We’ve closed the loop* on MagAO-X!

* with two DMs! But not at the same time…yet.

Late last week, after painstakingly recabling and aligning the BMC 2K following its relocation to the MagAO-X instrument, we closed the loop at 3.6kHz with 2040 actuators. See Jared’s video below:

MagAO-X Closed Loop 3.6 kHz

What exactly are we looking at in this video?

On the far left is the image from our pyramid wavefront sensor. It’s tricky to interpret, but the four pupil images are a bit (but not exactly) like a 2-axis knife-edge test, with the key difference that aberrant rays are refracted into the different pupil images rather than simply blocked or passed. Through the magic of linear algebra (and lots of calibration), each frame from the wavefront sensor is converted into a map of voltages to apply to the DM to cancel this wavefront error.

The DM commands can be found along the bottom row of windows in the video. It’s split across multiple channels, but the important one is the image on the far right: each pixel is a command we’re sending to an individual actuator on the DM. With the loop open, we’re just creating simulated atmospheric turbulence. With the loop closed, it’s the same simulated turbulence plus the correction computed from the wavefront sensor.

And, finally, on the upper right is the “science” PSF, doing its thing. (That last image on the desktop—the pixelated one in the top middle—is the command sent to the ALPAO DM, but it’s not doing anything here other than holding a flat shape.)

Olivier gestures at a computer monitor while Jared looks on
Jared and Olivier debate the finer points of cacao.

Before we could close the loop, the 2K had to be aligned, which isn’t a trivial task when you’ve just dropped it into the middle of a rather complicated optical system and expect the beam to be centered on the DM to better than one actuator (which have a pitch of 400 microns). Enter Laird and Alex, experts on all things alignment. To aid in their efforts, we placed a pattern on the DM that could be seen both on the wavefront sensor and by eye in the beam reflected by the DM.

This isn’t the first time we’ve closed the loop on MagAO-X. A month ago, we closed the loop on the low-order ALPAO DM-97 (the woofer). We have video evidence of that too:

MagAO-X woofer, closed loop at 2 kHz

And finally, to procrastinate studying for finals a few moments more, here’s my half-micron (peak to valley) entry into the ongoing MagAO-X logo contest, imprinted on the 2K and measured on our Zygo interferometer:

MagAO-X logo animated on the 2K BMC

Placing our 2040-actuator deformable mirror in MagAO-X

As of today, our 2040 actuator Boston Micromachines MEMS deformable mirror (BMC-2K DM, for short) has been moved to MagAO-X instrument optical table. With a cost of roughly three houses, it’s by far the most expensive piece of the whole project. (If you don’t count paying half a dozen graduate students for half a decade.)

View of the BMC-2K DM mounted in its holder on our MagAO-X optical bench.
The BMC-2K in all its glory, finally in place atop the MagAO-X optical balcony.

So, why is it important? And what makes it so expensive?

Adaptive optics involves first sensing the shape of an incoming wavefront of light to determine aberrations, then deforming a reflective surface to perfectly cancel out as much of the aberration as you can. So, as you might guess, a deformable reflective surface is key.

Extreme adaptive optics is an informal term for the next stage in the evolution of adaptive optics for astronomical high-contrast imaging. We’re running our system faster than predecessor systems like MagAO (in terms of the number of measurements and corrections each second), placing more stringent tolerances on all of our optical surfaces, and using more actuators on our DM. Unlike the MagAO system, which deforms the telescope’s secondary mirror directly, MagAO-X uses three DMs placed at images of the pupil within the instrument enclosure.

Dr. Jared Males leaning over some cable assemblies to disconnect them.
The P.I. disconnects one of the 16 sets of ribbon cables necessary to interface the DM with its high-voltage drivers.

The first DM in the optical path, an ALPAO DM97, is a large-stroke device, meaning it can deform a whole 80 µm from one edge to the other. This is about the diameter of a human hair, which doesn’t seem “large”, but for H-alpha (0.656 µm) photons 80 µm is over 120 wavelengths. The flip-side is that it has only 97 actuators. We call this the “woofer” by analogy with speaker systems, since it can only correct aberrations with low spatial frequencies.

The last DM the light will encounter before being imaged onto a detector is another ALPAO DM97. This one is tasked with squashing “non-common path” aberration: basically, any aberrations we’re introducing ourselves within the instrument that aren’t being sensed by our wavefront sensor.

Interface plate where ribbon cables from the DM driver meet ribbon cables from the DM.
Not for nothing is this thing called the octopus.

The device we moved today is the “tweeter”, responsible for correcting the high-spatial-frequency modes that generate speckles in our images. These speckles can look awfully similar to planets, and can even persist in a quasi-static way in a series of images. After we’ve taken out the low-frequency content with our woofer, the residual aberration is smaller amplitude but higher frequency.

Our BMC-2K DM lets us cancel out these aberrations to a high degree, resulting in more control over speckle-causing aberrations and less light lost from the core of the image of each star or planet.

Thanks to Jared Males, Kelsey Miller, and Lauren Schatz for the patient explanations that informed parts of this writeup.

Meet the new blog, same as the old blog

To simplify maintenance, we’re making the MagAO-X blog a continuation of the MagAO blog. Going forward, we will be doing our project updates here on xwcl.science. I did my best to migrate things without breaking references to our previous nine years (!) of posts. However, if you notice any broken links to visao.as.arizona.edu, do let me know.