HARMONI at ELT: LPOU calibrations for a micron-level pointing accuracy

Imagine the night sky as a giant, cosmic dance floor where stars twirl and galaxies spin. For centuries, astronomers have tried to take crisp, clear photos of this dance, but Earth's atmosphere acts like a wobbly, heat-hazed window, making the stars blur and jitter. To fix this, scientists are building the Extremely Large Telescope (ELT), a massive mirror the size of a football field, designed to peer through that haze. But even a giant mirror needs a steady hand. If the telescope moves even a tiny bit, the image blurs. To solve this, the telescope uses a "guide probe" system—think of it as a robotic arm with a camera eye that constantly watches a nearby star to tell the telescope, "Hey, you're drifting left, move right!" This system is part of a super-sensitive instrument called HARMONI, which acts like a high-tech camera that can split starlight into a rainbow of colors to study what stars are made of. The challenge is that the robotic arms themselves can get a little "sloppy" with their movements, and if we don't know exactly how sloppy they are, we can't tell if the telescope is moving or if the robot is just having a bad day.

This paper tackles that specific "robotic sloppiness" problem for the HARMONI instrument. The authors, a team of engineers and scientists, are worried that the robotic arms used to point the guide cameras might have tiny, hidden errors in their motors. If these errors aren't measured and fixed, the telescope might think it's pointing at a star when it's actually looking slightly off-target. To fix this, they needed a way to calibrate the robots inside the instrument, right where the light enters, rather than relying on outside references that could be distorted by other parts of the telescope. They proposed a clever solution: a special "calibration mask" (a pattern of tiny light dots) that could be dropped into the light path, along with a special lens to bring those dots into focus. However, since this mask would be placed in a spot where the light isn't perfectly focused, they had to design a new lens to handle it. The team didn't build a physical prototype for this test; instead, they used a computer simulation software called RayZaler (which they helped create) to model the light paths. They compared their new model against a trusted, existing software called Zemax to make sure their math was right. Their simulations showed that their new design works: the lens successfully focuses the calibration dots, and the system demonstrates a linear response with a deviation from perfect linearity of just 150 nanometers (nm) in the simulation—a value dominated by how the computer sampled the light beams rather than a flaw in the design. While they didn't prove this works in the real world yet, their computer models suggest that this approach is a solid, viable way to keep the telescope's robotic eyes perfectly sharp.

The Cosmic Jitters and the Robotic Arms

To understand why this paper matters, we first need to look at the ELT, or Extremely Large Telescope. Imagine trying to take a photo of a firefly on a windy night while standing on a boat that's rocking back and forth. That's what looking at stars from Earth is like. The air in our atmosphere is always moving, which makes stars twinkle and blur. The ELT is a massive telescope designed to fix this, but it needs a "steady hand" to keep its eyes locked on a target.

Enter HARMONI. Think of HARMONI as the telescope's super-powered camera. It doesn't just take pictures; it splits the light from stars into a rainbow (a spectrum) to tell us what those stars are made of, how hot they are, and how fast they are moving. But for HARMONI to work perfectly, the telescope has to be incredibly steady. If the telescope moves even a tiny bit, the image gets blurry.

To keep the telescope steady, it uses a system called "adaptive optics." This is like a magic mirror that changes its shape hundreds of times a second to cancel out the blurring caused by the air. But to do that, the mirror needs to know exactly where the telescope is pointing. This is where the "guide probes" come in.

The Problem: When the Robot Gets Clumsy

The guide probes are like robotic arms with eyes. They reach out into the sky to grab a "guide star" (a bright star near the object the telescope wants to study). By watching this guide star, the system can tell if the telescope is drifting. If the star moves in the camera, the system tells the telescope to move back.

However, the robotic arms themselves are made of motors and gears. Just like your own arm might shake a little when you're tired, these motors can have tiny, systematic errors. They might move 1 millimeter when they think they moved 1.001 millimeters. If the telescope doesn't know about this tiny error, it might think the telescope is moving when it's actually just the robot being a bit clumsy. This leads to bad measurements.

The team needed a way to check the robots' accuracy inside the instrument. In the past, they tried using a "calibration mask" (a sheet with a pattern of dots) that could be dropped into the telescope's main view. But this had a problem: the light had to pass through other parts of the telescope (like the MORFEO system) before reaching the guide probes. Those other parts could distort the image, making it impossible to tell if the error came from the robot or from the other parts of the telescope.

The Solution: A New Lens and a New Spot

The authors proposed a new idea: put the calibration mask and a special lens right next to the guide probes, deep inside the instrument. This way, the guide probes would only see the mask and the lens, ignoring everything else.

But there was a catch. The spot where they wanted to put the mask wasn't a "focal plane" (a place where light naturally comes together to form a sharp image). If they just put the mask there, the dots would look like blurry blobs. To fix this, they needed to add a "deployable lens" (a lens that can be popped in and out of the light path) to bring those blurry dots into focus.

They had to be careful, though. The space inside the instrument is very crowded. They couldn't just put the lens anywhere; it had to fit between existing mirrors. This meant the lens might not make the image as big as they wanted (a "magnification ratio" less than one), but it had to be big enough to be useful. They set a rule: the image couldn't be shrunk by more than 10 times, or it would be too small to measure accurately.

The Test: A Virtual Experiment

Since building a real prototype is expensive and time-consuming, the team used a computer simulation. They used a software called RayZaler, which they developed themselves, to model how light would travel through the new setup. They also used a trusted software called Zemax to double-check their work.

They simulated a star shining into the telescope and then simulated the calibration mask with five dots: one in the center and four around it. They moved the center dot in tiny steps (1 millimeter) and watched where the image of that dot landed on the detector.

The Findings: It Works!

The results were promising. First, they checked that their RayZaler software matched the Zemax software. The images of the star spots were almost identical, with differences of less than 1 micrometer (1 µm). This gave them confidence that their model was accurate.

Then, they tested the new calibration setup. They found that:

1. The lens worked: It successfully brought the blurry calibration dots into focus.
2. The magnification was good: The system shrunk the image by a factor of 3.69 (meaning the image was about 1/3.69th the size of the real object). This was well within their safety limit (they only rejected anything smaller than 1/10).
3. It was highly linear: As they moved the dot, the image moved in a nearly perfect straight line. The simulation showed a deviation from perfect linearity of only 150 nanometers (nm). That is incredibly small—about the width of a few hundred atoms. It is important to note that this tiny deviation was primarily due to how the computer simulated the light beams (called "ray sampling"), not a flaw in the physical design.
4. The results were consistent: The simulation confirmed that the system could resolve positioning errors smaller than 1 micrometer, a key requirement for the instrument.

Conclusion

This paper doesn't claim to have built the final hardware yet. Instead, it shows that the idea works in a computer simulation. The authors suggest that placing a calibration mask and a special lens right next to the guide probes is a viable way to measure and correct the tiny errors in the robotic arms. By doing this, the HARMONI instrument can ensure that its "robotic eyes" are pointing exactly where they think they are, leading to sharper, more accurate views of the universe. The simulation results match up well with established software, giving the team confidence that this design is ready to be built and tested in the real world.