preferred solution from a noise and reliability standpoint).
Q: What if
there are no spare TDC's for the system after all?
A: While funding
issues will need to be settled, both Argonne and the University of Chicago have
indicated an interest in producing new boards. This would probably double the
cost, but they could still be produced by the 2005 installation deadline. Since
they are not to be installed downstairs their installation is not coupled to the
Run IIb shutdown.
Q: What if we can only procure a fraction of the
components?
A: From a schedule point of view, since the system is
modular we can instrument parts of the system as pieces become available and
still have physics impact along the way. From a final system point of view, for
example, if we only get half the ASD's or TDC's we would consider not
instrumenting the PEM, or only instrument half of the tubes if we could show
that the turn-on curves are acceptable.
Q: What are the research and
development risks?
There are very small risks associated with the
research and development phase of the project. The parts costs are small
compared to the overall project and the labor costs are all University people.
The schedule is driven by the Run IIb shutdown date, so the probability of our
R&D phase extending beyond 2005 is extremely small. The scope and technical
aspects of the project shouldn't be affected unless the prototypes don't work,
and there is only a small chance of that. If the current splitter solution
doesn't work, we have lots of time to fix it, and in the worst case scenario we
can investigate the possibility of going back to the base modification scheme,
which will cause schedule slippage, but will still fit within the current Run
IIb shutdown time. If the ASD->TDC cable does not perform as expected in the
final tests, we can upgrade to a higher quality cable for a small cost increase
relative to the overall project cost. The ASD's, TDC and PEM harnesses are
already in use in the existing HADTDC system to the probability of them not
working as expected is very small.
Q: What are the production risks for
the various parts?
All pieces of the various components are off the
shelf parts, the dominant costs are the ASD's and the TDC's, and the schedule is
driven by availability of the detector for installation. All the parts are
either purchased or we have quotes from reputable companies and the lead times
are small compared to the time until the detector is open. Therefore even a
large increase in the percentage cost (which is low probability) would have a
small impact on the overall cost. The labor and testing of the components is
small both in terms of time (compared to the shutdown time) and overall cost.
The dominant risks are the low probability occurrence that we cannot
create all the ASD or TDC boards since they have larger technical design
requirements. In this case, there is risk to the cost (it would cost more to
make new ones from scratch), and if we couldn't make all of them there is a risk
that the scope of the project would have to be reduced.
Q: What are the
other risks?
There is a risk that we can't get or easily get the crate
parts. While these have low probability, we can potentially generate spares from
test stands. A second option is to slightly modify the crates downstairs and
install the TDC's in those crates. However, many tests to insure noise problems
would have to be studied.