The TMX-U data system is a general-purpose system for acquiring, analyzing, and output of TMX-U data, and can also apply to any pulsed controlled thermonuclear experiment. In its present implementation, it routinely acquires 3 Mbytes of data per shot at an average rate of one shot every 8 minutes. For increased throughput, the system uses 5 CPUs accessing shared disk memory with a capacity of about 500 Mbytes. All acquisition, analysis, and output of data is handled by a collection of standard program modules (processors), which can be linked together to form chained calculations. Processing for various shots may be allowed to overlap, with higher priority results being available quickly, while lower priority results being allowed to lag behind and catch up in natural lulls in the experiment. Selection of tasks (processor invokations) is done independently by each of the 5 CPUs in the system, and each task can run in any suitable CPU.

This paper describes a fast rise, low-impedance pulse generator that has been developed at the Lawrence Livermore National Laboratory. The design specifications of this generator are: 50-kV operating voltage, 1-ohm output impedance, subnanosecond rise time, and a 2 to 10 nanosecond pulse length. High repetition rate is not required. The design chosen is a parallel-plate, folded Blumlein generator. A tack switch is utilized for its simple construction and high performance. The primary diagnostic is a capacitive voltage divider with a B probe used to measure the current waveform.

A breeding ratio of at least 1.2 was a design goal for CRBRP. The value for the initial core (using plutonium with 11% /sup 240/Pu) calculated with ENDF/B-IV data is 1.27. Engineering mock-up studies for CRBRP were made in ZPPR-11. Analysis of ZPPR-11 using ENDF/B-IV data showed consistent underprediction of K/sub eff/ by about 1.5% and overpredictions of the /sup 238/U capture to /sup 239/Pu fission ratio (C8/F9) between 5% and 8%. These results are typical for all LMFBR critical assemblies at ANL. The following approach was used to determine the breeding ratio: sensitivity analysis of a range of fast reactor benchmarks and a fit to the experimental data by data adjustment; tests of the adjusted data against experiments in ZPPR-11; calculations for CRBRP with ENDF/B-IV data and the adjusted data to predict the breeding ratio bias; and estimates of k/sub eff/ and breeding ratio uncertainties using data sensitivities for CRBRP.