An Efficient and Highly Accurate Technique for Periodic Planar Scanner Calibration with the Antenna Under Test in Situ

Size: px

Start display at page:

Download "An Efficient and Highly Accurate Technique for Periodic Planar Scanner Calibration with the Antenna Under Test in Situ"

Sabina Underwood
6 years ago
Views:

1 An Efficient and Highly Accurate echnique for Periodic Planar Scanner Calibration with the Antenna Under est in Situ Scott Pierce I echnologies 1125 Satellite Boulevard, Suite 100 Suwanee, Georgia sierce@mi-technologies.com arion Baggett I echnologies 1125 Satellite Boulevard, Suite 100 Suwanee, Georgia mbaggett@mi-technologies.com ABSRAC his aer describes the develoment, testing and evaluation of a new, automated system for calibration and AU alignment of a lanar near-field scanner that allows the calibration system to remain in lace during AU measurement and which can be used to suort AU alignment to the scan lane. During scanner calibration, robe aerture osition measurements are made using a tracking laser interferometer, a fixture that ositions the interferometer retro reflector at a recise location relative to the robe aerture and a robe roll axis that maintains the roer orientation between the retro reflector and the interferometer as the robe osition is moved. Aerture scan ath information is used to construct a best-fit scan lane and to define a Cartesian, scanner-based coordinate system. Scan ath data is then used to build a robe osition error ma for each of the three Cartesian coordinates as a function of the nominal osition in the scan lane. hese error mas can be used to imlement software-based corrections (K-corrections) or they may be used for active Z-axis correction during measurements. By using a set of tooling oints on the antenna mount, an AU coordinate system is measured with the interferometer. he system then directs an oerator through a set of AU adjustments that align the AU with the lanar near-field scanner to a desired accuracy. his aer describes the imlementation and testing of the system on an actual lanar scanner and AU test environment, showing the imrovement in effective scanner lanarity. Keywords: Planar, Near-field, Error Correction, Scanner, Interferometer 1.0 Introduction Planar near-field scanning has become a standard method for antenna testing. his method is articularly imortant in testing high frequency, lanar hased-array antennas. At very high frequencies the scan-lane accuracy needed for lanar near-field testing begins to exceed the accuracy that can be achieved strictly through hysical alignment and control of machining tolerances. hus, some secondary correction mechanism is needed to imrove scan lane lanarity and osition accuracy. his secondary correction can take the form of a software correction or it may be achieved through active correction of the osition of the RF robe aerture during testing. In either case, error correction requires a set of mas of errors in the robe aerture osition as a function of robe osition and orientation. In this work, we describe a method for the construction of a set of lanar scanner error mas using a calibration system that integrates a tracking laser interferometer with an I-echnologies microwave measurement and analysis system (Figure 1). he geometry of this calibration system is designed so that error mas can be constructed without moving the antenna under test from its osition in front of the scan lane. as are constructed by making direct measurements of the satial osition of the RF robe aerture. We demonstrate the construction of X-axis, Y-axis and Z-axis error mas for a lanar scanner with a 88 cm. x 88 cm. scan lane. We then demonstrate the use of the Z-axis error ma to erform a hysical correction of the scan lane geometry and comare the toograhy of uncorrected and corrected scan lanes. We also demonstrate the use of the system to align a test antenna to the corrected scan lane.

he Z-axis (which is used for error correction) is a relatively inexensive, commercially available ositioning slide.

2 axis. he lanar scanner uses linear servomotor drives and otical linear encoders in the X and Y-axes. he olarization (χ) axis is a rotary, worm-drive rotary stage driven by a DC-servomotor and an otical rotary encoder. he Z-axis (which is used for error correction) is a relatively inexensive, commercially available ositioning slide. It utilizes a DC-servomotor with an otical, rotary encoder driving a linear ball-screw slide. 2.2 he Coordinate etrology System Figure 1: he test aaratus 2.0 System Architecture Figure 2 shows a schematic reresentation of the measurement system. he major comonents of the system are the I-6900 lanar scanner, the I-2097 icrowave easurement System, the I-3001 Data Acquisition and Analysis Workstation and the coordinate metrology system which utilizes a FARO racker tracking laser interferometer. 2.1 he Planar Scanner In this work we use an I echnologies 6900-Series lanar scanner with scan lane dimensions of 88 cm. in the X (horizontal) axis and 88 cm. in the Y (vertical) FARO racker Laser Beam I 6900 Scanner In order to measure the satial osition of the robe aerture, we utilize a FARO racker tracking laser interferometer, a FARO sherically-mounted retro reflector target (SR) and a fixture that facilitates accurate ositioning of the SR relative to the aerture of the robe (Figure 3). he SR consists of a corner mirror mounted at the center of a steel shere. By comaring the relative ositions of the outgoing and reflected laser beams the racker can be made to follow the osition of the SR as it moves through sace. he SR mounting fixture consists of a air of recisionmachined comonents that attach to the end of the RF robe and hold the SR using a magnetic mount. he relative osition between the SR and the robe aerture is measurable and highly reeatable, thus if the osition of the SR is known the osition of the robe aerture is known to nearly the same accuracy. he advantage of this aroach is that we are very nearly making a direct measurement of the osition of the robe aerture. his is in contrast to less direct measurement aroaches that measure the osition of intermediate oints on the scanner carriages. hese less direct measurement aroaches can fail to detect motion of the robe aerture due to carriage rotations. AU Servo Control Signals he absolute osition accuracy of the FARO racker is a function of distance from the racker laser head and the angles of rotation of the racker rotary axes. In this exeriment the 2σ accuracy of the racker varies across the scan lane from 11 µm. in the area nearest to the racker to 27 µm. in the area farthest from the racker. rigger I-4193S Controller/ PAU GPIB I-3001 Workstation I Figure 2: Schematic Diagram of the easurement System In order for the racker to measure the osition of the SR, the aerture of the SR must oint at the laser head of the racker to within an angle of +/- 20. his means that the lanar scanner control algorithm must rotate the χ axis of the scanner as the SR is moved through the scan lane. Imlementation of this coordinated motion is discussed in further detail in Section 2.3.

3 tracker, reventing loss of tracking during the collection of the grid. At each oint, when the scanner axes have been ositioned, the software generates a trigger that is sent to the racker to trigger a measurement of the osition of the SR. he entire measurement rocess is controlled by a Visual Basic scrit that uses built-in low-level instrument controls for the I-4193S. Scan lane correction verification is erformed by the same scrit with the corrected Z-axis used instead of the raw Z-axis. Figure 3: SR ositioned in the mounting fixture. 2.3 he icrowave easurement System he I-2097 Automated icrowave easurement System consists of the I-1797 Receiver, the I-3100 Signal Source and the I-4193S Position Controller. he I-4193S is actively involved in the scan lane error maing rocess as follows: he I-4193S rovides the control of motion for the X, Y and Z axes of the scanner, along with the χ-axis. he I-4193S can imlement an error correction ma through the use of its coordinated motion feature [1]. A similar ma has been used for three-axis robe correction in a large sherical near-field scanning system [2]. his error ma can be alied real time to the Z as either the Y or X-axis is scanned or steed. he error ma is a simle text file in X, Y and δz loaded into the I-4193S. Once the error ma is loaded, the I-4193S can be commanded to use the raw Z-axis or a corrected Z axis based on X and Y ositions. 2.4 he Data Acquisition and Analysis System he I-3001 Data Acquisition and Analysis Workstation and software rovide the comuter control of scanner error ma data collection and verification oerations. he un-corrected scan lane data is collected as a raster scan in X and Y, where the X and Y ositions are obtained through a sequence of discrete stes. hrough the entire grid, the Z-axis is held at a fixed osition. As the X and Y-axes are steed to a new osition in the grid, the olarization axis is moved to an angle that best oints the SR toward the racker at that articular location. In this way, the SR is always ointed at the 3.0 Error aing he goal of lanar scanner error maing is to ma errors in the Cartesian coordinates of the RF robe aerture osition as a function of the desired aerture osition. hus, we would like to find: x y ( x, y, z, χ ) ( x, y, z, χ ) z ( x, y ) where: = z z x, y, z = deviations from the target Cartesian coordinates x, y, z = target Cartesian coordinates Figure 4 is a flow chart of the scanner calibration and antenna alignment rocess. For the work described in this aer, the 88 cm. x 88 cm. scan lane was samled using a square grid of oints that were saced 2 cm. aart. he choice of samling grid size was based uon an earlier measurement of the uncorrected scan lane. his measurement showed that the scanner exhibits a roughly eriodic error in X that has a eriod of ~10 cm. and an amlitude of ~120 µm. (Figure 5. Note that the lanar scanner was urosely laced in an imerfectly-aligned state for the uroses of this exeriment.) A 2 cm. samling grid is fine enough to ensure that this eriodic error is measured. As shown in Figure 4, error maing and antenna alignment begins with calibration of the racker and mounting of the SR on the RF robe. Once this rocedure is comleted, error maing roceeds as follows: 1. Home the scanner. Using the racker, measure the satial osition of the SR. his oint will be used as the origin of the Cartesian,

4 scanner coordinate system. his is the system in which all error terms will be defined. 2. easure the nominal Y=0 line in order to define the X-axis of the scanner coordinate system. 3. easure a grid of oints across the scan lane: a. ove the robe to one corner of the scan lane. b. easure the satial osition of the SR for the first grid oint. c. Iteratively ste through the grid, measuring the osition of the SR at each grid oint. As the robe moves through the scan lane the χ-axis is rotated so that the SR faces the racker. 4. Using the grid data, find the best-fit scan lane using a least-mean squares algorithm. he normal vector to this scan lane defines the direction of the Z-axis of the scanner coordinate system. 5. a x and y errors as a function of χ: a. Ste the χ-axis through the range of angles needed to kee the SR ointed at the laser during scan lane maing. b. At each χ ste, measure the osition of the racker target. Use these measurements to calculate X(χ), Y(χ), Z(χ). 6. For each grid oint, calculate the comonents of the x, y and z errors that are caused only by translation of the robe across the scan lane. hese are the terms that are used to generate the x,y and z error mas. For a given grid oint these terms are: ( x y ) = x x x ( χ ) ( x y ) = y y y ( χ ) x, y, z ( x where:, y ) = z z z P P ( χ ) x, y, z = measured coordinates of oint x y, z =,, P ( P, ( ) ( ) ) target coordinates of oint x χ y χ z χ = osition errors caused by rotating the olarization ositioner to the angle χ needed to kee the laser target ointed at the laser when the robe is at oint At this oint, the satial orientation of the scan lane has been defined and the error mas have been constructed. Iterate hrough Point Grid Start racker Calibration Attach the SR to the Probe ove to the Center Of the Scan Plane easure Errors Due to Polarization Rotation: δx(χ), δy(χ),δz(χ) ove to a Grid Point easure and Record: x, y, z, x, y, z, Calculate a Best-Fit Scan Plane and Scanner Coordinate System Calculate Error as: δx(x, y, χ), δy(x, y, χ), δz(x, y, χ) Find a Best-Fit Corrected Scan Plane and Corrected Scanner Coordinate System easure AU ooling Points to Establish AU Coordinate System Is Alignment Accurate? Yes No Calibration/Alignment Comleted Aly Rotations and ranslations Needed to Align AU to Corrected Scanner Coordinate System Figure 4: Flow chart of the calibration and alignment rocess In this work, the Z-axis error ma is utilized to imlement active error correction using a linear stage. he effect of this active correction was measured by re-measuring the oint grid with the correction enabled. he best-fit lane to this corrected data was then calculated and used as the basis of a corrected scanner coordinate system. AU alignment was then erformed relative to this corrected system. In order to align the AU, tooling oints that defined an AU coordinate system were measured using the racker. he osition of the tooling oints relative to the antenna geometry were known from revious racker measurements. Once the AU coordinate system was found, the values of the two rotations about the AU system X and Y axes that would make the AU Z-axis arallel with the scanner Z axis were reorted by the racker software. hese two rotations were alied to the AU and the tooling oints were re-measured. his

5 rocess of measurement and adjustment was iterated until the Z axes of the two coordinate systems were within the required recision. In order to make the Z-axes of the two coordinate systems collinear, the relative ositions of the scanner system origin and the AU system origin were calculated and these values were entered as offsets in the I-4193S controller. his had the effect of shifting the scanner system into alignment with the AU system without moving the AU. corrected scan lane. We then used the racker to align an AU to this corrected scanner coordinate system. 4.0 Results Figure 5 shows the toograhy of the scan lane before the imlementation of active Z-axis correction. he lane has a eriodic error across the X-axis that has a eriod of ~10 cm. and an amlitude of ~120 µm. In addition to this eriodic error the nominally lanar surface has some curvature about an axis that runs diagonally across the lane. his curvature is caused by a slight arallelism error between the front and rear X-axis rails. he absolute Z-axis error of the uncorrected scan lane is +170/-93 µm. and the RS error is 40 µm. Figure 6 shows the toograhy of the scan lane after the imlementation of active Z-axis correction. Both the eriodic error and the overall curvature error have been reduced to levels that are near the limits of the racker measurement resolution. he absolute Z-axis error of the corrected scan lane is +35/-37 µm. and the RS lanarity is 10 µm. Figure 5: oograhy of the uncorrected scan lane 5.0 Summary In this aer we have described a lanar scanner error maing and calibration system that is based on the integration of an I-2097 Automated icrowave easurement System with a FARO racker tracking laser interferometer. We have shown that this integrated system can be used to make direct measurements of the osition of the aerture of an RF robe as it is moved through the scan lane. We have used these measurements to construct error mas of the aerture osition. In order to utilize the ma of lanarity errors, we imlemented active Z-axis correction using a relatively inexensive linear slide and the coordinated-motion caability of the I-4193S motion controller. Using this error correction scheme we were able to imrove the RS lanarity of the scan lane by a factor of four. An advantage of our aroach is that error maing can be erformed with the AU in lace and the racker can then be used to align the AU. In this aer we demonstrated the use of the measurement system to establish a Cartesian coordinate system based on the Figure 6: oograhy of the corrected scan lane 6.0 REFERENCES [1] Alications for Coordinated otion in Radome esting, cbride, et. al., AA Symosium 2002, g [2] Imlementation of a Geometric-Error Correction System for Extremely High Probe Position Accuracy in Sherical Near-Field Scanning, Pierce, R.S. and Langston, J., AA Symosium 2004 (Acceted for Publication).

6 7.0 ACKNOWLEDGENS he authors wish to recognize Dean Darsey, Becky Kaffezakis, Jim Langston, Scott cbride and Don Weiss, all of I echnologies, for their assistance in the develoment and oeration of this rocedure.

Optics Alignment. Abstract

Optics Alignment. Abstract Otics Alignment Yi Qiang Duke University, Durham, NC 27708 Xiaohui Zhan Massachusetts Institute of Technology, Cambridge, MA 02139 (Dated: July 9, 2010) Abstract yqiang@jlab.org zhanxh@jlab.org 1 I. INTRODUCTION