What software and files do I need to operate my deformable mirror?
There are three software items required to properly operate your deformable mirror; including two configuration files, specific to the device you have, that should be copied to their respective folders within the Boston Micromachines software folder.
The Boston Micromachines Deformable Mirror Software Development Kit (DMSDK) provides a common interface to all BMC products. It allows users to write one code base that can be used with any product. Command line and graphical applications are provided for demonstration of the DMSDK and hardware functionality. A C API is provided for versatile development of high performance applications.
Adaptive Optics (AO) Software Development Kit (SDK) available as an upgrade. Find out more about our AOSDK.
A configuration file that is matched to a unique DM-Driver combination. A profile contains information such as deformable mirror array size, actuator mapping, voltage limits, the driver PC interface type, and other settings. Profile file names contain the serial number of the deformable mirror device to which they apply. Contact us if you have any questions regarding your DM profile.
Flat map file
A type of Command Map that drives the deformable mirror surface to a plane that is parallel to the silicon wafer base. This file is optional and useful for calibration and alignment of your optical component.
Flat maps for deformable mirrors are established by Boston Micromachines and are unique and custom to only that DM-Driver combination. Boston Micromachines typically supplies a flat map with each deformable mirror at 50% deflection for use in establishing a mirror surface for optical alignment and calibration. Other flat map files are also available — please inquire for more details.
How do deformable mirror actuators work?
The concept behind MEMS deformable mirrors is pretty simple. This simple design enables them to be fast and efficient devices.
There is an electrode layer, an actuator layer and a mirror layer.
You ground the actuator layer. When a voltage is put on an electrode, the actuator layer above it is attracted and it pulls the mirror layer down. Voila! A deformable mirror. The motion for every device is pretty much the same.
The interesting thing about deformable mirrors is that there are many ways that you can fabricate the mirror layer and therefore many types of mirrors that you can make. If you want a big, smooth surface, simply make the mirror layer continuous across the entire mirror. If you want each actuator to control a small region that is independent of the others, a segmented mirror is for you. And, if you want larger segments which are controlled by multiple actuators, just divide the surface up the way you want to and you can have separate mirrors that can be controlled with a number of actuators, like our Hex-TTP (Tip, tilt, piston) version that we offer.
In the past, versions of various sizes and segments has been created. And, if you want something more esoteric, we’re happy to entertain your ideas.
Some of the common configurations that Boston Micromachines has created for applications from astronomy to laser communication:
Continuous Deformable Mirror Surface
Segmented Deformable Mirror Surface
Tip-Tilt Deformable Mirror Surface
What does my deformable mirror surface layout look like?
MEMS deformable mirrors are typically fabricated in a cartesian layout format. This is simple to layout and stays consistent from the smallest to the largest sizes. That is where the layouts stop being similar.
Within the cartesian layout, there are two types of mirrors:
Continuous Deformable Mirror
A Deformable Mirror where the reflective mirror surface extends continuously (without any breaks) over the entire actuator array. In Continuous DMs, actuator arrays can be linear, square grid, or other type of pattern.
Typical configurations (actuator maps):
Multi Deformable Mirror
Multi-C Deformable Mirror
Kilo-C Deformable Mirror
Segmented Deformable Mirror
A Deformable Mirror where the reflective mirror surface is segmented and the number of mirror segments is equal to or less than the number of actuators in the entire actuator array. In Segmented DMs, actuator arrays can be linear, square grid, or other type of pattern. Coupling between the mirror surfaces of adjacent segments is zero. Mirror segments in Segmented DMs can be controlled by one or more actuators.
- With one actuator per mirror segment, the segment can move in a single direction.
- With two actuators per mirror segment, the segment can move in a single direction and tilt in one plane.
- With three actuators per mirror segment, the segment can move in a single direction and tilt in any plane.*
More information about standard models can be found on the Standard Deformable Mirrors page.
*Certain type of segmented mirror that many have found useful across an array of applications: The Hex-TTP deformable mirror.
Hex-TTP Deformable Mirror
Breaking up the surface into hexagonal segments, with three actuators controlling each, enables tip-tilt-piston (TTP) functionality and provides the user with the ability to achieve precise angular and phase control with each optical segment. Typical sizes of mirrors are shown below.
Hex-111 Deformable Mirror
Hex-1011 Deformable Mirror
Hex-3K Deformable Mirror
Boston Micromachines has published a whitepaper on the architecture. It can be accessed on our Hex Class Deformable Mirrors page.
How big is my deformable mirror and how much of the mirror surface is actually useful?
Deformable mirror surfaces are classified by Boston Micromachines in three different ways: Full Aperture, Active Aperture and Recommended Optical Aperture. Below we will describe these different apertures for both continuous and segmented deformable mirrors.
Full Aperture (FA)
The largest circular region that fits within the deformable mirror’s reflective surface. The Full Aperture is what you see when looking directly at the deformable mirror, and is also known as the Unpowered Surface Figure. As it is not entirely controllable, the Full Aperture is typically not used, but is useful to know if the application calls for overfilling the useful area.
Full Aperture For Continuous Deformable Mirrors
Continuous deformable mirrors have two or more rows of inactive actuators outside of the controllable actuators. The reflective area extends to the outermost inactive actuators.
The diameter of the Full Aperture is [(actuators across) + 3] x (pitch)
Full Aperture (red circle) overlaid on the actuator map
Surface scan of the Full Aperture
Active Aperture (AA)
The largest circular region that fits within the displaceable controllable area of the DM surface. The Active Aperture is a circle inscribed within the outer controllable actuators. The area within the Active Aperture is displaceable but not fully controllable at the edges, because the controllable outer actuators are coupled to actuators further out that cannot be controlled.
The diameter of the Active Aperture is ([actuators across] -1) x [pitch]
Active Aperture (red circle) overlaid on the actuator map
Surface scan of the flattened Active Aperture
Recommended Optical Aperture (OA)
The largest circular region that fits within the displaceable and fully controllable area of the deformable mirror surface.
For continuous DMs, the Recommended Optical Aperture is a circle inscribed within the next actuator in from the outer controllable actuators. Like the Active Aperture, the area within the Recommended Optical Aperture is displaceable but is also fully controllable at the edges, because the controllable outer actuators are coupled to actuators further out that can also be controlled.
The diameter of the Recommended Optical Aperture is ([actuators across] -3) x [pitch]
Recommended Optical Aperture (red circle) overlaid on the actuator map
Surface scan of the flattened Recommended Optical Aperture
How fast are deformable mirrors?
There are a few areas that need to be addressed with respect to the speed of a MEMS based deformable mirror. The best way to do this is by separating out the different forms of latency/delay in the system.
We define the latency of our electronics as the time between the first command written to the last DAC (digital to analog converter) updated. In the case of our USB drivers, the data is sent over the USB protocol in 125us packets which leads to the calculation of 8kHz frame rate.
For our X/S-Driver architectures, much higher frame rates of up to 400kHz are possible, depending on the number of actuators addressed. This is the first source of delay with deformable mirror systems.
Amplifier slew rate
The next delay in the system involves the slew rate of our amplifiers. For most of our devices, we utilize a custom Supertex amplifier that slews at 5.1V/us. Therefore, this is completely dependent on the size of the step taken for a given command. This should be added to the electronics latency when evaluating the system. For our X-Driver architecture, we utilize improved amplifiers that yield a slew rate of 11.6V/us, which is what makes the higher speeds possible.
Actuator Mechanical response
The final consideration when calculating the system latency is the mechanical response time of the DM. For some users, this is not a consideration as they are more concerned with sending commands as quickly as possible and are not worried about the position of the DM. However, for others, this is important since they would like to detect a settled wavefront for analysis in the control loop. The mechanical response is summarized in the documents below for each of our DM architectures.
Finally, the entire surface of the mirror must be taken into account when considering how fast your system can go. The major concern here is air damping. While individual actuators can move with sub-100 microsecond speed, as the air trapped under the surface moves around, it can affect the settling time of the entire surface. The surface response time for continuous deformable mirrors is on the order of 150 to 300 microseconds, depending on the stroke and spacing of the actuators.
How stable is my deformable mirror?
Boston Micromachines deformable mirrors exhibit highly stable operation with no hysteresis and no measurable drift, down to the sub-nanometer level. Our MEMS based deformable mirrors have been extensively tested for various parameters, including stability. We currently have a 4096 actuator count deformable mirror integrated into the Gemini Planetary Imager (GPI). Before its installation, MEMS stability was tested by applying a flattened shape to the device with successive wavefront measurements taken every 38 seconds, for 60 iterations. Short term stability was measured over 9 minute intervals, within the long term stability test.
The shorter time scale is comparable to typical closed loop operation times. Stability was measured to be 0.08 nm RMS phase by replacing the MEMS with a flat mirror. Average long term stability of the MEMS was measured as 0.16 nm RMS phase. On the shorter time scale, the system is more stable with an average RMS deviation of 0.13 nm RMS phase for the MEMS, and 0.07 nm RMS phase for the flat mirror. Previous tests had indicated less stability because of errors produced by the MEMS drive electronics that have since been corrected. The figure below is a curve of growth showing that most of the actuators are quite stable. Of the 500 actuators tested, there was 97% stability of better than 0.16 nm RMS over 38 minutes.
Curve of growth for stability data (standard deviation of surface over 60 measurements taken in 38 minutes)
Inter-Actuator Coupling and Influence Function
The type of deformable mirror you have affects how actuators behave with one another. Powered actuators have an influence on their neighboring actuators in certain type of deformable mirrors.
Segmented deformable mirrors
The actual displacement is determined only by the magnitude of the command sent to the actuator. By their nature, segmented deformable mirrors have no inter-actuator coupling.
Continuous deformable mirrors
The actual displacement is determined by:
The magnitude of the command sent to the actuator.
An offset caused by inter-actuator coupling.
The magnitude of this offset is related to the stroke, the actuator spacing, and the relative displacement of adjacent actuators.
As the mirror surface that is directly controlled by an individual actuator is attached to the mirror surfaces over adjacent actuators, the position of those adjacent surfaces negates the full effect of the actuator. The closer the surrounding surfaces are to the actuator’s driven position the greater the coupling.
While quantifying inter-actuator coupling becomes complicated when the adjacent mirror surfaces are at differing positions, it is relatively straightforward when the adjacent surfaces are un-powered.
The BMC deformable mirror architecture allows for local deformation of the mirror membrane with an influence function from 11-25% depending on the specific device design. Influence function, defined as the ratio of the deflection induced on a neighbor of a powered actuator and the maximum deflection of the powered actuator, with all other actuators at half bias, for the three different deformable mirror actuator types is shown below.
These profile measurements of the deformable mirror surfaces at half bias with an actuator pulled down and poked up show the influence functions of these devices to be 15%, 13%, and 22% for the 1.5, 3.5, and 5.5μm stroke deformable mirror actuator types respectively.
Deformable mirror actuators at half bias
with single actuators at 0 and maximum command, respectively
Influence function of a 1.5 μm stroke deformable mirror
Influence function of a 3.5 μm stroke deformable mirror
Influence function of a 5.5 μm stroke deformable mirror
Deflection and pitch
For BMC deformable mirrors, deflection is the distance moved by an actuator from its unpowered state. Due to the architecture, actuators are pulled down when voltage is applied, which is why deflection (height) is represented below in negative values. On the other hand, pitch is the distance between each actuator within a deformable mirror.
Deformable mirror with every other row actuated at maximum command
Maximum inter-actuator stroke of a 1.5 μm stroke deformable mirror
Maximum inter-actuator stroke of a 3.5 μm stroke deformable mirror
Maximum inter-actuator stroke of a 5.5 μm stroke deformable mirror
Open Loop Control
Boston Micromachines offers open loop calibration for your deformable mirror as an upgrade. This allows you to precisely control the position of your mirror, without the need for a feedback sensor and control algorithm. We provide calibration files specific to your device that allow you to control your deformable mirror without feedback. This is useful in a variety of applications, such as high-resolution microscopy and laser communication.
Due to the method used for creating calibration files, currently we offer open loop calibration for deformable mirrors with up to 492 actuators. Contact us if you would like to purchase the open loop calibration upgrade for your deformable mirror, or if you have any questions about the process.
Open loop calibration file
A calibration file contains the exact voltages necessary for a series of specific actuator deflections over the range of 0 to 1.0 (full stroke). For continuous deformable mirrors, at each deflection the position of the surrounding 8 actuators is also recorded.
Calibration files are used in open loop applications where there is no feedback on actuator deflection. For a specific mirror shape, i.e., an array of desired actuator deflections, the calibration file contains the actuator voltages necessary to drive the deformable mirror to that shape by taking into account both the position of each actuator, and the influence of the surrounding actuators. For deflections between those in the calibration file, the necessary voltages can be interpolated.
The Calibration File is created by Boston Micromachines through interferometric analysis of a single representative actuator and is valid for only that DM-Driver combination. Calibration file names contain the serial number of the DM device to which they apply and have file name extension .mat; these files are stored in a folder within Program Files\Boston Micromachines\Calibration. As installed, the Calibration folder contains a few generic calibration files for demonstrative purposes.