SpUpNIC exposure time information

For users wishing to estimate exposure times from real SpUpNIC data, the following files give extracted 1D spectra from a representative subset of a suite of spectrophotometric standards observed with SpUpNIC during 2015 Nov-Dec.  The most reliable way to estimate exposure times is probably to measure S/N directly in the wavelength range of interest, with the setup which most closely matches yours, and then to scale accordingly.

Measurements are given for the two most commonly used gratings, G4 and G7, and at typical grating angles.  These conversion factors should be approximately unchanged for small changes in grating angle.  Other gratings will be added as the information becomes available.

Each _1D.txt file contains wavelength (A) and [sky subtracted] counts summed over the extraction window in the spatial direction.  This is the raw counts observed in the total exposure time.  Exposure times and other useful info are given in the NOTES file.  These files are the output from V2 (May 2016) of the quick look GUI and include automatic wavelength calibration.  These are not yet finely tuned, but should be accurate to better than ~2A, and thus fine for comparing with flux standards.

Also included are files containing the conversion factors to convert from **counts/second/pixel** into flux units in 10^-16 erg/s/cm^2/A.  So, in order to go from the observed counts in the _1D.txt files, it is necessary to divide by the exposure time and then multiply by the appropriate conversion factor.

All of these observations were made in “Faint & Slow” mode, which has a gain of 1.145e-/ADU.  No aperture correction has been made.

Filename Grating Grating Angle Object EXPTIME Comments


G4 4.1 LTT_377 600  2.1″ slit in 1.5″ seeing


G7 17.0 LTT_377 300  1.35″ slit in ? seeing


G7 15.3 EG21 300  2.1″ slit in 1.2″ seeing

A more detailed version of this table: NOTES


The observed S/N may be simply estimated in the wavelength range of interest by, for example in python:

x,y = np.loadtxt(‘a0071110_1D.txt’,unpack=True, usecols=(0,1)) # read in data

ok = np.where( (x>5100.) & (x<5200.) )[0] # choose wavelength range relatively free of absorption lines

print median(y[ok])/std(y[ok])


Therefore, for LTT 377 at ~5150A (flux ~1.2e-13 erg/s/cm^2/A) SpUpNIC reaches a S/N ~76 PER CCD PIXEL in 600s.


Flux calibration curves





The Hamuy standards used in the calibration can be found here:

ftp://ftp.eso.org/pub/stecf/standards/ctiostan/  NOTE: LTT377 is cd_34d241

Approximate polynomial fits to a series of standard stars, giving the factor by which to multiply the raw counts (per second), to get to flux, as described above.  The plots below show the stars used (and the variation from object-to-object) for G4 and G7 respectively.  Note: LTT1020 seems to be an outlier and has been rejected from the fits.  It may be necessary to interpolate the wavelengths onto a common grid before multiplying.


Quick python example:

x,y = np.loadtxt(‘a0071110_1D.txt’,unpack=True, usecols=(0,1)) # read in data

y = y / 600. # convert to counts/s

a,b = np.loadtxt(‘gr7_calfac_grang15.3.txt’,unpack=True, usecols=(0,1)) # read in appropriate flux conversion factor

# interpolate to same wavelength bins:

ib = np.interpol(x,a,b)

plot(x, y*ib*1e-16 ) # flux calibrated in erg/s/cm^2/A


Comparison with pre-upgrade spectrograph

For users preferring to scale their exposure times from previous experience, the following may be more useful.


Comparison of pre-upgrade (blue) vs SpUpNIC (green) in e-s per second **per A**.  Note: the dispersion has changed from 0.495 A/pix to 0.622 A/pix and this conversion factor has been included.  The old gain was 1e-/ADU.  To compare values per CCD pixel, SpUpNIC will be 0.622/0.495 times higher than plotted. To compare old and new counts, SpUpNIC will be 1.145 times lower (the new gain).


Relative throughput (ratio of the above). Not smoothed.



As above, but for G7.