(OE Magazine - May 01, 2005)
BY GLENN STUTZ, LINCOLN LASER CO.
A polygonal scanner consists of a uni-directional rotating device that incorporates a polygonal mirror, a drive system, and a bearing system. These devices are used to steer a light source or the instantaneous field of view of a pixel. When assembled with the proper light source, optics, modulator, and slow-axis (y-axis) scan mechanism, the polygonal scanner (x-axis scan) can act as a useful data output device in applications like laser projection or marking. When assembled with the appropriate detector, optics, and slow-scan mechanism, the polygonal scanner forms a useful data acquisition device in applications like web inspection or fluorescence microscopy.
Polygonal scanners compete as output devices with galvanometers, acousto-optic deflectors, holographic scanners, and micromirror devices. As with many design decisions, the choice of scanners involves tradeoffs. Polygonal scanners offer the advantages of high line rates, high pixel counts per line, economical pricing, and uniform velocity. Of the available scanner technologies, galvanometers compete most frequently with polygonal scanners (see oemagazine, May 2004, p. 30). The other technologies tend to have niche applications within which they excel; they are not in broad use in many of the markets served by polygonal scanners.
Polygonal scanners dominate markets like prepress, laser printing, film printing, and some medical imaging areas that require high pixel counts and high line rates. Galvanometers excel in applications requiring vector positioning and/or low line rates. Laser marking is an application in which the two technologies overlap. Polygonal scanners tend to be used when the information content is high for a given frame and raster imaging makes sense. Galvanometers tend to be used when the information content is low to medium and vector scanning is quicker.
Data acquisition applications pit the polygonal scanner against technologies like CCD line- and area-array cameras. CCD camera technology has advanced to the point that polygonal scanning is an alternative only for applications in which it can provide distinct advantages over camera technologies, as in the area of high pixel counts. Many web inspection applications require high pixel counts that would need several CCD cameras coaligned, whereas a single laser scanning system could acquire the same data set. Another area where laser scanning excels is 3-D laser imaging, in which a laser pulse probes a target to obtain distance information that is assembled into a 3-D map of the target. Polygonal scanners relay the outbound and return pulse, scanning the scene or target in a raster format.
Another application that uses polygonal scanning for input is fluorescence imaging. Polygonal scanners are well-suited to this application since they can steer an extremely bright source to the target and collect the fluoresced energy efficiently. This fluorescence imaging technique has been used in applications like web inspection, printed-circuit-board inspection, agricultural inspection, and biological imaging.
In a polygonal scanner, a rotating polygonal mirror relays a collimated input beam through a scan lens so that it translates across the image plane. The beam deflects through the beam-feed angle alpha, and as the polygon rotates, it shifts the beam through the active scan angle theta to cover a scan length L.
Specifying the proper polygonal scanner for a given application requires knowledge of certain system requirements. Important information in the design of the motorized scanning polygon includes the duty cycle, operating speed, active scan angle, beam-feed angle, image-plane scan length, scan-lens focal length, spot size, wavelength, and pixel placement accuracies. The formulas needed to size a polygon are straightforward.
1 One would typically start with a desired scan angle theta and duty cycle C, which is the ratio of the active scan time to the total scan time (start-of-scan to start-of-scan for successive scans). The number of facets n is given by
n = 720C/theta
The resulting number of facets can be rounded up, yielding a slightly higher duty cycle, or down, yielding a slightly lower duty cycle.
The incident beam diameter D (measured at the 1/e2 level) to the polygon is needed to complete the polygon sizing. We choose the diameter based on the final spot size required in the image plane, the laser beam quality M2, and the scan lens chosen for the application. Assuming that the beam size has been established, the projected footprint D´ is given by
D´ = 1.5D/cos(alpha/2)
where alpha is the beam-feed angle. The constant 1.5 ensures that the clipping is minimized and virtually of all the energy from a Gaussian beam is reflected from the surface of the polygon.
We calculate the facet length L as L = D´/(1 - C). The polygon's approximate inscribed diameter is
Diam = L/Tan(180/n)
Once the polygon has been sized, we have a few options if the polygon is too large. Reducing the facet length will increase vignetting at the edge of the scan line but may be a good tradeoff depending on the application. Reducing the duty cycle will allow the use of fewer facets and a smaller diameter at the expense of a higher speed and a faster burst data rate (the data rate during the scan). The time between scans is longer with the lower duty cycle, so the average data rate remains constant. The third option is to reduce the feed angle by as much as possible to reduce the beam footprint on the polygon.
Mirrors used in scanning applications come in every conceivable configuration. The application determines the performance requirements. Those requirements determine the size and shape of the mirror, which steers the design direction of the entire scanner. Facet counts can range from one (referred to as a monogon--one side of a pyramid) to more than 150 facets, although applications with greater than 30 facets are rare. Diameters typically range from 10 to 350 mm, with thicknesses from 2 to 100 mm. Mirrors can be designed so that all of the facets are parallel to the rotation axis (prismatic) or at a common angle to the rotation axis (pyramidal). They can also be fabricated so that each facet has a different tilt angle relative to the rotation axis (irregular pyramidal).
Polygonal mirrors are constructed primarily from diamond-turned aluminum, but glass and beryllium have been used in certain applications. There are instances, however, in which a diamond-turned mirror is not the best choice; for instance, a scatter-sensitive writing application in the UV or violet spectral region may not be able to tolerate the near-angle scatter produced by diamond-turned surfaces. In these cases, a conventionally polished mirror can be used. Aluminum is not an ideal material for conventional polishing because it is easily scratched. One solution is to plate the aluminum polygon with electroless nickel and then polish the nickel surface. This process yields surface roughnesses below 10 Å rms.
Polygonal scanners can be constructed in either a cantilevered or captured design. In captured designs, the rotating spindle is held on two ends with the polygon inboard of the bearings. Other scanners use a configuration in which the polygon is outboard of the bearing system. The choice depends on the weight of the mirror and the operating speed. It is recommended that the captured design approach be used to improve spindle performance at higher speeds.
Understanding Error Sources
If polygonal scanners could be manufactured perfectly, the system designer's job would be much easier. Unfortunately, real-world manufacturing tolerances cause the path of the reflected beam to deviate from the theoretical. Critical fabrication errors that require analysis include dynamic track, jitter, and facet-radius variations. It is critical that the system desi
(OE Magazine - May 01, 2005)