1. The ALMA Correlator Status The ALMA Correlator underwent its Preliminary Design Review on 2000-01-20. The official committee agreed that they were impressed with progress on the NRAO design and the thoroughness with which it has been carried out; the design meets the scientific specifications spelled out in the MMA White Paper "Astronomical Requirements for the Millimeter Array Correlator" by Rupen, Shepherd and Wright (1998). 1.1 Design The ALMA Correlator is being built to accommodate 64 antennas. The design may be changed to accommodate larger or smaller arrays until production units are begun (1.3 below). There are 8 baseband inputs per antenna, each with a maximum sampling rate of 4 GHz (2 GHz bandwidth) digitized at 3-bit 8-levels. The signal is digitized and transmitted over fiber optic cables to the correlator. Station cards provide bulk delays suitable for a 30 km range. At the digital FIR filter, bandwidths and fine delays are set using 4-bit quantization before being passed at 2-bit 4-levels through packetization to the 2-bit 4-level correlator. The resulting bandwidths per baseband input range from the 2 GHz maximum down to 31.25 MHz, providing 512 channels of 31.25 MHz resolution over 16 GHz at the lowest resolution broadest spectral range setting, and 15.3 kHz at the highest full polarization spectral resolution. For single polarization work, 1.9 kHz can be obtained. Details are given in ALMA Memo. No. 194. Power estimates are decreased 25% from the numbers in the project book, where further details may be found. The cost of the full ALMA correlator will be $13M. 1.2 Schedule The correlator continues to be on schedule, with working test fixtures and FIR filter card in 2000-12, working station card, control card and long term accumulator in 2001-05. The prototype correlator chip will be deliverd in 2001-07, followed by the critical design review on 2001-07-31. The minimally populated correlator will be complete by 2002-04, with a commitment to a production run of correlator chips in the summer of 2002. A single baseline correlator for the test array will be working in the laboratory by the end of 2002, to be delivered to the VLA site in 2003-05. The first quadrant correlator, fully capable for 32 antennas will be delivered to Chajnantor on 2004-06-16. 1.3 Options At the Correlator PDR, there was a unanimous recommendation that the change from 0.25 to 0.18 micron technology in the correlator chip be carried out, provided that the ongoing study of consequences resulted in confirming the conclusions which are now believed correct: lower power and lower overall cost. It was recommended that the decision to go to an 8K chip not be made until a detailed evaluation of the power dissipation of the 4K, 0.18 micron chip was known. Currently, the project remains committed to the 4K lag/chip .25 micron technology, on the basis of expected lower yield for the 8K lag/chip, .18 micron technology chip, its higher estimated power dissipation requirements, and the $1.8M higher cost for that technology. The major scientific gain expected for the 8K lag/chip, .18 micron technology, would be 15.625 kHz resolution over 1024 channels for the broadest 8 GHz full polarization mode (and more channels in other modes also). Furthermore, the foundries which can produce 0.18 micron chips are fully booked and prototype chips would not be available until 2002 it this technology were pursued now. If the Enhanced ALMA is made possible with Japan entering as a third major partner, the ALMA Correlator may be expanded to accommodate additional antennas. The cost for this will go as (N/64)^1.6 so, e.g. the cost of the correlator would increase 40% for 80 antennas rather than the current 64, with an additional design time of a few months.