Figure 1. Laser Retroreflector
Feedback
Form for this paper
Author Resume
International Journal of Small Satellite Engineering - received
11 October 1995
email: M.Unwin@surrey.ac.uk
Satellite laser ranging (SLR) is a technique using a pulsed laser and a telescope to measure the range of a satellite from the ground with an achievable resolution of millimetres. Through fitting ranging measurements made by multiple SLR stations, the entire orbit of some low Earth orbiting satellites can be determined with an accuracy of centimetres. This highly precise orbit determination capability can be used to tune the models of the forces acting on the satellite in orbit to minimise the discrepancies between the orbit model and the measured data. The model of Earth's gravitational field can be improved by tracking satellites in different orbits that are affected by different gravitational harmonics. A better knowledge of the Earth's gravitational components can assist in geophysics, geodesy and oceanography. Furthermore, temporal changes due to tides, seasonal variations, tectonic plate motion and global warming can be investigated through the precise tracking of satellites over time. SLR and precise orbit determination are invaluable tools for understanding the physics of the Earth.
Several satellites dedicated to SLR have been launched into different orbits, these include GFZ-1, Stella [1] and Lageos. A dedicated SLR satellite is a completely passive satellite surrounded by retroreflector mirrors so that light is reflected back irrespective of the satellite attitude. The satellite has a high mass and is spherical so that the atmospheric drag and solar radiation forces are minimised with respect to gravitational accelerations.
Other satellites which are not completely dedicated to SLR, e.g. ERS-1 and TOPEX/Poseidon, carry small arrays of retroreflectors. This permits precise tracking to support on-board remote sensing experiments, such as altimeters or synthetic aperture radars. Precise tracking information can also be used to support technology development. An example of this is the mounting of retroreflectors on the GPS satellites to assist the calibration of the system in addition to gravity modelling [2].
Although dedicated laser tracking satellites are perhaps the most technically simple and inexpensive satellites orbiting the Earth, there is still a significant cost involved in the launch procurement and practicalities. It is common for such satellites to be launched as auxiliary passengers on other launches, but there is still the expense of the design and implementation of the deployment mechanism.
A number of small satellites, and in particular microsatellites (10-100 kg), are launched every year for research, remote sensing and communications purposes. For example, there have been 11 UoSAT-class microsatellites placed in orbit in the last 11 years [3]. Such satellites are designed and built rapidly, are low cost, and make use of existing launch opportunities to be deployed in various low Earth orbits. These platforms may offer an opportunity for a very low cost way to fly retroreflectors on existing missions. Small satellites rarely have deployed solar arrays, and so have a lower surface area to mass ratio than the large remote sensing satellites, making the orbit more easy to model. The attitude is usually controlled to within a few degrees, and so the number of reflectors required can be minimised. The size of the satellite means that the reflectors are inherently close to the centre of gravity.
From the view-point of satellite manufacture, the laser reflectors are easy to accommodate, so long as there is enough physical space for the configuration. No power is required, and the temperature and radiation effects need little or no consideration. The precise tracking can be used to support on-board experiments and technologies in addition to the benefit to be gained from the SLR community.
This paper describes as a case study the investigation into the integration of laser reflectors into the first Chilean satellite, FASat-Alfa microsatellite. The reflectors were not eventually flown on FASat-Alfa due to the low priority of the experiment, but the study has general application to the use of reflectors on other similar missions. Part I describes the general design study, and Part II addresses the specific problem of damage to the sensors from laser power.
The scheduled orbit for the FASat-Alfa mission was circular at 82.5° inclination and an altitude of 650 km. This orbit was of interest to the international SLR community and so a number of spare retroreflectors from the Stella mission were made available by CNES in support of the FASat experiment. The laser reflector experiment on FASat-Alfa had three main objectives.
In the design study, the laser reflectors comprised a relatively simple payload. Nevertheless, there were issues of concern:
The laser reflectors are approximately 3.5 cm cylinders of glass containing a 3-surface corner mirror. The principle of operation is that a light ray that reaches the reflector will always be reflected back the same direction, independent of the angle of incidence (Figure 1).
Figure 1. Laser Retroreflector
A single retroreflector has a field of view within which it can reflect most of the power back to the transmitter. To increase the tracking capabilities to lower elevations, multiple reflectors can be mounted at slightly different angles. SLR satellites are completely spherical and covered in reflectors (Stella has 60 reflectors), so that there is always a reflection no matter what the elevation of the satellite above the horizon or its attitude. The FoV (Field of View) of these SLR satellites is therefore 360°, or omni-directional. A satellite with limited room for reflectors, however, must make a trade-off between minimising the number of reflectors and obtaining a good FoV.
Figure 2. Satellite Laser Ranging Operation
The Satellite Laser Ranging (SLR) Station generates a very short pulse of light which travels to the satellite and is reflected back (Figure 2). The transit time can be measured to a resolution of less than 20 picoseconds, which corresponds to a precision of 3 mm. The data from a series of such measurements are combined together and fitted to an orbit model in order to determine the orbit.
Before the SLR station can track the satellite, it must first acquire using a real-time orbit tracking model. This acquisition can be difficult, because NORAD elements, which are often used, are sometimes wrong by several kilometres. Typically, the divergence of the laser beam will give a width of only a few hundred metres at the satellite's altitude. This width would be most narrow when the satellite is passing directly overhead. Therefore, longer passes are more likely to give useful data.
The mounting and the size of the reflectors is shown in Figures 3 and 4.
Figure 3. Retroreflector Mounting (CNES)
The mechanical housing for the retroreflectors is relatively simple and has been specified clearly in the drawings from CNES. The main mechanical requirement is that there must not be a temperature gradient of greater than 3° across the glass.
Figure 4. Seat for Retroreflector (CNES)
The reflectors can be mounted in individual blocks, with the seating tilted to an appropriate angle. These reflector blocks can then in turn be mounted onto the satellite.
For the reflectors under consideration, the reflected power falls to 40% level about 25° from the bore-sight, giving a total field of view of 50%. Therefore, using a single nadir-pointing reflector on a satellite, an SLR station could only track the satellite if it was at an elevation of over 65°, an occurrence that only happens rarely for a 650 kilometre orbit.
The ideal placement of multiple reflectors is on a spherical section with the centre of the sphere (common reflection origin) coincident with the centre of mass of the satellite. A common reflection origin is necessary so that as the satellite rotates, the measured range is consistent from one reflector to the next. Unfortunately the mechanical constraints of microsatellites may prohibit the reflection origin from being coincident with the centre of mass.
An example configuration of reflectors is illustrated in Figure 5. To increase the coverage, a realistic implementation would be to mount four reflectors on the four edges or corners of the -Z-facet (Earth pointing). The reflectors can be tilted from the vertical at an appropriate angle to optimise tracking. A hemisphere with the common reflection origin at the centre of mass dictates that the tilt angle be 15° from nadir. A considerable improvement to the FoV could be achieved, however, by tilting the reflectors to 30°. If the reflectors are tilted further than 30°, then gaps in the reflection will occur when the satellite is directly overhead and also when the satellite is viewed at low elevations, and the viewer is at an equal angle between two reflectors.
Figure 4. Location of Retroreflectors
This reflectors in this configuration are aligned 30° from the nadir, and would be visible from a further angle of 25° due to the FoV of the reflector. Therefore, the half-angle FoV would be 55° (from the nadir), and the total FoV of the system would be 110°. This would mean that the satellite could be tracked at elevations as low as 35° by SLR. Through modelling the orbit, it was found that the satellite would typically give SLR pass lengths of up to 4 minutes at 650 km, which should be adequate for SLR tracking.
The centre of mass of the satellite is about the centre of the +Z facet when the boom is deployed, i.e. about 530 mm above the reflectors. The reflectors are positioned about 150 mm from the centre of the -Z facet, and a tilt of 30° will put their apparent common reflection origin at 260 mm above the reflectors. Therefore, the vector from the reflection origin to the centre of mass is about 270 mm in the Z-axis through the centre of the satellite.
The configuration presented above uses separate reflectors, rather than a single SLR unit. The advantages of this are that the reflectors have smaller mounts, they can be moved away from the centre and so are more easily accommodated on the satellite platform. The precise location of the reflectors is not very critical: it should be possible to ensure that the reflectors are located within a millimetre of each other. The angle of the mounting need not be exact, as the light will always be reflected back the same direction. The mass of an individual reflectors is 20 g, and therefore the full complement of four reflectors and mounts has a low mass even by microsatellite standards. Obviously, the experiment is completely passive and requires no power, no commanding and will produce no on-board data.
With every new satellite design or configuration, there is a different perturbation to the antenna pattern, especially with lower frequency communications. The effect of the inclusion of reflectors on the -Z-facet must be accounted for, but could be minimised through location and mounting arrangements. The antenna pattern of a satellite is usually numerically modelled and tested with a scale mechanical model before the design is finalised.
With gravity gradient stabilisation and magnetorquer control, a reasonable attitude control is possible. The nutation has been assumed to be within 5° from the vertical with a period of over 30 minutes. This has the effect of increasing the minimum SLR tracking elevation to 40°. However, it also means that on some passes, the satellite will be visible with elevations as low as 30°. The yaw rotation is not critical, but has been assumed to be one revolution every 3-5 minutes. Therefore, the satellite will typically make one rotation in the duration of the tracking pass. If a reaction wheel is used to give stabilisation, then no yaw motion will take place. In both cases, there should be continuous tracking while the satellite is at an elevation of above 40°. To obtain very precise tracking data, the vector between the reflector centres and the centre of mass will be required. Given the 5° nutation, this vector will have a lateral change of about 26 mm. If necessary, most of this attitude error can be recovered through post-processing.
The laser retroreflector experiment cannot be flown on a satellite mission if there is a question over whether the laser could damage optical sensors. This is a valid concern, as the operators of Spot-3 refused to let the tracking stations use their lasers on Stella until its orbit had diverged sufficiently far from Spot. In general, all remote sensing payloads have large aperture lenses that could focus the received laser power onto one pixel of a sensor with potentially damaging results. To give an insight into the power levels from SLR stations, the intensity of the transmitted laser-beam would be sufficient to blind passengers in an aeroplane more than 30,000 feet above the station (in practice, radars are used to prevent this occurrence). However, this is certainly not the first time that SLR has been operated on satellites carrying cameras: of the 18 satellites currently being tracked by SLR stations, seven are remote sensing satellites or have onboard optical sensors of varying characteristics (ERS-1, ERS-2, TOPEX/Poseidon, GPS-35, GPS-36, and ADEOS, MSTI-2).
In this safety study for FASat-Alfa, the lenses and filters of the payloads under concern are listed, and the characteristics of typical and high power lasers found in SLR are investigated. The mechanism of damage to the sensors is discussed, and calculations are made which are confirmed by actual laboratory tests.
The cameras and the solar cells are the systems that seem most likely to be affected by the laser beam. Other systems considered here are the Sun Sensors and the Earth Underneath Detector (EUD).
Low cost remote sensing has been pioneered by the UoSAT missions from UoSAT-5 launched in 1991 to the current FASat-Alfa design [4]. A common CCD camera architecture is used for Earth imaging, ozone and star imaging, although photo-diodes are also used for ozone measurements. The optical and ozone cameras are the most vulnerable instruments to the laser beam, as the light is focused by a lens prior to the CCD (charge-coupled device) sensors. The laser will appear as a point source from the ground, and so all the laser light reaching the aperture of the lens could potentially be focused onto one pixel of the CCD. Very closely related are the photo-diodes (part of the ozone monitoring experiment), as these will also be fronted by a lens. The cameras have different-sized apertures and filters, and the laser tracking stations produce laser beams at different power levels, with different pulse characteristics and at different wavelengths. The worst cases must be investigated in order to confirm that there is no danger of any of the lasers at their characteristic wavelengths burning a pixel in any of the cameras.
Table 1. Cameras and Photo-diodes on FASat-Alfa
From Table 1, it can be seen that the ozone experiments have the largest aperture lenses, but have considerably greater attenuation out of band. All of the ozone experiments have the same lens geometry, and only differ regarding the exact interference UV pass-band. At a silicon level, photo-diodes are less vulnerable than CCDs to lasers due to their simpler structure (see Section 4.3). Therefore, only one representative case needs to be considered for all the ozone experiments.
There is a reasonable chance that a laser may exist in-band of the optical systems, but there are no tracking stations in the UV bands. The narrow-angle optical camera has a larger aperture than the wide angle, and so only the former will be considered, and treated for the worst case where there is no attenuation from the filter.
Some concern was voiced that the solar cells could be damaged by laser beams. The FASat-Alfa solar cells are made from Gallium Arsenide, and are particularly vulnerable to damage from partial illumination, or shadowing. The laser intensity itself will not do any damage, but if, for example, the light only falls on a few cells, the unequal charge could be potentially harmful on the string of cells.
The diffraction of the beam is adjustable, and with tracking a fast-moving low Earth orbit (LEO) satellite, the beam is spread wider to make it easier to find the satellite. The spread at 650 km is around 100 metres, with a gaussian fall-off [5]. Therefore, there appears to be no danger of the beam falling on one solar cell and not another, and the cells will not be damaged by the laser beam.
If sensors have no focus mechanism, then there is a considerable safety margin over the focused sensors. Both the Sun-Sensors and the Earth-Underneath on FASat-Alfa are unfocused, and therefore are not at risk.
The power levels of the different SLR stations that might track the satellite must be investigated so that the worst possible cases are considered. The following information was provided by Dr. John Degnan of the Laboratory for Terrestrial Physics at NASA GSFC [6], with additional facts from Dr Andrew Sinclair, Royal Greenwich Observatory, and Dr Richard Biancale, CNES [7].
There are currently about 43 SLR stations world-wide that track some 18 satellites equipped with reflectors. All of the high peak power laser systems have an operational wavelength of 532 nm. Some two colour experiments are being carried out at additional wavelengths, notably 1064, 355, and 680 nm, but it can be arranged through the CSTG that these lasers are not used on particular satellites if this presents a problem. The power levels of the 532nm SLR stations are typically 100 mJ per pulse or less. Only three or four systems are known to operate at 10 Hz; other repetition rates are 5 Hz or lower. Systems which operate at pulse-widths shorter than 100 psec (down to about 30 psec) generally have significantly lower output energies at the 10 to 30 mJ level.
A typical SLR station that will be tracking FASat is the SLR operated by the Royal Greenwich Observatory (RGO) at Herstmonceux. This has a wavelength of 532 nm (green) and is pulsed 10 times per second beam in 100 picosecond bursts with a transmitted power of 30 mJ. At the receiver on the ground, they expect to receive 1-100 photons per burst, depending on the satellite being tracked.
A worst case in the global SLR network can be represented by the Lunar Laser Ranging System at Grasse. This has an energy of 400 mJ, a pulse-width = 100 ps, with a PRF of 10 Hz. Relative to Grasse, most SLR systems produce a minimum factor of four less energy density at the satellite and a minimum factor of eight less average power. This laser is capable of operating with a divergence of as low as 4 arcseconds to track reflectors on the moon, and it can track high satellites only by increasing divergence to 100" (1 min 40 sec of arc). Although it probably cannot track low Earth orbiting (LEO) satellites, it will be assumed for this worst case study to be able to track a LEO satellite with a divergence of 30 arcseconds.
There are some high repetition rate (kHz) systems which give a high average power. However, these are all US and Russian military systems and they are not part of the international SLR network. Furthermore, the US military will never track a satellite without first obtaining permission from the satellite owner.
Unfortunately, a list of the different tracking stations around the world along with their different laser characteristics has never been assembled, so it is hard to be certain that all the lasers have been considered.
The divergence of the beam is varied according to the circumstances. In the case of a fast-moving low Earth orbiting satellite, for example, the divergence will be set wider to enable the acquisition of the satellite. The narrowest divergence ever to be possibly used on FASat is 30 seconds of arc.
The usual assumption made is that the beam propagates with a Gaussian intensity profile. Then the peak intensity is
(1)
where
Taking the worst case Lunar laser values applicable in this case:
More typical worst case values, based on RGO characteristics:
As a comparison, the eye-safe level is about 2.4 x 10-4 Joules/sq. metre.
There has been much practical work on the field of optical material damage due to lasers, including an annual conference devoted to the subject. A general review of the subject [8] is given by Wood [9]. The majority of the work seems to be concerned with the survivability of transparent materials and reflectors and there is only a small amount of specific information on the case of damage to photodetectors. Further information is given by Zhang et al [10]. on some tests undertaken with CCDs.
The Laser Induced Damage Threshold (LIDT) is usually given for a material in terms of either power density (Wm-2) or energy density, or fluence, (Jm-2). In general, it has been found that damage may correlate with:
The specific mechanisms through which the damage occurs are numerous and by no means fully modelled in theory, but can be classified into two types:
Dielectric breakdown: bulk and surface effects, material defects, avalanche breakdown.
Thermal Absorption: bulk and surface absorption, material impurities, induced conduction.
The case of a photodetector is more specific, in that damage will occur at much lower threshold than, say, an optically transparent glass lens due to the sensor's absorbent nature. The damage to such sensors is closely connected to the thermal absorption, rather than the high voltage dielectric breakdown effects [11], although it would be feasible for damage to occur through high induced currents causing semiconductor thermal runaway.
Of the six LIDT parameters above, the average energy density in a pulse will form the focus of the investigation, together with the CW Power density or average power density. The peak power is not a useful parameter for very short pulse lengths as explained in the next section, 4.2. Unfortunately, the spatial energy density of the focused pulses on the CCD is difficult to assess, and will be approximated as a gaussian profile. Neither can the pulse shape be judged exactly, as the characteristics of all the SLR lasers are not known, so again an assumption is made that the pulse is a perfect rectangular one. It will be sufficient for this study to ensure that a large safety margin is assured to allow for such unmeasurable effects.
Wood explains that as the pulse becomes shorter, the peak power becomes less useful in determining LIDT of semiconductors, and the energy density becomes the important parameter. Figure 6 is an example showing that as the pulse length decreases, the peak power damage (P) threshold increases until it appears to becomes inversely proportional to the pulse width. On the other hand, the energy density (E) flattens off to give an LIDT value characteristic of the material which is independent of pulse-width. When the time-constants in the semiconductor become longer than the pulse, then the time during which the energy is deposited appears irrelevant, as long as the average energy is below the LIDT.
Figure 6. LIDT - Photovoltaic CMT (Wood)
The relationship has been formulated into an equation:
(2)
where ET is the energy density required to raise the crystal surface to its melting temperature, is the density, R is the reflectance, k is the thermal conductivity, W is the radius of the laser beam, and is the pulse-width.
For longer pulse widths, it is clearly more sensible to use the power, or average power density as the parameter for measuring LIDT. This is also true for evaluating the LIDT with a repeating pulse, as long as the interval is long enough for the heat from the previous pulse to have mostly dissipated. In Figure 6, it can be seen that the power density LIDT has flattened off by the time the pulse is 10 msecs long. The PRF of the SLR stations under consideration for this study is 10 Hz. This is a very low repetition frequency, and so will be considered in terms of the average CW power.
Although Figure 6 is not directly applicable to CCDs due to their differing structure, the same principles will apply, and the curves will be of similar shape, characteristic of the silicon thermal time-constants. It is useful to note that the LIDT in terms of CW power density is several orders higher than the short pulse-width LIDT in terms of energy density.
Zhang et al report on a detailed study into laser-induced damage to silicon photo-sensor arrays, considering the degradation and failure of both CCDs and photo-diodes under laser power. It was found that CCDs have far lower damage thresholds than bulk silicon photo-diode arrays as a result of their different structures. A CCD consists of thin-filmed poly-silicon lines which are separated from the bulk silicon by an oxide layer (see Figure 7). The poly-silicon absorbs the laser energy, and for the energy to dissipate, heat must flow down the poly-silicon lines due to the higher thermal resistance of the insulator below. It is this heat flow that appears to cause the damage to CCDs, firstly degrading the breakdown voltage and increasing the leakage current above 0.3 J/cm^2, and then causing noticeable morphological damage at above 0.7 J/cm^2. The damage is expected to firstly occur in the non-active area, where the poly-silicon is separated from the bulk silicon by the thicker field oxide layer.
Figure 7. Cross-Sectional View of CCD (Zhang et al.)
Unfortunately, it is not easy to directly scale the results from this paper to extrapolate the required information for FASat, as there are a few different conditions. The CCD pixel size described by Zhang is 13 x 13 µm, with 0.5 to 0.6 µm thick poly-silicon lines over field oxides of 1.25 µm and thin oxides of 0.1 µm. The laser used had a wavelength of 1060 nm, pulsed at 10 ns at a 10 Hz rate. The EEV CCDs used on the FASat mission have a similar structure, but with larger pixels. Also the lasers to be encountered have different wavelengths and shorter pulses. Unfortunately, silicon absorption varies significantly depending on the wavelength of the light: silicon is very absorbent at 354 nm, but this absorption falls off as the wavelength increases. However, certainly a rough order of magnitude can be gained from the 0.3 J/cm^2 fluence required to cause degradation.
To find out more information specific to the behaviour of the FASat CCDs under laser power, the manufacturer, EEV was contacted. David Morris of EEV confirmed that the first damage expected would be the polysilicon electrodes of the CCDs melting under the heat. Unfortunately, they had no specific data of the laser damage threshold, but lasers are used as one of EEV's manufacturing processes to melt the backs of the sensors. This laser produces 0.8 J/cm^2 over 30 nsec at 354 nm. (It is believed to be at 10 Hz). There is a monitor CCD sensor that has an attenuator of 105, and works optimally. In fact once part of the attenuation was left out, giving a factor of 103 with no ill effects.
While the CCDs may have slightly different architectures, the laser fluence required in both cases to damage the CCDs agree in their order of magnitude. The difference could be attributed to either the different technologies or the different wavelengths of the laser. We now have an idea of the order of magnitude for LIDT in terms of energy density, but not in terms of the power density.
Some benchmarks have been set, but the knowledge of the CCD characteristics is by no means complete. It is possible that a detailed investigation of the LIDT of CCDs has been performed as part of a military program, for example. A more extensive review of the literature may yield more information on the relative sensitivity of different CCD technologies and the effects of different laser wavelengths, and perhaps give more quantitative data on CCD LIDTs.
To assemble these facts, we must consider the energy deposited, (the fluence), the average power and take into account the filtering of the camera. For the purposes of comparison, several reference points are defined: the EEV processes and sunlight.
As discussed earlier, the behaviour of silicon in different wavelengths is not constant, so therefore, many worst case examples will be used. The subsequent calculations will make use of the EEV burn process as a 'danger' reference point, and the 10^3 attenuated safe power level as the upper limit for safety. However, Zhang's results suggest that power levels can be tolerated much more closely approaching the EEV 'burn' process laser levels.
The sun's power is also used as a reference, as it is known that the optical cameras in the past have never had problems with direct sunlight, and no provision for solar protection has been made in the case of FASat-Alfa. It is also a matter of interest for the future to investigate the safety margin of damage from the sun.
Figure 8. Focusing of Laser onto CCD
Maximum intensity of laser at satellite = 57 x 10E-6 J/m^2 in
100 psec.
Near-IR Filter attenuates 532 nm by more than 10E3, therefore, 57
x 10E-9 J/m^2.
Lens size is 10 mm. Assume all light focused onto one pixel.
Therefore, energy deposited = 57 x 10E-9 . pi . (0.01/2)^2 = 4.5
x 10E-12 Joules onto one pixel.
Pixel size is 15 x 22 µ, so energy density (fluence) = 1.36 x
10E-6 J / cm^2.
Average Power = 10 pulses x 4.5 x 10E-12 / 1 second = 4.5 x
10E-11 Watts.
Maximum intensity of laser at satellite = 57 x 10E-6 J/m^2 in
100 psec.
Filter 105 attenuation, therefore 57 x 10E-11 J/m^2.
Lens size is 12.5 mm, energy on one pixel is 57 x 10E-11 . pi .
(0.012/2)^2 = 6.4 x 10E-14 Joules.
Pixel size is 22 x 22 µ, so energy density (fluence) = 1.32 x
10E-9 J / cm^2.
Average power = 10 x 6.4 x 10E-14 / 1 second = 6.4 x 10E-13
Watts.
Maximum intensity of laser at satellite = 1.075E-6 J/m^2 in
100 psec.
Filtered by more than 10^3, therefore, 1.075 x 10E-9 J/m^2.
Lens 10mm lens, energy on one pixel = 1.075 x 10E-9 . pi .
(.01/2)^2 = 8.4 x 10E-14 Joules.
Pixel size is 15 x 22 µ, so energy density (fluence) = 2.5 x
10E-8 J / cm^2.
Average Power = 10 pulses x 8.4 x 10E-14 / 1 second = 8.4 x10E-13
Watts.
Laser capable of burning CCD: Fluence = 0.8 J/cm^2 deposited
in 30 nsec.
We assume there is no focusing, therefore this is the intensity
that reaches the actual pixel.
Energy = 0.8 J/cm^2 * x 15 x 10E-4 cm x 22 x 10E-4 cm = 2.64 x
10E-6 J onto one pixel.
Average power = 10 x 2.64 x 10E-6 J / 1 sec = 2.64 x 10E-5 W.
Known 'safe' level is 10E3 below Case 4:
Fluence is 8 x 10-4 J/cm2.
2.64 x 10E-9 Joules onto one pixel.
Average power = 2.64 x 10E-8 W.
We assume the narrow angle sensor can survive the sun:
The Sun produces 1353 W/m^2, assume that 70% is optical and will
be received by CCD.
The sun subtends about 0.53° from Earth, corresponding to 650 km
x sin0.53° =6 km on the ground.
Assuming that the camera focuses 100 m onto one pixel, then sun's
diameter will be focused onto 60 pixels. This will cover a total
of pi x 30*30 = 2830 pixels
Lens receives from sun 1352 x .7 x pi x (0.010/2)^2 = 0.074
Watts.
This is spread over 2830 pixels => 2.6E-5 W (26 µ) per pixel
continuous power.
Table 2. Laser Power Calculation Summary
In terms of energy, it is clear that the energy received by FASat sensor is far lower than is received by the sensor undergoing the burn process at EEV, but a safety margin is required to allow for effects due to differing wavelength.
An important issue to note is that we expect the CCD's LIDT in terms of power density to be much higher than that of the energy density (see fig 1). In case 4, for example, the damage is caused by the high energy density in a very short time. If the average power 8 W/cm2 was applied, then no damage would be done.
The Spectroscopy laboratory at Surrey's Electronic Engineering Department has a 5 W Argon laser with a choice of wavelengths (351, 454, 457, 480, 514, and 528 nanometres). The usual operating wavelength is 480 nm.
For the purposes of the tests, we used a CCD camera which had the same sensor as the CCD cameras on FASat-Alfa. The camera was configured to enable real time operation viewed with a video monitor. The lens chosen was 10 mm aperture (16 mm focal length). Clearly, the claculated power levels must take into account the aperture of the lens if it is different from those on FASat.
The laser was unfortunately unable to be pulsed at high powers, so therefore we used an electro- mechanical shutter, so that the laser was operated at lower powers, but at a much higher duty cycle. Hence we could be sure that the CCD sensor was receiving equal or greater power levels than the sensor will on FASat.
The experimental set-up is shown in Figure 9. The laser is operated continuously, but is shuttered at 3 kHz by an electric rotating shutter. The shutter was adjusted to give a duty of about 33%, with pulse widths of 100 microseconds. The shutter was used as an attempt to approach the short pulse-width energy domain. However, the pulse-width is still significant with respect to the time constants, so the tests were still essentially in the power density domain.
To ensure that the profile of the laser beam was even, a lens was used which diverged the beam significantly wider than the aperture of the camera. The power level was calibrated through the use of a laser power meter. A 10 mm diameter iris was placed in front of the meter to enable the measurement of the power received by the camera. Using this power meter, we were able to confirm that the shutter was giving an approximate 33% duty cycle.
Figure 9. Laser Testing Set-up; a) Power level calibration; b)
Camera testing
From this measured reference power, it was possible to lower the power received by the camera to the required level through the use of calibrated neutral density filters. For example, an attenuation of 104 could be achieved by inserting two filters of factor 1 and 3.
The camera was evaluated by imaging test-cards of white and black in between each test. The real-time video output enabled the tests to be done surprisingly rapidly. However, for the formal testing, pictures were taken before and afterward the tests using the same electronic capture method as used on the satellite. The powers were then stepped up by the use of different filters from below expected power to well above the maximum expected powers, each time checking that there were no permanent visible effects on the video screen.
A matter of some importance was whether the camera lens was capable of focusing the light onto one pixel. This could be determined by examining the laser image while the power level was still very low. It could be seen that the laser was being focused onto approximately two by two pixels. At higher power levels, the laser saturated the surrounding pixels, and quadrupled the area. However, it was known that the energy was still concentrated on the central 2x2 pixels, as the profile of the laser still had to be the same.
Table 3: Laser Power Levels Subjected on CCD Camera
Operating Conditions: Average Power Power density Compared to Compared to
(onto 4 pixels) (W/cm2) worst case worst case
(Lunar) (1.36 x (RGO) (1.32 x
10-5 W/cm2 ) 10-7 W/cm2)
(Calibrated with no 0.5 mW - - -
filter)
Attenuation: 106 0.5 nW 3.8 x 10-5 2.7 290
Atten: 105 5 nW 3.8 x 10-4 27 2,900
Atten: 104 50 nW 3.8 x 10-3 270 29,000
Atten: 103 500 nW 0.038 2,700 290,000
Atten: 5 x 102 1 uW 0.076 5,400 580,000
Table 3 shows the experimental procedure. Increasing powers were applied to the camera until it was calculated that the average laser power levels were 5,400 times higher than the very worst case predicted for FASat in orbit. The camera was not damaged by this testing, and so the limit to the power density capability was never found. There was no intention of continuing the tests until destruction, and so the tests stopped at this level.
In this study, it has been proven that the cameras on FASat in a 650 km orbit can survive the average laser power level from the lunar SLR station. The safety factor is well over 3 orders of magnitude in the worst case.
In addition, the results from two other short pulse-width tests (EEV and Zhang) were used to prove that the cameras are safe from the energy density levels expected in orbit, again with a large safety margin of well over three orders of magnitude. Unfortunately the precise threshold of damage to these CCDs is not known, and so the damage threshold has been grossly underestimated. It is likely that the cameras have a safety margin of over 5 orders of magnitude from the lunar laser case.
These calculations have made use of known characteristics of the most common high power lasers at 532 nm. Although the other lasers (at 1064, 355 and 680 nm) have lower powers, it would be recommended that these lasers are not permitted to track FASat until their characteristics are known and a similar study should be made on the power levels.
In all of these calculations and tests, several significant worst case scenarios were used:
Lunar Laser: was assumed to be used while in practice, it will never be used on a low Earth orbit. It was chosen as the worst possible case for current and future SLR stations. Worst case calculations for a more typical SLR station (RGO) have been shown also.
Maximum Aperture Narrow Angle Camera: The narrow angle camera was used in these calculations, as it is clear that this is the most vulnerable due to its simple filtering and large aperture. The aperture of the camera is assumed to be set to the maximum size: 10 mm. In practice, this will be reduced according to optimal exposure levels, etc.
Filtering: The values for the out-of-band attenuation have been chosen again as the worst case values. If the filter characteristics are examined more closely for the laser specific case, the attenuation might well be another order greater.
Minimum Distance Assumed: It was assumed for these calculations that the satellite will only be 650 km away from the laser tracking station. This occurrence will almost never occur, and during which time, the laser tracking station finds it hardest to keep track of the satellite, so widening the divergence.
Minimum laser divergence: We have assumed 30 seconds of arc for tracking LEO satellites. In practice, this divergence will probably be 1 minute of arc or wider.
Considering that our tests indicate that levels that are caused by all these worst case scenarios are still safe, then we must conclude that there is no chance that the lasers on the ground can damage the CCDs.
In the course of this study, a few assumptions have been made. It is believed that the safety margins obtained account for these margins more than adequately. However, if these assumptions are resolved, it will inevitably add to the confidence in the security.
To complete these studies, it would be useful to investigate the power budgets for the other SLR stations that operate at different wavelengths. It is not anticipated that there should be any problems, but the exact characteristics are not known as yet.
A more exact investigation into the filters might be beneficial. All the safety margins will almost certainly increase, and it would be interesting to know the safety factor of the CCDs from the sun for future missions which may carry cameras with larger apertures.
The laser reflector experiment is technically a very simple passive project. An example configuration of four retroreflectors has been presented that provides a reasonable SLR tracking capability while taking very little on-board space. A study has been presented which gives a high degree of confidence that the UoSAT cameras are safe from all possible SLR station lasers. Numerical analysis was backed up by the laboratory testing of representative CCD sensors under a laser beam. To a large extent, the analysis presented here can be applied to satellites with different optical sensors in other orbits. If the safety and operational issues are resolved, the inclusion of laser reflectors on a small satellite has very little technical or financial impact on a project, while adding significantly to the satellite's research value.
Many thanks to Alvaro Valenzuela for reviewing this study, ensuring that the analysis was rigorous and the assumptions were valid. Thanks are due to Kevin Homewood and colleagues for their generous assistance in the Spectroscopy Laboratory, and thanks to Andrew Sinclair of the Royal Greenwich Observatory, Richard Biancale of CNES and John Degnan of NASA and David Morris of EEV for their help.
International Journal of Small Satellite Engineering - received 11 October 1995 [Author Resume] [Feedback Form for this paper]