After the technique of self-calibration (Pearson and Readhead 1984) was first introduced to radio astronomy, it was apparent that it might lift the ionospheric limitation on baseline length for low frequency interferometers. That barrier had restricted the aperture size of previous connected element low frequency synthesis telescopes to less than 5 km, thereby greatly restricting their angular resolution, and because of confusion, their sensitivity as well. Since then a number of instruments have operated successfully at low frequencies on baselines well beyond 5 km, including the Multi-Element Radio-Linked Interferometer Network (MERLIN, down to 150 MHz at baselines up to 217 km), the Very Large Array (VLA, down to 330 MHz at baselines up to 35 km), and most recently the Giant Metre Wavelength Radio Telescope (GMRT, down to 150 MHz at baselines up to ~25 km).
The first proposal to extend such techniques to frequencies below 100 MHz with a connected element, synthesis imaging array was made shortly after self-calibration was first introduced, when Rick Perley and Bill Erickson proposed development of a large, dipole-based array to work alongside the VLA in New Mexico. (G. Swarup's original concept for the GMRT also appeared at about this time.) Funding to implement the sensitive, broad-band, ambitious system originally envisaged in VLA Technical Memorandum #146 (Perley and Erickson 1984) was not readily available at that time. However their proposal inspired the Naval Research Laboratory (NRL) and National Radio Astronomy Observatory (NRAO) to work together in the early 1990s to implement a narrow-band, modest version of the Perley-Erickson proposal using the existing VLA dishes and infrastructure. (Note that some of the earliest VLBI work was done in the 1960s at frequencies as low as 18 MHz on baselines in excess of 100 km, e.g. Brown, Carr, & Block 1968a,b)
An initial 8-antenna 74 MHz system developed at the VLA was very successful, being the first connected-element imaging interferometer operating below 100 MHz to "break the ionospheric barrier" (Kassim et al. 1993). It successfully demonstrated that self-calibration could, at least to first order, remove ionospheric effects and permit imaging on baselines out to at least 35 km even at the very lowest frequencies where the ionospheric effects were largest. Its reliance on an over-determined problem in which antenna-based corrections to ionospheric phase distortions could be readily extracted from the self-calibration process worked well at the VLA.
The required antenna-based phase corrections were derived from simultaneously obtained 330 MHz data that utilized all 27 VLA antennas and possessed intrinsically much greater signal to noise. Using this trial system, several of the best known sources in the sky were resolved and imaged for the first time (Fig. 1), and a number of unique scientific results were extracted. At the same time as this system was being developed, solutions to key challenges common to all low frequency interferometer observations, such as RFI-excision and wide-field imaging, were being developed on the VLA 330 MHz system that NRL had also worked with NRAO to develop during the 1980s.
Based on this success, NRL obtained additional funding to build the receivers and to work with NRAO to extend the 74 MHz system to all 27 antennas of the VLA. When the techniques developed at 330 MHz, aided by the ongoing revolution in computational power, were applied to the completed 27-antenna 74 MHz data stream, it quickly demonstrated the ability to map thousands of sources, achieving a significant leap forward compared to past capabilities. In fact the full system made self-calibration so much more robust that the previous prerequisite for "phase-transfer" from simultaneous 330 MHz observations was no longer required. Today the 74 MHz VLA system is still the most powerful interferometer in the world working below 150 MHz (Fig. 2), and has attracted a wide variety of scientific projects in the areas of solar system (planetary emission, solar bursts), Galactic (supernova remnants, ISM), and extragalactic (clusters, radio galaxies) astrophysics. An ongoing major project is the VLA Low Frequency Sky survey (VLSS) a 74 MHz complement to the successful NVSS 20 cm VLA sky survey; the growing VLSS catalog now contains a list in excess of 65,000 sources.
The full 74 MHz VLA system has been available to the general scientific community since 1998, and has a growing international user community conducting unique observations in many different areas of astrophysics. Many of the technical innovations developed during the course of its maturity have also had tangible benefits for higher frequency observations, for example including the application of calibration, RFI excision, and wide-field imaging procedures to data from the GMRT and VLA at frequencies up to 1400 MHz. In 2002 a 74 MHz receiver was added to the Pie Town VLBA antenna, and successful images synthesized from baselines up to 73 km represented another major milestone in long wavelength radio astronomy (Fig. 3).
While the improvement in resolution afforded by the expansion to Pie Town is impressive, the restriction to a narrow band and the continued poor sensitivity relative to the higher frequency VLA and GMRT systems reveal that future steps forward require development of new instruments with much more collecting area and longer baselines. Nevertheless the 74 MHz VLA represents a major step forward in low frequency radio astronomy, and its success continues to play an important role in inspiring the development of the emerging much larger instruments such as LOFAR and the LWA.
Low-Frequency Observing and Data Reduction with the VLA