We describe the operation and commissioning of the Jefferson Lab UV FEL using a CW SRF ERL driver. Based on the same 135 MeV linear accelerator as the Jefferson Lab 10 kW IR Upgrade FEL, the UV driver ERL uses a bypass geometry to provide transverse phase space control, bunch length compression, and nonlinear aberration compensation necessitating a unique set of commissioning and operational procedures. Additionally, a novel technique to initiate lasing is described. To meet these constraints and accommodate a challenging installation schedule, we adopted a staged commissioning plan with alternating installation and operation periods. This report addresses these issues and presents operational results from on-going beam operations.

In this contribution we describe various systems and aspects of the electron beam diagnostics of the JLab UV FEL. The FEL is installed on a new bypass beam line at the existing 10 kW IR Upgrade FEL. Here, we describe a set of the following systems. A combination of OTR and phosphor viewers is used for measurements of the transverse beam profile, transverse emittance, and Twiss parameters. This system is also used for alignment of the optical cavity of the UV oscillator and to ensure the overlap between the electron beam and optical mode in the FEL wiggler. A system of beam position monitors equipped with log-amp based BPM electronics. Bunch length on the order of 120 fs RMS is measured with the help of a modified Martin-Puplett interferometer. The longitudinal transfer function measurement system is used to set up bunch compression in an optimal way, such that the LINAC RF curvature is compensated using only higher order magnetic elements of the beam transport. This set of diagnostic systems made a significant contribution in achieving first lasing of the FEL after only about 60 hours of beam operation.

We report on the performance of a high gain UV FEL oscillator operating on an energy recovery linac at Jefferson Lab. The high brightness of the electron beam leads to both gain and efficiency that cannot be reconciled with a one-dimensional model. Three-dimensional simulations do predict the performance with reasonable precision. Gain in excess of 100% per pass and an efficiency close to 1/2NW, where NW is the number of wiggler periods, is seen. The laser mirror tuning curves currently permit operation in the wavelength range of 438 to 362 nm. Another mirror set allows operation at longer wavelengths in the red with even higher gain and efficiency.

Many novel applications in photon sciences require very high brightness and/or short pulses in the vacuum ultra-violet (VUV). Jefferson Lab has commissioned a UV oscillator with high gain and has transported the third harmonic of the UV to a user lab. The experimental performance of the UV FEL is much better than simulated performance in both gain and efficiency. This success is important for efforts to push towards higher gain FELs at short wavelengths where mirrors absorb strongly. We will report on efforts to characterize the UV laser and the VUV coherent harmonics as well as designs to lase directly in the VUV wavelength range.

We discuss the use of multipass recirculation and energy recovery in CW SRF drivers for short wavelength FELs. Benefits include cost management (through reduced system footprint, required RF and SRF hardware, and associated infrastructure - including high power beam dumps and cryogenic systems), ease in radiation control (low drive beam exhaust energy), ability to accelerate and deliver multiple beams of differing energy to multiple FELs, and opportunity for seamless integration of multistage bunch length compression into the longitudinal matching scenario. Issues include all those associated with ERLs compounded by the challenge of generating and preserving the CW electron drive beam brightness required by short wavelength FELs. We thus consider the impact of space charge, BBU and other environmental wakes and impedances, ISR and CSR, potential for microbunching, intra-beam and beam-residual gas scattering, ion effects, RF transients, and halo, as well as the effect of traditional design, fabrication, installation and operational errors (lattice aberrations, alignment, powering, field quality). Context for the discussion is provided by JLAMP, the proposed VUV/X-ray upgrade to the existing Jefferson Lab FEL.

When operating JLab high current ERL a strong reduction of the FEL efficiency was observed with the increase of the average current of the electron beam. Investigating the FEL efficiency drop-off with the electron beam average current we have measured the electron beam phase noise and the fast energy modulations. The phase noise is a variation of the time arrival of the electron bunches to the wiggler. It could be a very effective way of reducing the FEL efficiency especially when the driver accelerator for the FEL is operated with the RMS bunch length of about 150 fs. Under a fast energy modulation we denote a modulation which can not be followed by the FEL due to its time constant, defined by the net FEL gain. Such a modulation also could be a possible cause of the efficiency drop-off. Making the measurements we could rule out the FEL efficiency drop-off due either the fast energy modulation or the phase modulation. We also have learned a lot about instrumentation and techniques necessary for this kind of beam study. ELECTRON BEAM PHASE NOISE MEASUREMENTS Investigating the FEL efficiency drop-off with the electron beam average current we have measured the electron beam phase noise and the fast energy modulations. The so-called phase noise is essentially a variation of the time arrival of the electron bunches to the wiggler. That could be a very effective way of reducing the FEL efficiency if one takes in to account that the accelerator is routinely operated with the RMS bunch length of about 150 fs [1]. Under a fast energy modulation we denote a modulation which can not be followed by the FEL due to its time constant, defined by the net FEL gain. Such a modulation also could be a possible cause of the efficiency drop-off. The two effects are strongly connected in the FEL driver accelerator due to the longitudinal phase space transformation, i.e., longitudinal bunch compression. The simplified view of the longitudinal phase space transformation is a rotation of a long and low energy spread beam at the injector by ~90 degrees in the longitudinal phase space so that the bunch length minimum is located in the wiggler [2]. Under such a transformation an energy modulation in the injector would get transferred in to a phase modulation at the wiggler and a phase modulation in the injector would gets transferred in to an energy modulation at the wiggler. The technique we use for the phase noise characterization of the electron beam was originally Figure 1 a: Single sideband spectrum

An Ethernet based embedded system has been developed to upgrade the Beam Viewer and Beam Position Monitor (BPM) systems within the free-electron laser (FEL) project at Jefferson Lab. The embedded microcontroller was mounted on the front-end I/O cards with software packages such as Experimental Physics and Industrial Control System (EPICS) and Real Time Executive for Multiprocessor System (RTEMS) running as an Input/Output Controller (IOC). By cross compiling with the EPICS, the RTEMS kernel, IOC device supports, and databases all of these can be downloaded into the microcontroller. The first version of the BPM electronics based on the embedded controller was built and is currently running in our FEL system. The new version of BPM that will use a Single Board IOC (SBIOC), which integrates with an Field Programming Gate Array (FPGA) and a ColdFire embedded microcontroller, is presently under development. The new system has the features of a low cost IOC, an open source real-time operating system, plug&play-like ease of installation and flexibility, and provides a much more localized solution.

The Free Electron Laser Project at Jefferson Lab is based on a comparatively small accelerator driver. As it's systems continues to grow and evolve, strict configuration control has not been a programmatic goal. Conversely, as the IR-Demo FEL and the 10kW IR FEL have been built and operated, hardware and software changes have been regular part of the machine development process. With relatively small component counts for sub-systems, changes occur without requiring much formal documentation and in-situ alterations are common-place in the name of supporting operations. This paper presents an overview of the web-based software tool called DevLore which was first developed to be a library for embedded programming and then became a tremendously effective tool for tracking all changes made to the machine hardware and software. This tool is maturing into an easy-to-access, common point-of access for the status of all of the installed system components, technical drawings, embedded firmware, procedures, operations notes, maintenance history, inventory tracking and computer-files associated with the FEL Project.

A high current beam transmission test has been successfully completed at the JLAB FEL Facility, culminating in very low-loss transmission of a high current CW beam through a small aperture. The purpose of this test was to determine if an ERL is capable of meeting the stringent requirements imposed by the use of a 1018/cm3 internal gas target proposed for the DarkLight experiment*. Minimal beamline modifications were made to create a machine configuration that is substantially different from those used in routine UV or IR FEL operation. A sustained (8 hour) high power beam run was performed, with clean transmission through a 2 mm transverse aperture of 127 mm length simulating the target configuration. A beam size of 50 um (rms) was measured near the center of the aperture. Experimental data from a week-long test run consistently exhibited beam loss of only a few ppm on the aperture while running 4.5 mA current at 100 MeV -- or nearly 0.5 MW beam power. This surpassed the users? expectation and demonstrated a unique capability of an ERL for this type of experiments. This report presents a summary of the experiment, a brief overview of our activities, and outlines future plans.