Nick Konidaris

Search this site

SED Machine


The SED Machine

Nick Konidaris, Robert Quimby, Chow Chong-Ngeow Sagi Ben-Ami, Richard Dekany, & Shri Kulkarni
NSF Award #1106171

No Transient Left Behind

SED Machine Rendering on a 2-m class telescope



Why the SED Machine?

The existence of our solar system is a result of the production of elements produced by cosmic explosions. One of astronomy's greatest triumphs is the story explaining the production of these elements, and "transients" play a central role. Due to advances in detector technology and computing, surveys designed to discover transients have popped up: Catalina Sky Survey, Palomar Transient Factory (PTF), SkyMapper and Pan-STARRS; as well as up-and-coming surveys: SASIR and LSST. The discovery power of these surveys is immense. The discovery rate for PTF exceeds one new transient every fifteen minutes,  but this rate is too high to systematically classify all transients. In short, because our ability to classify is so poor, we are in an era where the  systematic classification of transients is impossible.

The scale of this problem, however, is such that moderate aperture telescopes can play the critical role of classifying transients. To play this role, however, a new and highly efficient type of spectrograph is needed. Efficiency here is not limited to throughput, but rather designing an instrument and system that works together to produce the highest rate of classifications per square-centimeter of telescope. The SED Machine optimized specifically to maximize the transient classification rate.

Maximum Throughput: Instrument Design

Though the spectrograph itself is efficient, components of the SED Machine works together as a system to maximize transient classification efficiency.

  • Acquisition: In order to minimize acquisition time, the SED Machine has a wide field of view integral field spectrograph. It takes 3600 spectra simultaneously (e.g., TIGER of Bacon et al. 1995).
  • Science spectra: Science spectra are taken at the ideal resolution for transient classification, R~100. The spectrograph is designed to have extremly high throughput. Example spectra are shown below.
  • Calibration: To reduce the number of calibration spectra taken, we simultaneously (see figure above) take multi-band images (Bgri) using a four-color imager known as the rainbow camera. 
  • Data Reduction: A high-quality data reduction pipeline is being produced by the National Central University in Taiwan.

Laboratory Testing at Caltech

July 2011: SED Machine project begins. Laboratory testing starts at Caltech:
Sagi & Katie


Detector Testing in Israel

Detector testing is organize by our graduate student, Sagi Ben-Ami. Sagi has borrowed both an Apogee, as well as a Princeton Instruments PIXIS detector and has tested instruments on sky, as well as in the lab. Two example photos are shown below.