2014A Day 21: Loops of MagAO

A closed feedback loop is when you are monitoring some output so that you can control some input. How many closed loops does MagAO run? Here we present: The Loops of MagAO.

The top-level loop: The AO loop.

1. The AO System’s Pyramid WFS and ASM

The top-level loop is the adaptive optics (AO) loop. This is the loop that all the others are here to serve. We are making flat wavefronts so that our science cameras can take sharp images, and it is a serious business.

How a closed-loop AO system works: Your flat wavefront in space is distorted by turbulence in the atmosphere.  The distorted wavefront encounters the deformable mirror at the telescope (the ASM), and a beam splitter sends the bluer light to the wavefront sensor (WFS), where a control system calculates the shape of the wavefront, then applies the opposite shape to the ASM.  The corrected, flat wavefront is then sent to the science camera (Clio2 or VisAO).
The AO system in control
Me running the AO system a few nights ago. Dear Laird: Do you notice how I’m taking logs?
And when everything is running smoothly, this is the AO Interface that the AO operator can use to close the loop

2. The Camera Lens

This loop is my favorite, because it’s one of the subtle calibrations we do that keeps our AO system one of the best in the world. The camera lens loop keeps the positions of the Pyramid pupils aligned to the pixels on the WFS CCD to a tenth of a pixel. This means our AO system is always calibrated, in the way that it measures brightness and on the CCD and converts it to slopes to send to the ASM.

The camera lens loop is my favorite. (Well of course, besides the AO loop). Left: The light falls in these 4 pupils after it hits the pyramid, one for each facet. Center: We measure the position of the pupils in software (red cross-hairs and thin circles). Right: We compare the measured positions to where the software is expecting the light (blue and red lit-up pixels), and the camera lens loop moves the camera lens to line up the pupils with the pixels we want them to fall on.

3. The 585 ASM Sensors

The ASM has 585 actuators to control its shape at 1000 times per second, and they have sensors to control their current and check their temperatures.

The subtlety is that the mirror shape is actually controlled by the DSPs upon the back of the ASM itself.
The control electronics for the MagAO ASM
Keeping house
The ASM housekeeper tracks the temperatures, currents, and forces of the 585 ASM actuators

4. Telescope Off-loading

We send some of the wavefront correction to the telescope — we call this off-loading. For example, if the ASM has to tilt too far to the side and starts to use up all its “throw” or stroke, then we just send a little nudge to the telescope and re-point the whole telescope, flattening out the ASM. We do this once per second, and we off-load focus once every minute.

5. VisAO Coronagraph Guider

Jared wrote a little opto-mechanical loop for VisAO in coronagraph mode. He nudges the VisAO gimbal mirror to keep the star aligned precisely behind the coronagraph. The loop runs once every few to tens of seconds.

Jared running the coronagraph guider loop on VisAO
Here we see the offsets scrolling by, and finally a gimbal command at the end
The coronagraph guider in action

6. Clio Temperature Controller

The Clio2 optics are kept at 77K via the outer dewar, by the LCO staff who refill its liquid nitrogen dewar every morning. The Clio2 detector is kept at 55K by a pump that lowers the pressure of the liquid nitrogen and makes it solid inside the inner dewar. However, the pump could keep lowering the pressure and thus the temperature even more, but it’s important to keep the temperature stable. Therefore, we have a heater that senses the current temperature, and turns on a bit when the temperature is below 55 K, and keeps it always at 55K. This is a closed feedback loop.

The Clio temperatures

7. Mechanical Loops with Encoders:

We also control a lot of mechanical components using encoders. On the WFS/VisAO board, called the “W-unit”, we have the Bayside stages X, Y, Z; the PI piezo Tip/Tilt mirror X, Y; the camera lens X, Y; the two atmospheric dispersion compensators (ADCs) and the re-rotator (K-mirror); the beamsplitter and the two VisAO filter wheels; and the gimbal motors X, Y. That’s 15 encoders:

All the things on the wavefront sensor and VisAO that move mechanically and with encoders

8. Finally, the telescope itself has several mechanical loops: Elevation; Azimuth; the Dome; and Active Optics (the primary mirror M1 has ~150 actuators controlled via a closed-loop Shack-Hartmann (plus the 5-d vane ends (x,y,z, theta, phi))

The Shack-Hartmann guider loop
The back of the primary mirror, where there are actuators controlling the active optics

Well, I lost count, but that’s a lot of control loops! And when it’s all working, this is what we get:

60 milli-arc-second PSF at H-band on a 7th-magnitude guide star. That’s really good! Also it has a faint companion…

Well, that’s it for tonight, suffice it to say we had a good busy night on sky.

The moon to the west, at sunrise, from Clay
The moon setting as the sun was rising, the morning after the lunar eclipse

The song of the day has an astronomical theme, is by a top South American artist, and it came out on Vevo the day we left Tucson for this trip:

Here’s another good one by Shakira, from when the World Cup was in South Africa, it’s in the top ten most viewed Youtube music videos of all time:

Speaking of the World Cup, I’m happy to report that there is a soccer field at LCO! But it’s near the gate and we never go by there, so I’ve never seen anyone playing soccer here.