From: Adolphsen, Chris E. [star@slac.stanford.edu] Sent: Wednesday, August 19, 2009 3:05 PM To: ilc-ops-availability@desy.de Cc: Nantista, Christopher D. Subject: LLRF Configuration for Availibiltiy Below I outline a possible LLRF configuration for the Klystron Cluster scheme that ensures high availability at a low cost. For each cryomodule (CM), the probe signals from the 8 or 9 cavities are each dowmixed to baseband (I and Q) and digitized at 10 MHz in a 'CM Sum Unit' that is located near the CM in a radiation shielded enclosure. After calibration and filtering are applied, a digital sum of the cavity voltages is computed in the CM Sum Unit and sent via a fiber optic cable to the associated KlyClus building at a rate of 1 MHz for FB purposes. The raw digitized data for the probe rf and other signals, such as reflected rf, are sent between pulses for diagnostics purposes. If the CM Sum Unit fails to send this info, the cavities in that CM would be detuned and the power to this CM would be diverted to a load (both would take on the order of minutes). Note that there would be some failures, such as an individual digitizer, where the CM could continue to operate. At the KlyClus building, the fiber optic signal from each of the ~ 100 CMs would be split three ways and sent to three identical Drive Processors The output rf for each of the klystrons (36 or 29 - this includes one spare in each case) would also be digitized at 1 MHz and split three ways, one going to each Drive Processor. The electronics to do this would be separate for each source, and if it failed, at worst that klystron would be shut off (the klystrons run at saturation, and a few would be run off phase to adjust the pulse shape of the combined power). Others signals, such the klystron reflected power and power launched in the TE01 pipe, would be also digitized and used for monitoring and for slow FF control - however, the system could run without them on the time scale in which they could be repaired. Each Drive Processor would use the CM sum signals, the klystron output signals and dynamic FF tables to compute the digital drive signal to each of the klystrons. There would be one digital output per klystron. At each klystron, the three digital signals from the Drive Processors would go through 2-of-3 majority logic before being used to upmix a LO source that gets amplified (to 100 W) to drive the klystron. With this logic, if any one of the Drive Processor fails (both hard or soft), the results of the other two are used. The clock signals and LO for this system would be provided by redundant sources so that if one source fails, the signals would still provided by the others. Also, the AC power to the Drive Processors would be provided through different breakers. The fiber optic splitters could perhaps have battery backup. With this scheme, one may have to turn off individual CMs or klystrons due to LLRF failures, just as in the baseline design, but rarely the entire cluster (e.g., in the case when all three of the Drive Processors disagree for more than one klystron). For the DRFS scheme, I believe the FB control would be done locally in the tunnel for the cavities associated with each klystron. By design, this system is less prone to failures that turn off more than one source. However, a modulator or klystron failure also 'take downs' the associated cavities (2 or 4), and the klystron would have to run below saturation if operated independently of the others - if neighboring klystrons were paired, then they could run off crest in saturation.