Customer Solutions
NASA Stabilizes a Fabry-Perot Interferometer with LabVIEW and IMAQ
Author(s):
Kristie A. Elam, Akima Corporation, NASA Glenn Research Center
Industry:
Aerospace/Avionics
Product:
Data Acquisition, LabVIEW
The Challenge:
Developing a more user-friendly, automated control system that can stabilize, maintain, and monitor the alignment of parallel mirrors in a Fabry-Perot interferometer.
The Solution:
Using LabVIEW and IMAQ Vision to develop a program that adjusts for thermal, mechanical, and laser frequency drifts to maintain the mirror alignment of a Fabry-Perot interferometer over a long time with minimal user input.
NI LabVIEW and IMAQ program monitors Fabry-Perot interferometer in automated control system.
|
Previous Stabilization Methods
We use the Fabry-Perot interferometer (FPI) to analyze scattered light in jet noise studies at NASA Glenn Research Center. The FPI is an optical instrument that works on the principle of light interference between two highly polished, flat mirrors, which, when perfectly aligned and parallel, form interference fringes. The FPI is highly sensitive to room conditions, particularly temperature and vibration, which can cause the mirror alignment to change. When we use an FPI in a laboratory environment, it is crucial to maintain mirror alignment for accurate measurements.
Prior to the automated system, scientists used several methods to align the FPI to obtain sharp fringes. The first method involves manual adjustment of three knobs on the interferometer. This is both a tedious and time-consuming process that requires hours of practice before one is able to master it. This method also is not suitable in an experimental environment because continuous “tweaking” of the knobs is necessary due to the FPI’s sensitive nature. A second method involves using a ramp generator available from the FPI manufacturer, which also requires manual adjustment of three knobs. This adjusts mirror alignment by applying a voltage to three piezoelectric actuators (PZT) mounted on one of the mirrors. This is also tedious and does not hold the alignment, so it requires readjustments during a test run. The third method, a stabilization system available from the manufacturer, does not lock onto a finite diameter fringe, and is not readily incorporated into our automated data acquisition system. Because of this, we developed a DOS program that utilized a fuzzy logic control algorithm. However, this program was sensitive to light intensity and required two separate computers for stabilization and data acquisition.
Implementing the LabVIEW Program “Fabry Stabilize”
For our experiment, we passed uniform laser light through the interferometer to three prisms, which direct light from three regions on the mirrors, giving a set of fringes. The fringe images are acquired using an NI PCI-1407 frame grabber board and CCD camera. We used three analog outputs of an NI PCI-6703 D/A board to control the high voltages (HV) applied to the three PZTs.
Before running our program, we inputted parameters relating to the interferometer and the experimental setup, as well as the desired fringe target radius, in pixels on the parameter tab. We then chose to either load saved PZT HV values or get new values. If the FPI was recently aligned, we loaded the last saved values. If not, the subVI “Fabry HV” opened, and we adjusted the fringe radii slightly by mimicking the three-knob manual technique as we displayed a continuous image of the three fringes using the “IMAQ Grab Acquire” VI.
When the fringe radii were approximately equal, we clicked the “Get Centers” button on the mode tab, which stopped the grab acquire and snapped a single image. We then clicked in the centers of the three fringes to acquire reference center coordinates.
With the Vision VI “IMAQ Line Profile,” we obtained horizontal and vertical lines measuring 90 pixels from the reference coordinates for each fringe. We fed these into a nonlinear Levenberg-Marquardt fitting routine using an Airy function to obtain the best-fit coefficients for the horizontal and vertical fringe profiles. We then averaged the radii to obtain a best-fit radius for each fringe. We plotted a plot of the actual fringe profiles and the fit profiles on the program’s front panel.
We used these radii, as well as the current piezoelectric voltages and target radius, as inputs to a control algorithm. The algorithm compared the current radii to the target radius and determined the adjustment needed in the PZT voltages. The NI PCI-6703 D/A board used the “AO Update Channel” VI to apply new voltages to the three PZTs. A new image was then snapped to obtain the latest profiles and the sequence was repeated within a continuous loop until the program was stopped. The voltages were adjusted three to four times per second, thus enabling the interferometer to maintain alignment. At this point, the program could run for hours without any user input.
As the program ran, it monitored the difference between the current radii and the target radius and reported a status – locked, active, sleeping, or control loops off. Locked status occurred when the fringe radii were within the minimum tolerance when compared to the target radius. In other words, it was in alignment. Active status occurred while the program was actively seeking to reach the target radius. We used sleeping status while acquiring data. Using the global variable “Fabry Mode,” we incorporated the stabilization program into our National Instruments LabVIEW data acquisition program. While stabilizing, the global variable had the value “Change.” When the data acquisition program started, it put the stabilization program to “sleep” by sending the value “No Change,” which essentially disabled the nonlinear fit routine. After data acquisition, we “awakened” the stabilization program. The control loops off status occurred when we disabled the control algorithm so that the radii was allowed to drift. We used this for analysis purposes.
The program also logged information to a file for analysis purposes. Date, actual time, elapsed time, radii (average of the horizontal and vertical line profiles) for each of the three fringes, PZT voltages, and status mode were recorded. Two tabs showed running history charts of the three radii and the three PZT voltages to allow for quick viewing of alignment maintenance, while another gave current values from the fits.
Using LabVIEW and NI hardware, we developed a program that simplified the tedious task of FPI stabilization by maintaining alignment for long periods of time. With this new system, we were able to maintain mirror alignment and control the fringe radius to within 0.2 pixels, which led to more accurate data collection.
For more information, contact:
Kristie A. Elam
NASA John H. Glenn Research Center
Lewis Field
21000 Brookpark Rd.
Cleveland, OH 44135
Phone: (216) 433-4000
Web: www.grc.nasa.gov
