The Voyager Ultraviolet Spectrometers



1.    Instrument Description

The Voyager 1 and 2 Ultraviolet Spectrometers are nearly identical instruments. This discussion applies to both, except in a few instances in which important differences between the two are noted explicitly. The Voyager 1 Ultraviolet Spectrometer (UVS) is a compact Wadsworth mounted objective grating spectrometer that covers the wavelength range of 53.5 to 170.2 nm (51.3 to 168.0 nm for the Voyager 2 UVS). It records the entire spectrum within this range in a single exposure. It has no moving parts. A mechanical 'collimator' consisting of a series of 13 aperture plates defines the main 'airglow' field of view (FOV) of 0.10° full-width-half-maximum (FWHM) × 0.87° in length. Light passing through the collimator strikes the concave diffraction°grating at near normal incidence. The grating disperses and focuses light onto a 1-d array detector that records individual photoevents. An auxiliary field of view for solar occultation experiments is offset 20° from the airglow field by a small mirror near the front of the collimator. Using this 'occultation port', the UVS can view the Sun without pointing the main field, and those of other coaligned instruments, directly at the Sun. The occultation field is 0.25° FWHM × 0.87°. A sunshade prevents illumination of the main entrance aperture by the sun during occultation observations.


Voyager UVS Physical Characteristics


4.52 kg


43.2 cm


14.8 cm


17.2 cm

Power (HV off)

2.8 W

Power (HV on)

3.2 W


The UVS has two spectral resolutions depending on the nature of the source. An extended monochromatic source that fills the FOV ideally produces a triangular intensity distribution of 0.1° FWHM. (The actual response function is slightly rounded at the top and base, but a triangle is a satisfactory approximation for most applications.) The 0.1° corresponds to width of 3.5 anodes, or 3.3 nm. This inherent spectral resolution may often be improved by spectral analysis. A monochromatic point source is imaged onto a width of about 1 anode for a practical resolution of about 1.5 nm. Precise measurements of the relative response as a function of position within the FOV have been made by rastering the FOV across a star. At wavelengths longward of 135 nm there is a slight (~10%) asymmetry in the response on either side of the center of the FOV.


The effective sizes of the entrance apertures are (airglow port) 21.2 and (occultation port) 0.75 cm2.


The anode array is scanned at a rate of 3125 scans per second and the results are added into an internal memory. The UVS transmits the contents of this memory to the flight data system (FDS) on command of the FDS. The FDS retrieves values for a pair of channels each 5 msec, and so reads a complete spectrum from the UVS in 0.32 sec. For the fastest transmitted data rate (OC-1, see below) used for occultation observations, this readout proceeds continuously, producing a series of spectra separated by 0.32 sec. For lower data rates, the memory is read in bursts of 0.32 sec separated by the appropriate intervals. During these intervals, the UVS integrates the spectrum in its internal memory. As the data is transferred to the FDS it is logarithmically compressed from 16 to 10 bits.


The FDS determines the rate at which spectra are read from the UVS after being integrated internally in the instrument memory. Most planetary observations are made at one of two data rates, OC-1 (0.32 sec spectra, for occultation measurements) and GS-3 (3.84 sec spectra, for emission spectroscopy). Slower rates are used from time to time. Rates and their designations are:






Time (sec)




























A description of the UVS investigation is given by Broadfoot et al. (1977). Performance and analysis techniques are described by Broadfoot et al. (1981).

2.    Scientific Objectives

The primary goal of the UVS is to study the composition and structure of the atmospheres of the outer planets and their satellites. Secondary goals include the study of magnetospheric particle populations, magnetosphere-atmosphere interactions, the composition and distribution of the interplanetary wind, determinations of the solar flux, and stellar astronomy.

3.    Operational Considerations

The Voyager UVS instruments have operated nearly continuously since launch in 1977. With the singular exception of a decrease in the Voyager 1 microchannel plates (MCP) gain, due to a high radiation-induced count rate during passage through the inner Jovian magnetosphere, both instruments have remained photometrically stable at a better than 3% level since 1977. In-flight performance of the UVS from launch through the 1979 Jupiter encounters is reviewed in Broadfoot et al. (1981), and a description of astronomical observations is contained in Holberg (1990) and Linick and Holberg (1991).

4.    Calibration Description

Laboratory calibration of the UVS included measurements of:


1) sensitivity at a number of wavelengths throughout the spectral range and the positions of those wavelengths in channels,

2) response to scattered light,

3) off-axis response, including collimator transmission function, and

4) intrinsic dark count rate.


The wavelengths corresponding to each channel of the UVS may be calculated from


w = 9.26c + 525.57, (Voyager 1)


w = 9.26c + 504.19, (Voyager 2)


where c is the channel number. For integral c, the value of w corresponds to the short-wavelength edge of the channel. For example, the wavelength of the short-wavelength edge of Voyager 1's channel 76 is 1229.33 Angstroms (9.26*76.0 + 525.57). To find the central wavelength of a channel, add 0.5 to the channel number. For example, the central wavelength of Voyager 1's channel 21 is 724.66 Angstroms (9.26*21.5 + 525.57).


In-flight calibration has included assessments of absolute sensitivity by comparisons with stars, and measurements of the FOV response profile using stars.


Before the absolute calibration can be applied to a measured spectrum, three or four spectral analysis steps are required. These are flat field correction, dark count subtraction, and descattering, and (sometimes) sky background removal.


Channel-to-channel variations in sensitivity result from variations in effective count threshold among channels. Applying a 'fixed pattern noise' (FPN) correction adjusts the signal levels to their equivalents for a common threshold in all channels. The FPN correction involves a channel-by-channel multiplication by a correction spectrum. The correction spectra are different for Voyager 1 and Voyager 2. In fact, two spectra are in use for Voyager 1. The first is used for data acquired before Jupiter encounter and the second for data after encounter. The two differ to account for changes in the response of the Voyager 1 UVS induced by its operation in the intense Jovian radiation environment.


Channels 3 and 4 have large FPN corrections, i.e. they are less sensitive than the others. Therefore the statistical accuracy of the signal in these channels is lower than in the other channels.

4.1.           Dark Counts

In interplanetary space, detector dark counts arise mainly from effects of gamma radiation from the radioisotope thermoelectric generators that power the spacecraft. The count rate is about 0.02 counts per channel per second. The shape is approximately flat in wavelength, with a step near the edge of the filter in the detector. The shape is accurately known from long observations of the calibration plate mounted on the spacecraft. Almost no photon signal (except for a weak reflection of sky-background H Lyman alpha) is recorded during these observations. The absolute level varies slightly with scan platform position, because rotating the scan platform changes the shielding mass between the UVS detector and the generators. For data acquired outside a planetary magnetosphere, subtracting a scaled dark spectrum from a calibration plate observation is usually a satisfactory correction for dark counting.


Within a planetary magnetosphere, the dark count rate can include contributions from high-energy particles. For lower levels, scaling a calibration plate spectrum is again satisfactory, but for higher levels the shape of the dark spectrum changes and another method must be used to estimate dark levels. The best alternative is to use a spectrum taken at nearly the same time with no significant source in the FOV. Satellite observations often fill this need.

4.2.           Descattering

Light scattered within the instrument illuminates channels outside the ideal transmission function of the collimator. The effects of scattering are removed by a process called descattering. Descattering is accomplished through the use of a matrix operator, a 126×126 element matrix which describes the response of detector channel 'j' to the measured signal at channel 'i'. This scattering matrix is completely empirical, having been determined from laboratory measurements of 50 individual emission lines covering the entire passband. Descattering is a linear operation. Dark counts must be subtracted prior to descattering as the descattering algorithm is based on the assumption that only photon events are present. Descattering will also correct for second order response. Therefore, if the spectrum to be descattered contains artifacts, such as anomalously high or low counts in channels 3 or 4, a corresponding error will be introduced in the vicinity of both the first and second order positions.

4.3.           Sky Background Subtraction

When the UVS slit is not completely filled by the disk of a planet, the portion off the planet sees the sky background. Fortunately, in the far UV the sky is generally quite dark and diffuse starlight is seldom significant. However, bright emissions at H Lyman alpha, Lyman beta, and He 58.4 nm from the interplanetary medium (IPM) often must be taken into account. These lines arise from strong solar chromospheric emission lines that are scattered from neutral H and He of the local interstellar medium. The physics of this 'interstellar wind' is complex and leads to emission which is inhomogeneous in space and variable in time. The IPM responds to active regions of the solar chromosphere as the sun rotates. This means that the sky brightness as seen by the UVS can change noticeably on time scale of days. As with instrumental dark counts, there are two standard means of removing sky background: direct subtraction of an adjacent sky background suitably scaled, if available, and construction of a synthetic sky background spectrum.

4.4.           Calibration

Calibrating the spectra converts them from count units to absolute brightness units. This step has not been included in the data processing because the correct procedure depends on the type of source viewed. The data spectra represent count rates after correction for fixed pattern noise, a background subtraction, and descattering. Multiplying these spectra by one of the calibration spectra converts it to brightness units. There is a calibration spectrum corresponding to each of the two source types, namely point sources and extended sources.


Point Source: Multiplying a data spectrum by the calibration spectrum VxPTCAL.TAB (x=1 for Voyager 1 and x=2 for Voyager 2) converts the spectrum from counts/(channel) to photons/(cm**2-Angstrom-time), where time is the integration time of the spectrum.


Extended Source, continuum emission: Multiplying a data spectrum that has been normalized to an integration time of 1 second by the calibration spectrum VxFLCAL.TAB (x=1 for Voyager 1 and 2 for Voyager 2) converts the spectrum from counts/(channel-second) to spectral brightness in units of Rayleighs/Angstrom for a source that fills the field of view.


Extended Source, monochromatic emission: The finite spectral resolution (about 35 A) of the spectrograph must be considered in this case. For isolated lines (those that are not strongly blended with emissions at nearby wavelengths) it is sufficient to sum the channels that include light from the emission of interest (about 7 channels) and multiply by the appropriate calibration factor. This factor is the product of the dispersion (9.26 Angstroms/channel) and the value in VxFLCAL.TAB corresponding to the center channel of the wavelength of interest. The resulting quantity is the brightness of a monochromatic emission that fills the field of view. For more complex spectra that include blended emissions, the most accurate approach is spectral analysis by generating synthetic spectra. This technique uses an iterative approach to adjust an estimated brightness spectrum until the model spectrum computed from it matches the observed spectrum. The model can be fairly simple, but must include the triangular transmission profile of the collimator and the instrument sensitivity (calibration). The calibration factor described earlier in this paragraph is the correct one to use for this kind of synthesis.


Calibration and spectral analysis issues are discussed by Holberg et al. (1982) and Holberg (1986).

4.5.           Processing

These data have been processed by first applying the fixed pattern noise correction (flat field correction), and then correcting for internal scattering. Finally, background noise has been subtracted, by interpolating in a table of background noise versus time. For the reasons described above, we have not subtracted a sky background nor applied the absolute calibration. The values reported in the RDR in the positions CHANNEL_01 to CHANNEL_126 are 'counts' per INTEGRATION_TIME. The 'counts' may be non-integral because the fixed pattern noise correction, scattering correction, and dark subtraction are inherently floating-point processes.

5.    Platform Mounting Descriptions

The UVS is mounted on the scan platform. The instrument is approximately bore-sighted with the wide and narrow angle television cameras, and with the PPS and IRIS instruments. The alignment of the fields is not perfect; the following table gives offsets of the centers of the UVS fields relative to the centers of the ISS Narrow Angle Camera fields of view. Elevation is positive to the right within the imaging field of view, and cross-elevation is positive downward. The narrow axis of the UVS slit is aligned with the elevation direction.





Voyager 1




+18.9 pixels

-56.6 pixels

Voyager 2




0.0 pixels

+150.9 pixels



The RDR reports positions of three points within the airglow field of view. They are designated as P2, P5, and P8. The meanings are

P2  The center (in width) of the low-azimuth end of the slit.

P5  The center (in length and in width) of the field of view.

P8  The center (in width) of the high-azimuth end of the slit.


Field-of-view parameters for which no P-value is specifically mentioned refer to the P5 point.


The occultation field is offset from the airglow field by a small mirror. The offset is toward lower elevation. The elevation offsets are

Voyager 1:   -19.53°

Voyager 2:   -19.296°

6.    Instrument Detector

The windowless, photoevent-counting detector consists of an electron multiplier, a pair of microchannel plates (MCP) in series, and a 128-element linear self-scanned readout array. Photoelectrons ejected at the front of the MCP stack are amplified by a factor of about 1E6, and the resulting charge pulse is collected by the anode array. The 128 narrow aluminum anodes, each 3 mm long, are deposited on 0.1-mm centers for a collecting length (in the dispersion direction) of 13 mm.


The anodes are accessed sequentially by a shift register and FET switches contained on the single integrated circuit. The scanning circuitry discharges each anode into a charge sensitive preamplifier. The charge pulse is digitized and the information added into a shift register memory consisting of 128 16-bit words. The 128-anode array consists of two separate interdigitated 64-anode arrays scanned by two shift registers. The shift registers and memory are driven by a 200 kHz clock, so that an individual anode is accessed every 320 microsecond. The detector scan rate is therefore about 3125 Hz.


Wavelengths shorter than about 125 nm strike the MCP directly. Longer wavelengths first pass through a MgF2 filter with a semi-transparent photocathode of CuI. This serves to boost the quantum efficiency at long wavelengths and to reduce the response to second-order light.


The detector is heavily shielded to reduce its response to trapped particle radiation. A description of the detector may be found in Broadfoot and Sandel (1977).

7.    Instrument Electronics

The UVS electronics is housed in an enclosure integral with the optical section of instrument. Most of the electronics is in the base of the instrument, but clock drive generators for the anode array and the first stage of charge sensitive preamplification of the analog signal processing electronics are mounted in the detector housing so that they are near the anode array. Elements of the electronics complement include:


(1)  Low voltage power supply

(2)  High voltage power supply

(3)  Clock drive generator for the anode array

(4)  Analog signal processing electronics including A/D conversion

(5)  128×16 bit accumulation memory for spectrum

(6)  FDS interface


The FDS interface sends data to the FDS on demand and accepts mode commands from the FDS. The mode commands set the level of the high voltage applied to the MCPs of the detector and set the mode of analog signal processing (pulse counting or integration).


Radiation-hard electronics components were used where possible, and spot radiation shielding was used to reduce the fluence at certain critical elements.

8.    Instrument Optics

The optical system consists mainly of the mechanical collimator and concave diffraction grating. The 13 aperture plates of the collimator establish a field of 0.1°×0.87° for the airglow field and 0.25°×0.87° for the occultation field. The 0.1° and 0.25° dimensions are in the dispersion direction, and the 0.87° dimension is in the cross-dispersion direction. The collimator provides separate light paths for the airglow and occultation ports, and a small mirror diverts the occultation field by 20° from the airglow field.


The concave diffraction grating is a platinum coated replica, ruled at 540 lines/mm, blazed at 80 nm and having a spherical radius of curvature of 400.1 mm. Dispersion in the image plane is 9.26 nm/mm, or 0.926 nm/channel. The grating substrate is a 4×6-cm rectangle, and the useful ruled area is 21 square cm.


Telescope Diameter : 6 cm

Telescope f/ ratio : 4

Telescope Focal Length : 20 cm

9.    Instrument Mode 'Pulse Counting'

Two modes of electrical operation allow the detector to operate in a photon-counting mode for low source intensities, or in an integration mode for high source intensities. In the pulse counting mode, the number in the corresponding memory location is incremented by one if a charge above a fixed threshold is detected on an anode. The access time of 320 microsec implies that single random photoevents can be recorded on any one of the anodes at a rate of about 300 Hz with a coincidence of 10%. The pulse-counting mode is used for all measurements except solar occultations.

10.           Instrument Mode 'Pulse Integration'

In the pulse integration mode a 3-bit A-to-D converter is introduced ahead of the adder. In this case the charge on each anode is coarsely digitized and added to the previously accumulated signal in memory. The statistics of sampling these events is complicated by the logarithmic pulse height distribution of the events. There is also a logarithmic current limit function of the MCPs at the high event rates that normally obtain when this mode is used. Because both these characteristics lead to non-linear response, modeling of the detector response is needed to restore linearity. The integration mode is used for observing solar occultations.

11.           High-Voyager Power Supply Settings

The gain of the electron multiplier can be adjusted by setting the potential drop across the microchannel plates. The high-voltage level is commanded by the FDS. Levels 4-7 were intended to compensate for losses in supply output that could have occurred owing to radiation damage. They are not normally used for observations.


UVS High Voltage Levels






solar occultation


solar occultation


airglow, solar occultation


not used


not used


not used


not used