INTRODUCTION

This aerial view shows some of the main parts of the NPOI, located on Lowell Observatory's dark sky site at Anderson Mesa approximately 15 miles southeast of Flagstaff, Arizona:
the Control Building, the Lab Building, the Astrometric Array which includes 4 stations, the Imaging Array with another 27 stations, and the Long Delay Lines. Inside the lab building, there is also the Fast Delay Lines, and the Inner Room where the beam combining is done.

Each arm of the array is 250 meters (820 feet) long. The longest baselines will be 437 meters (1433 feet) in length, while the shortest baseline is 17 meters (56 feet). The current ~1 milliarcsecond resolution will be increased to 200 microarcseconds when completed.

This tour of the NPOI shows how various parts of the instrument work and how all the parts work together to produce stellar interference patterns. The tour follows the starlight path through the telescope, beginning at a station on the array...

ARRAY STATION

There are two types of stations on the NPOI array - Imaging, used for observing programs that image stars, and Astrometric, used for accurate measurements of the positions of stars. There will be a total of six Imaging stations which can be moved to different locations on the array. There are four Astrometric stations at fixed positions.
Inside an astrometric station, the main components are the Siderostat, with a 50 cm mirror (20 inch), the WASA (Wide Angle Star Acquisition) CCD camera, and the NAT (Narrow Angle Tracking) tip-tilt mirror. The movable Imaging stations have the siderostat under one cover, and the NAT and WASA under another.
When starlight reflects off the Siderostat mirror, the star is imaged by the WASA camera which has a view of the siderostat mirror. The image is analyzed and information fed back to adjust the siderostat pointing so that the light is reflected to the NAT mirror and into the feed system...

THE FEED SYSTEM

The Feed System is a series of vacuum pipes and "cans" containing mirrors which route the light from the array stations back to the Lab Building where the light is recombined. Each arm of the array has three levels of pipe allowing up to three stations on any one arm, or any combination of up to six stations total from the three arms to be used at once.
At each station on the array there is an "Elevator" can. Light is reflected into the elevator can, from the NAT mirror, through a window in the top of the can. Inside the elevator can, there is a moveable platform to direct the light onto any of the three pipe levels, while letting light pass straight through from stations farther out the arm. At pipe intersections, at the branches to the astrometric stations and at the array center, there are "Feed" cans which direct the light beams around or through the intersections. Inside the feed cans, there are mirrors on slides on each level that will intercept a beam and reflect the light in a different direction or slide out of the way to let the light pass through.
The 255 cubic meters (9000 cubic feet) of pipe in the feed system, spanning the three 250 meter (820 feet) arms of the array, is all under vacuum so the light will travel at the same speed across the array. The distance the light has to travel to the beam combiner is different depending on which stations are being used. This fixed optical path difference has to be compensated for before the light can be recombined. The projection of the starlight onto the array, creates an additional variable optical path difference, that also has to be compensated for. This is done by the delay lines...

THE DELAY LINES

The NPOI has two types of "Delay" lines. The Long Delay Lines, or "LDLs", compensate for optical path differences in incremental amounts. The LDLs are 110 meter (361 feet) long vacuum tanks with cans, that each contain two "pop-up" mirrors, spaced at 6 intervals. Inside the lab building the beams are brought to the same level through periscope cans and then will be sent, twice, out and back a delay line, reflecting back off the pop-up mirrors. The mirrors may be popped up in different cans creating a wide variety of incremental delay distances up to 440 meters (1444 feet).
The Fast Delay Lines, or "FDLs", are 18 meter (59 feet) long vacuum tanks in the Lab building. Inside each tank is an optical cart, riding on parallel steel rails. Each starlight beam travels to the optical cart where it is reflected back by a "cats-eye" reflector. The optical cart, with it's position monitored by a laser metrology system, can travel at speeds of several centimeters (~ 1 inch) a second with a precision of a few nanometers (~ 0.0000001 inches). This allows continuous delays lengths up to 36 meters (118 feet), covering the intervals between LDL pop-ups, and compensating, while tracking a star, the changing optical path differences due to Earth's rotation.
At the front of the cart is a piezo mirror that strokes in and out, at 500 Hz. Setting the piezo mirror strokes to different amplitudes, between 500-8000nm, on different FDL lines allows the signal from each baseline (a pair of stations) to be determined. This is necessary because each output beam contains up to six different baselines after the light is combined...

BEAM COMBINATION

The Beam Combining Table, located in the Lab Building, holds the Narrow Angle Tracker (NAT) quad cells, which feed back error signals to the NAT mirrors keeping the beams centered through the feed system and aligned into the beam combiner, the Spectrometer prisms and lenslet arrays, which disperse the light from the beam combiner and then feeds the spectral components, covering 550-850nm, through 16 optical fibers to Avalanche Photo Diodes (APDs) where the light photons are converted to electrical signals, as well as the heart of the system - the Beam Combiner optics. The Beam Combiner consists of two 3-beam wide beam splitters (BSA and BSB), two 3-beam wide flat mirrors (M3A and M3B), a 2-beam wide flat mirror (M2), and one single-beam flat mirror (M1). The light beams enter the Beam Combiner, from up to six different array stations, impinging on either BSB or M3A first, and then propagate through so that each of the three exit beams are the combination of light from four stations, or six "baselines" (pairs of stations). Then each of the three exit beams passes through a Spectrometer prism and is focused onto one of the three Spectrometer lenslet arrays, where the spectral components are sent through optical fibers to banks of APDs where the photons are converted to electrical signals - and that is the end of the light path.

TELESCOPE CONTROL

From the NPOI control room, using the primarily GUI based software, the observer can control the siderostats pointing, star acquisition, stellar tracking and flux into the instrument, the weather and "seeing" conditions (estimated from NAT tracking errors). The observer can control the delay lines, initiate fringe searching, adjust tracking parameters and monitor fringe SNRs (Signal to Noise Ratios), as well as monitor data recording and log observations.
In addition to numerous computers in the control building, the electronics behind the software for the NATs, FDLs and Fringe engine are located in the Lab Building. The electronics for each siderostat are located on the array, at or near each station, and are in the process of being upgraded.
While the observer can manually control all the telescope systems, most of the observing processes are also automated so that an entire observation sequence can run by simply selecting the star and pressing one button. This allows one observer to operate the entire NPOI and, in as little as three minutes, complete each stellar scan...

HOW IT WORKS

When the light from a star shines down on the NPOI array, unless the star is directly overhead, the light travels farther to reach different stations and this delay has to be compensated for before combining the light from the different stations. To observe the star, first the optical carts in the fast delay lines slew to starting positions and the siderostats point to the star so that the light propagates through the telescope. Then the fringe engine takes over control of the optical carts in the Fast Delay Lines, moving them, until the optical path lengths match the delay distances and the fringes are found.