Adaptive deformable mirror

fitting error

residual wavefront error that remains because the deformable mirror, with its finite number of discrete actuators, can only approximate the true shape of the distorted wavefront

temporal error

residual error that remains because the atmosphere keeps changing while the AO system takes a finite amount of time to measure and correct it

deformable mirror design

→ large number of actuators, high control bandwidth, low total power dissipation

deformable membrane

→ low mass, low out of plane stiffness, low actuator force, low power dissipation

influence function

how much the membrane deforms in response to a single actuator being pushed

coupling

relation between mirror membrane stiffness and actuator stiffness

• lower coupling == good (more control)

actuator plate goals

• low power

• extendable to large grids

• no hard points if one should fail

actuator plate - concept 1

→ reluctance actuator

magnetic force=srping force

• different => net residual parasitic stiffnes that needs to be balanced

(non) linear stiffness of the actuator membrane with uniform thickness

• linear stiffness has cubic dependence on thickness

• non-linear term due to non-linear force displacement relationship

→ tune unifrom thickness to manipulate force-displacement

Explain the saturation problem in cocept 1.

To avoid saturation: increase the inital gap and change magnetic material. Improvement, but only a few suitable combinations and combinations are very sensitive.

→ a non-saturated magentic design will result in high parasitic stiffness and hig I_rms and P_rms

→ a mechanical design with low parasitic stiffness will result in a saturated magnetic circuit

Solution: separation of the magnetic circuit and actuator stiffness in a new actuator design

actuator plate - concept 2

→ variable reluctance actuator - the coild current can add or subtract the permanent magnet’s force, modulating the net force on the moving ferromagnetic core, separated membrane suspension and ferromagnetic core

membrane stiffness measurement

• measurement of the non-linear stiffness of the membrane suspension

• measurement of the dynamic properties of the actuator

*mismatched between measured and predicted f:

1. stiffness too low

2. magnetic force lower than expected

measurement of the PM properties

with a Helmholts coil setup couple to a commercial flux matter

• fast and reasonably accurate

• no desctructive external magnetic field applies

• measurement system and measurements are low cost

→ 2 coils in series, magnet causes a flux change, output depends on volume, shape and relative permeability

actuator measurements

• measurement of each actuator in the grid (f, F, c, motor constat, I, R as a funciton of the position)

• spreading in the grid

→ are all actuator consistent with each other

mirror actuator connection

through rods

Is the mirror usable? - raw data correction

1. correct for the low spatial frequencies

2. fitting with the analyticla influence function

3. end up with the non-correctable error