SLAC is creating radically new modulators to power the next linear collider accelerator. Modulators convert wall-plug energy from PG&E into a high-voltage pulse that fuels the klystrons, which in turn propel particles down the linac.

The next linear collider project (NLC) aims to operate at 10 times the energy of SLAC’s accelerator. This will require a 20-mile accelerator and enormously better modulators and klystrons, which are the heart of the radio frequency (RF) system that accelerates particles to nearly the speed of light.

In collaboration with LLNL and Bechtel-Nevada, SLAC has designed a prototype solid state modulator that recently met two major milestones: driving four XL4 (50 megawatt X-Band) klystrons at full peak current and voltage, and creating the ideal shape pulse required for constant RF power. Richard Cassel (ESD) is the project leader at SLAC.

The Solid State Induction Modulator team beside the 8-Pack Modulator. Front row, left: Richard Cassel (ESD), Dan Moreno (ESD), Chris Pappas (ESD), Minh Nguyen (ESD). Rear row, left: Marc Larrus (ESD), Brad Hickman (LLNL), Ray Larsen (TD), Piotr Blum (ESD), Ed Cook (LLNL), Jeff de Lamare (ESD), Craig Brooksby (Bechtel) (Photo by Diana Rogers)

The old, shed-size modulators, which have served SLAC since its inception, use switch tubes that are now obsolescent and would last just over a year of full-time running in the NLC. When triggered, the switch tube sends a short high-voltage pulse from a pulse-forming network (PFN) to a transformer, where it exits at 350 kiloVolts (kV) to drive the klystron cathode. These modulators are sufficient for operating the roughly 240 klystrons on SLAC’s linac.

But the old design won’t work for NLC. Just the thought of procuring enough of the short-lifetime switch tubes for over 4,000 more powerful klystrons sent the R&D team to the drafting board.

"The old modulators are totally impractical for NLC, they’re not flexible and reliable enough. We’d worry whether we could keep the machine running," said Ray Larsen, NLC Program Manager. "They’re also expensive, and we challenged ourselves to build a more reliable, energy-efficient and cheaper design."

The new modulators put solid state switches on standard circuit boards. The prototype currently being tested at the lab’s NLC Test Accelerator (NLCTA) employs 76 boards stacked in a cabinet with a unique 3-turn, tri-axial transformer that steps up the stack’s voltage by a factor of three. Each board handles 2.2 kV, enabling the modulator to generate the 500 kV the NLC needs to feed its new 75 megawatt klystrons, which were also designed and built at SLAC.

"The prototype has a lot of beautiful features," said Larsen over the modulator’s buzzing noise. Each board handles relatively low voltage, so is easier to insulate. The solid state switches—Insulated Gate Bipolar Transistors (IGBTs)—turn on and then off to make the short pulse. The pulse length and shape are easily adjustable, a major advantage over the PFN design. The modulator is powerful enough to drive eight klystrons at a time. It runs even if a few boards aren’t working.

And, the prototype has successfully pumped four XL4 klystrons to provide the required RF peak power for testing the critical RF distribution and accelerator beamline components.

The modulator team’s next challenge is to achieve reliable 24/7 operation to support the NLC full-power demonstration program scheduled over the next one to two years. Meanwhile, the team has just manufactured the first test section of a new prototype geared to run two klystrons instead of eight, because the latest design calls for a more spread out RF system. The cylindrical ‘2-Pack’ modulator relies on 12 boards with fewer switches operating at 4.0 instead of 2.2 KV, and 11 turns in the transformer to generate the 500 kV.

The group continues to improve the IGBT switch to make it faster and better suited for short pulse work. Originally designed to drive electric trains in Europe, the switches will drive even faster objects wherever the new machine gets built.