What the Large Hadron Collider scientists teach us about avoiding collisions

-October 14, 2013

Do more with less. This is the working reality today for engineers, with projects getting bigger and resources limited. Even in extremely large projects like the Large Hadron Collider, where project schedules are not measured in “need-by” dates, this is the reality. No one organization, group, or team has either the luxury of reinventing from scratch, or assembling all the expertise in a single desired location, and we end up having to innovate to complete our projects on time and on budget.

It turns out that collaboration is one of the keys to success. An innovative story worth learning about occurred in the ATLAS experiment at the Large Hadron Collider project. Not only did these scientists put active collaboration into practice, they did it across institutional boundaries. Find out how they were able to seamlessly accelerate their own path to success without running into each other along the way.

The Large Hadron Collider project

The Large Hadron Collider Project is the world’s largest and most powerful particle accelerator. The design started in 1994 at CERN, the European Organization for Nuclear Research, located just outside Geneva, Switzerland. The Large Hadron Collider (LHC) accelerates protons (hydrogen ions) through a 27 km long circular tunnel from an initial energy of about 50 MeV (50 mega electron volts) to about 4 TeV (4 tera electron volts), which is an approximate acceleration of 105!  When it starts up again in 2015, after brief maintenance, the final energy will increase to near 8 TeV.

When two protons traveling in the exact opposite direction, each with an energy of 4 TeV, collide inside this tunnel at four observable points, it is believed that the energy density and temperature that are made available are similar to those which existed a few moments after the Big Bang. This allows physicists to study how the universe formed.

According to current theories, shortly after the universe was born, it went through a stage when matter existed in a sort of extremely hot, dense soup called quark gluon plasma, or QGP, composed of the building blocks of matter. When the universe cooled, these quarks became trapped in composite particles such as protons and neutrons. The LHC is also able to reproduce the QGP by accelerating and forcing the collision of beams of heavy ions such as lead, instead of protons.

For the sake of comparison, one must realize that in real terms the energy that we are talking about is not very impressive. For example, 1 TeV is about the energy of motion of a flying mosquito. What makes this impressive is that the LHC squeezes this energy into a space a billion times smaller than one single cell of a mosquito.

Last year around this time, the ATLAS and the CMS experiments at the LHC were able to detect the existence of a Higgs boson particle, one of the cornerstones that explains the origin of mass and why some particles are heavier than others.

Now that we know what the LHC is all about, with collisions being a necessary part, why would these scientists want to avoid collisions?

The ATLAS detector

We briefly mentioned that when these two opposing protons collide, they are observed at four points. One of those detectors (or observers) is the ATLAS detector.

large hadron collider atlas detector
Figure 1 The 27 km tubes of LHC with detectors1

This detector has many different components, which include 40 or 50 different ASIC designs as well as boards, FPGAs and such. A few of these ASICs have the complexity that might require about 10 to 20 IC engineers to design, and no research institute can even find, never mind collocate, that many.

One such chip, labeled FEI42 and described below, is an imaging chip for a hybrid pixel detector built in a 130-nm process. While the basic concept is similar to that of a digital camera, the actual requirements are very different. This pixel-detector SoC contains a large front-end matrix to detect the particle collisions; digital processing for control and readout sits around the periphery. While a camera image sensor might serialize the output of each pixel and read out serially, this FEI4 chip does this in a massively parallel manner, amplifying and digitizing each detector pixel simultaneously to provide an effective frame rate of 40 MHz! Lawrence Berkeley National Lab (LBNL), in Berkeley, CA led the design effort for the FEI4 chip across several worldwide research institutes, including University of Bonn (BONN) in Germany, the NIKHEF Institute for Subatomic Physics (NIKHEF) in Amsterdam, Netherlands, the National Institute of Nuclear Physics in Genoa, Italy (INFN G) and the Center for Particle Physics of Marseilles (CPPM) in Marseilles, France.

Figure 2 shows the block diagram for the FEI4 chip2. The color-coded legend on the right indicates the sites that developed the blocks. (Please note that the diagram is not to scale and is here primarily to show the blocks.)

large hadron collider atlas FEI4 block diagram
Figure 2 The block diagram of ATLAS pixel detector chip FEI4.

Lead scientist at LBNL, Maurice Garcia-Sciveres, explained that the development of this chip did not provide a guaranteed secret recipe for success, but was typical of many R&D projects with evolving specifications. However, some steps they took early in the project to set up management for distributed collaboration were critical to their success.
Next: Next steps

Loading comments...

Write a Comment

To comment please Log In