SLAC Today logo

SLAC Accelerator Expertise at Work in the LHC

(Photo)
Themis Mastoridis and Claudio Rivetta characterize LHC radio-frequency stations from their SLAC office using the software they helped design to manage the LHC RF system. (Photo courtesy of Dan Van Winkle.)

CERN's Large Hadron Collider became the world's highest energy particle collider when it began colliding protons at 1.18 tera-electron volts late last month. This was no small feat; it takes hundreds of mechanical and software systems to run such a machine—including a system governing the radio frequency waves that give particles their boost inside the accelerator. With their combined 25 years of experience on just such a system, a SLAC team of researchers helped CERN solve an unanticipated complexity with this LHC system, helping get the giant machine up and running.

"Our team is the place where technology and accelerator physics come together," said SLAC accelerator physicist John Fox. "We know particle beams and their dynamics, we have technical expertise with accelerator RF systems and feedback, and we’ve learned a lot understanding their interaction." Under Fox's leadership, SLAC researchers Dan Van Winkle, Themis Mastoridis and Claudio Rivetta used their knowledge of RF systems to design new software tools needed to commission and optimize the LHC RF system.

As in other accelerators, it's very important to keep both the LHC's beam and RF systems stable. The RF system governs the radio waves that push particles to ever higher energies, much as ocean waves help a surfer gain speed. But if an RF system is unstable or poorly optimized to the particle beam, it can inadvertently change the beam's shape or trajectory. These systems are very complex and require constant observation and adjustment—even more so at the LHC than at most other accelerators. Particles enter the LHC at a relatively low energy and then ramp up to the target collision energy. Substantial RF adjustments are needed throughout this ramp-up process.

The LHC team had originally planned to tune and configure their RF systems from a room in the LHC cavern 100 meters underground. But after a September 2008 magnet quench, CERN officials instituted a new safety policy: from then on, access to the cavern was very restricted once the magnets were cooled and, as always, forbidden when particles were being accelerated. Under this new policy, LHC accelerator physicists couldn’t tune and configure the RF system by hand as originally planned. They needed a Plan B.

Fortunately, the SLAC team was already working on a solution. For more than a year before the magnet quench, they had been collaborating with LHC RF physicists to model interactions between the RF and beam systems. As an offshoot of these discussions, the possibility of also writing software—similar to that used for SLAC's PEP-II collider—to remotely characterize and configure the RF system came up. When the regulations were changed at CERN, "suddenly this work was really necessary, not just a convenience. So we cranked up the effort," Mastoridis said.

"We developed some very advanced techniques to run RF systems at their limits during PEP-II," Fox added. "The technical heart of the RF systems at both PEP-II and the LHC is very similar because both systems are based off [the same] designs. That allowed us to apply knowledge gained here at SLAC to the LHC."

Funded by the Department of Energy through LARP, the U.S. LHC Accelerator Research Program, Van Winkle, Rivetta and Mastoridis began reconfiguring the SLAC RF algorithms and models for the LHC architecture and characteristics in 2007. Working closely with CERN collaborators Philippe Baudrenghien, Andy Butterworth and John Molendijk, the researchers sought to understand the likely behavior of the systems at their limits, and to anticipate possible configuration and operational options for high-current and high-energy operations.

Thanks to long hours, very hard work and many trips to Geneva, the SLAC team designed the software in time for LHC startup. Not long after, they received an unexpected e-mail from their CERN colloborators. "We were thrilled to get the note from Philippe telling us how well the tools worked, and how well the collaboration with CERN was going," said Van Winkle. "We’ve had quite a few e-mails over the years about RF systems, but this is our very first thank you note."

The software works by first monitoring the complex dynamics of the LHC's RF system and the beam. As a tightly packed bunch of particles travels through the accelerator, if everything is timed correctly, it surfs along the crest of an RF wave each time it enters an "RF cavity," gaining energy. Yet because the particles themselves are charged, each bunch changes the electric field in the cavities, muddying the RF wave. With more than 11,000 bunches traveling through each cavity every second at the LHC, the electric field gets very complex very quickly, affecting the shape, trajectory and acceleration of the bunches.

By measuring signals at various points in the RF system, the SLAC-designed software gathers enough data to characterize the system with a mathematical model. Based on that model, the software calculates the RF system parameters needed to minimize the changing interaction between the RF system and the particle beam, and to get the best performance for beam conditions.

"From PEP-II, we know that as you change the operational conditions of an experiment, you need to make continual adjustments to the RF system," said Mastoridis. "Together with our CERN collaborators, we have introduced ideas that could make the [LHC] system more easily configurable, closer to an optimal setting, and provide the option to continuously monitor the system."

The software also allows LHC accelerator physicists to adjust the RF system from afar while the machine is running and make any needed changes quicker than the previous, more manual method. For example, it took the CERN team four weeks to commission the RF cavities in 2008. This time around, using the SLAC-designed software, it took 34 hours.

In the coming months, the SLAC team will continue to refine the software. In addition, Mastoridis and Rivetta are developing a computer simulation for the interaction between the particle beam and the RF system's technical components. This simulation, an important part of Mastoridis' Stanford PhD thesis, will allow accelerator physicists to test any planned alterations to ensure they won't cause any detrimental side effects—before they ever touch the real accelerator. Such a system could also be useful for future upgrades, as it can help researchers understand how noise and imperfections add up at ever-higher beam energies, and also to understand the required specifications of new or expanded RF station architectures and control schemes.

"Through simulations we can make predictions on how various configurations or algorithms will affect performance in future operational conditions," Rivetta said. "We did this for PEP-II, and now we're doing this for the LHC. Hopefully these tools will be used there for a long time to come."

—Kelen Tuttle
  SLAC Today, December 10, 2009