You are here


Content owned by aadamson

Acquisition and Overheads

The acquisition steps for GeMS when working with 3 Canopus star + 1 ODGW are described in detail in: " Science Readiness of the Gemini MCAO System: GeMS" B. Neichel et al.

The acquisition procedure includes the following steps / overheads:

  • 1. Telescope slew and mechanisms in follow
  • 2. Laser alignment
  • 3. NGS alignment and Tip-Tilt loop
  • 4. High-order loop
  • 5. Focus loop
  • 6. Rotator and Tilt-Anisoplanatic (TA) loops
  • 7. Flexure loop

Ready for science: total overheads 30 min.

Acquisition details

Despite its complexity,GeMS is operated by a crew of only two people: the telescope operator manages all the AO systems, while the laser operator is in charge of the Laser and BTO. The instruments are still operated by an astronomer.

Most of the complexity of GeMS resides in the control and optimization of all the loops. There are as many as 20 loops and offloads that must be closed, monitored, and controlled. Further complicating the process, most of these loops are interacting together. Because managing all of these loops is out of the scope of the telescope operator, we have developed a layer of software called GeMS SMART tools. These SMART tools simplify the interaction with the AO system and integrate automation in the operation, assisting the operator both in the acquisition procedure and during the observation.

Below we detail the steps required for the acquisition. Complementary information regarding overheads can be found in the GSAOI page.

1. Telescope slew and mechanisms in follow

When the telescope operator slews to a new GeMS target, the Myst SMART tools also receive the slew command. The SMART tools then set all mechanism offset positions to zero and, based on the telescope position and temperature, puts all mechanisms into follow according to their LUT (or with probe mapping as is the case of the NGSWFS). Once the telescope is in position, the NGS and LGS alignment are done simultaneously by the telescope and laser operator maximizing on-sky efficiency. In most cases during a slew, the GeMS mechanisms reset and reach position before the telescope, thus the overhead on this step usually comes from the telescope-side.

2. Laser alignment

With the telescope now in position, the laser operator begins alignment of the laser inside the BTO only, at low power. This alignment consists of a handful of steps, which in total take between 3 to 6 minutes to complete. First, the laser must be aligned to the LLT. For safety reasons, this alignment is executed at a very low laser power of only a few mW. Using the video signal coming from the Pre-Alignment Cameras (PACs) the laser operator can accurately center the beam on each of the BTO mirrors leading up to the LLT. Once this critical alignment has been completed, the laser operator requests permission from the telescope operator to propagate onto sky.

The telescope operator then checks for any beam collisions that might impact the science of other telescopes, and requests permission from the laser spotter stationed outside monitoring for aircraft. With the permission to propagate, the laser operator increases the laser to medium power (2-5W) and then checks the alignment of the laser beams on the LLT. Once the beams are properly centered on the LLT, the laser operator increases to maximum power and using the Rayleigh light pattern, starts the alignment of the laser constellation on the LGSWFS.

A simple geometric model of the Rayleigh backscattering has been implemented and a minimization routine provides the geometric solution consisting of several parameters, of which the global de-pointing is the most important. This global de-pointing is applied to the BTO pointing and centering mirrors (PM/CM) to center the LGS constellation. With the constellation aligned, the laser operator closes the FSA loop keeping the constellation fixed on the LGSWFS. Finally, when everything has stabilized, the laser operator gives the green light to the telescope operator to close the high-order loop.

3. NGS alignment and Tip Tilt Loop

To maximize on-sky efficiency, the telescope operator starts the NGS acquisition in parallel with the laser alignment. Depending on the field and brightness of the NGS, this can take between 2-10 minutes (a crowded field, faint or diffuse NGS targets are more difficult setups). As the Gemini South telescope pointing accuracy from slew-to-slew is on average 10 arcseconds, it is necessary for the telescope pointing first to be corrected. This is a standard procedure for any telescope slew, but all the more important for GeMS acquisitions as the NGSWFS probes have only a 1.5 arcsecond FoV.

To do this, the telescope coordinates are swapped from the science object to those of the C3 coordinates thereby putting the C3 star on-axis. Next, the wide-field Acquisition Camera (AC) is used (1x1' when used with Canopus) and the pointing adjusted to put the C3 target onto a pre-determined telescope hot spot. With the pointing now corrected, the telescope is reverted to the science object position and the Canopus probes, using the probe mapping model, will move into position, ensuring that the star corresponding to C3 will be properly centered on the probe. Next, the Tip-Tilt loop is closed with C3 only, while C1 and C2 APDs are readout to check the counts. Assuming the NGS catalog coordinates are all correct, probes C1 and C2 will likely be very close to their NGS targets, and flux should be seen on the APDs.

The telescope operator can then start the Myst astrometric mode, which automatically adjusts the probe positions in a more accurate manner. Once the remaining two probes are centered, the Tip-Tilt loop can be closed on all three probes. If one of the stars does not appear on the APD (e.g. because of catalog errors), the telescope operator can start an automated spiral search for the star. In the event that a star still does not show up, it is possible to swap the telescope coordinates with that guide star to look for it in the AC (as was done first with C3). If necessary, the AC images can also be compared with finder charts, helping the operator identify the correct guidestars. These additional steps require the C3 offload to Tip-Tilt to be stopped, and add additional acquisition overhead. In practice, bright guidestars in relatively sparse fields require the least amount of overhead.

4. High-order loop and dependencies

With the laser FSA offloading and the NGS Tip-Tilt loop closed, the telescope operator can now close the high-order loop. While the command to close the loop is sent from the Telescope Control System (TCS), it is the Myst SMART tools that actually handle the proper sequencing. Default matrices and gains are automatically loaded and frame rate is optimized based on the NGS and LGS flux. Offload of the static DM0 shape to M1 is also started based on the loop performance. This entire process happens in 1-2 minutes and is monitored by the telescope operator for stability.

5. Focus loop

With the high-order laser and NGS loops closed, the telescope operator can now focus on the SFS loop. Depending on the C3 target brightness (SFS receives 30% of the C3 light), this process can take 2 to 5 minutes. First, the telescope operator starts the readout of the SFS camera with the best guess for integration time (based on a table), adjusting to optimize the SNR on all four spots. In the case of a very faint GS, the integration time can be as long as 5 minutes, explaining the large overhead for this step. When the SNR is sufficient, the telescope operator closes the SFS loop, which applies a focus term to the DM0 that is immediately offloaded to the secondary mirror of the telescope (M2). The telescope operator can see real-time the trend of the M2 focus, visually showing when the focus has converged.

6. Rotator and Tilt-Anisoplanatic (TA) loops

At this point the telescope operator can close the Rotator and Tilt-Anisoplanatic (TA) loops. The rotator loop compensates for any tracking errors of the telescope rotator, while the TA loop compensates for any differential Tip-Tilt errors between the three NGS probes, thereby fixing the plate scale. Information about this loop performance is also provided by the SMART tools GUI.

7. Flexure loop

The final step of the acquisition procedure is to close the flexure loop. This can take from 2-5 minutes. First, the telescope operator reads out the ODGW that will be used for flexure compensation. Exposure times for the ODGW are pre-defined based on flexure star magnitude and filter used by GSAOI, and can be adjusted as needed . The size of the ODGW is first set to the largest value (128x128 ∼ 2.5 arcseconds on-sky). If the star does not appear on the ODGW, a full-frame image can be taken with GSAOI to check its position with respect to the guide window, however this adds overhead. Once the star is properly centered on the ODGW (by re-defining the ODGW X,Y detector position, not the telescope position since we are at this point fully closed loop with the LGS and NGS T/T stars), the size of the ODGW window is reduced to 16x16 (∼0.3") and the flexure loop is closed.

Science ready

Finally, everything is ready for science. The telescope operator will give a final check on all the loops and ofloads, and give the green light to the astronomer to start the science sequence.

The large range in setup overheads can be almost entirely attributed to the NGS acquisition. For bright NGS without catalog errors, the full GeMS acquisition has been as fast as 10 minutes. For faint NGS, or more complex objects (high background or a crowded field) the acquisition procedure can be 20+ minutes. We have also experienced several occurrences where the NGS brightness as listed in the catalog were incorrect by several magnitudes thus a better NGS constellation needed to be chosen on the fly, greatly increasing the setup time. Because of this, we are considering the possibility of pre-imaging GeMS fields. While this would add overheads to the program, pre-imaging would be fast, no more than 5 minutes for each field (accounting for the telescope slew and image with the AC), and would not require good conditions.

Dither-Filter-Sky sequence

A science observation sequence can include telescope offsets for image dithering, a change in filters, and large telescope offsets for sky background images. For all of these events, specific GeMS loops must either be paused or opened, and then resumed automatically after the event. This is handled by the SMART tools. The sequence of events for each case is described below.

Dither sequence

When a telescope offset is required, the observation sequence executor (Seqexec) sends the information to the TCS. The SMART tools then receive a message "dither" and "pause" thereby pausing all the NGS loops and dependencies (Tip-Tilt, TA, rotator, flexure and focus), while all the laser loops are kept closed. The telescope then offsets, and once all the subsystems report that they are in position, the NGS loops are resumed, the "science ready" flag is green, and the next science exposure starts. Depending on the size of the offsets, a telescope dither take ~30 seconds. The dither pattern is set by the NGS and ODGW acquisition fields and cannot be larger than 30 arcseconds.

Filter sequence

Depending on the mode of observation, 1+3 or 3+1, a change in the instrument filter can have different consequences. For a 3+1 setup (i.e. 3 Tip-Tilt NGS and 1 OI for flexure) a change in the instrument filter only implies a pause/resume of the flexure loop. In the case of fast-guiding with the GSAOI OGDWs (1+3), all the NGS loops and dependencies must be paused before the filter is changed. Note that depending on the filter, the exposure time (and frame time) of the ODGW should be adjusted to keep the SNR constant. A change in filter also induces focus offsets, necessitating a shift in ODGW window positions as defined in a LUT. The overheads for a filter change are set by the instrument, for instance, GSAOI will take ~15 seconds to perform any filter change.

Sky sequence

A sky sequence is somewhat similar to a dither, except that the telescope offset is usually much larger (up to 5 arcminutes). In this case, all NGS loops and dependencies are paused, and the probes are frozen and do not follow telescope offsets (as a large offset would put the probes into a hard limit if they remained in follow). We found that large telescope motions could create instabilities in the high-order loop, as the LGS may be lost for a few seconds. Because of this, we have decided to also pause the high-order loop, but keep the LGS stabilization loop closed (FSA loop).

Large telescope offsets

If the observation requires offsets larger than 5 arcminutes from the base position, the laser propagation must be stopped due to Laser Clearing House restrictions . When the telescope returns to the original base position, the laser operator must re-acquire the LGS. Moreover, since this offset will also be unguided, the telescope operator must re-check the NGS acquisition and correct for any telescope pointing errors. In this case a separate observation block is required and extra overheads are introduced.

Loops and Offloads

Click on the arrows to navigate through the different steps.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Control Loops

LGS loop

This is the heart of the MCAO correction. This loop (also called high-order loop) controls the 3 DMs shape based on the measurements from the five LGS WFSs. We use a tomographic algorithm process that optimally finds the best shape to apply to each DM in order to have the best performance. This reconstructor is updated on the fly depending on the atmospheric conditions (seeing, wind speed, etc.).

NGS loop

Tip-Tilt and Plate Scales modes

This loop controls the Tip-Tilt and Plate Scale (anisoplanatism) modes. It is well known that the position of the LGS WFS cannot be trusted on the sky. This is called the "tilt indetermination problem" and is basically due to the fact that the laser beam uplink tip-tilt is unknown. Thus, we need NGSs to provide this part. In classical, one LGS AO systems, the only modes we care about are tip and tilt. Here with MCAO, because we want to correct not only on axis but also off-axis on an extended field (that is, we correct for anisoplanatism), we need - in addition to tip and tilt - to correct for Tip and Tilt anisoplanatism. These anisoplanatism modes are plate scales modes. Then, from the three TT WFS, we can deduce the following:

  • Global (average) Tip and Tilt (2 modes)
  • Plate scale modes (global shrinking of the image, or stretching in X and shrinking in Y, or same rotated 45 degrees = 3 modes)
  • Image rotation (1 "mode")
Below is an animation of these 5 modes on an star field:




Astigmatism 0

Astigmatism 45


These plate scale modes are not produced by the TT mirror, but by the DMs. It turns out that by combining a de-focus D0 on DM9 and the opposite d-efocus -D0 on DM0, one can control plate scale in the output focal plane. Note that rotation does not corresponds to anything that is correctable with the DMs. It is not a mode which can be induced by any physically possible phase aberration. It can only be a rotation, and thus needs to be corrected by a element that can induce a rotation. The only such element is the Cassegrain rotator. Thus this fifth TTWFSs measurement will be low pass filtered and send to the Cassegrain Rotator as an error term, that has to be added to the current target.
The NGS can be located anywhere in the 2 arcmin Canopus input field (defined by the canopus mechanical input port), but should preferentially be as widely spread as possible. At first approximation, the field inside the area defined by the 3GS triangle will be best corrected, and the correction will degrade outside this area.

Slow Focus Sensor (SFS)

The range (the altitude) of a Laser Guide Stars depends on the altitude of the Sodium Layer. When looking at zenith, this altitude is ~90km. As we are using the LGS to compensate for atmospheric focus, we cannot disentangle any changes in the sodium layer altitude from real atmospheric focus changes. In other word, if the altitude of the LGS varies (by a few hundreds of meters), we will interpret this as focus, and the DMs will compensate for it. As a matter of fact, the image appear out of focus in the science instrument. To cope with this effect, we need a "truth focus sensor" which monitors the true focus on an NGS. This is done by the SFS, with a 2x2 Shack-Hartmann sensor that uses 30% of the light of one of the NGS probes. This error focus signal is sent to the LGSWFS zoom, as offsets around the best focus position.

Secondary Loops and offloads

LGS Tip-Tilt stabilization

The natural TT only have little to do with the position of the LGS. The position of the LGS on the LGS WFS will also be a result of the laser beam direction, LLT flexures, etc. This overall spot centering of the LGSs on the LGS WFS need therefore to be somehow compensated, otherwise they will slowly drift and eventually get out of the LGS WFS subapertures field of view. To achieve this, we use the LGS WFS derived Tip-Tilt signal to drive the BTO Field Steering Array (one small TT mirror per beam). In effect, the FSA move the laser beams on the sky (independantly for each beam), and thus allow to recenter them on each LGS WFS. The average motion of the FSA is offloaded to the Poiting/Centering Mirror. The average rotation of the constellation is offloaded to the K-Mirror.

M1 Offload

By default, PWFS1 will not be used together with GeMS . PWFS2 cannot be used, as it occupies the same physical space as the AO fold (flat that sends the light into Canopus). Therefore, we are running "open loop" on the primary mirror figure, only using a LUT after the night initial setup. There might be, and probably will be, some drift, as the LUT are not perfect. In particular, astigmatism is likely to drift slowly. Correcting these slow drifts of the primary mirror figure is not a problem for the AO system, except for the fact that it will eat up some dynamical range from the DMs (more precisely, DM0). To avoid this, we offload at a low rate (typically every 10sec) the low order spatial modes of DM0 to M1 (typically 6-10 modes).

M2 Offload

Similarly, we do not use a tracker. Thus, if the open-loop telescope tracking is imperfect, this will slowly build up on the TT mirror. In a similar fashion as for the M1 offload, we offload the TTM average position to M2. In fact, because M2 accept commands up to 200Hz, we do the offload at this rate, albeit with a very low gain, which in effect means offloading only the very lowest temporal frequencies.

Elevation LUT to LGS WFS zoom

Following simple geometric considerations, the distance from the telescope to the sodium layer (and thus the physical LGS) vary with the telescope elevation. The average sodium layer altitude is 90km. Hence the range to the LGS is 90km at zenith, but it becomes 180km when looking down at zenith angle = 60 degrees. The LGS WFS zoom is a device that allows to maintain the LGS in focus on the LGS WFS themselves, to accommodate for this large LGS range variation. Because the main term in the LGS WFS zoom position is a (predictable) geometric term driven by elevation, this is taken in to account with a LUT.

Flexure Guide Star

There will be flexure of the science instrument vs. GeMS. For instance, flexure of the CANOPUS output optics (AOP2 and fold), or flexure of the science fold, or of the science instrument internal optics. Flexure of the CANOPUS TTWFS will also induce unwanted motion of the output beam, through the AO close loop. To avoid image motion on the science detector, the best approach is to actively sense its position on the science detector itself, or as close as possible to it. In GSAOI, we have the possibility to have On-Detectors Guide Windows (ODGW), that measure the position of a star on the detector. The error signal, if any, will be used to drive the TTWFSs probe position (probe#3 will move, #1 and #2 will move along with it, being positioned on top of #3), to compensate for possible flexure.

Non Common Path Aberrations Correction

An AO system can compensate for optical aberrations, but only the one that it can see with the WFS. All the optics that are after the WFS can introduce aberrations that will reduce the image quality on the science detector. On the other hand, all the aberrations that only exists in the WFS path will be compensated, when they are not supposed to be, degrading the IQ as well. The combination of all of these reduces significantly the static performance (when there's no atmospheric turbulence) and the Strehl Ratio we get are of the order of 20% in H-band. But as these aberrations are static, we can pre-compensate for them by adding a static shape on the DMs. To define this static shape, we optimize the image quality on the science detector. This is done by using focal plane wave-front sensing technique such as phase diversity. But for GeMS, this has been generalized to tomography.