A two-axis micropositioner for optical fiber alignment - internal construction of HP 71450A and 71451A optical spectrum analyzers - Technical

Hewlett-Packard Journal, Dec, 1993 by J. Douglas Knight, Joseph N. West

A positioning system with submicron resolution is used to keep the output fiber accurately aligned with the light coming out of the monochromator during movement of the diffraction grating.

The double-pass monochromator design used in the HP 71450A and 71451A optical spectrum analyzers offers a number of performance advantages over competing monochromator designs. Several of these performance advantages come from the secondary filtering effects of the optical fiber used at the output of the second pass of the monochromator. The limited cross-sectional area and limited numerical aperture of the fiber help reject stray light, giving good dynamic range performance and good spurious response rejection. Coupling the light into fiber also allows the use of a small, low-noise photodetector which results in excellent sensitivity even with rapid sweep speeds and minimal video filtering. In addition, going into fiber allows the instrument to have an optical monochromator output that offers both fixed-wavelength and swept-wavelength modes of operation with a full range of resolution bandwidths selectable by the user. These advantages are significant, but designing a system to keep the output fiber accurately aligned with the light coming out of the monochromator during sweeps proved to be a considerable design challenge.

The Positionlag Problem

Ideally, in a perfectly symmetric double-pass monochromator, the spot of light at the output of the second pass would not move. However, in reality, as the diffraction grating rotates and the instrument sweeps in wavelength, the spot of light at the output of the monochromator does move slightly in two dimensions. It is therefore necessary to track the moving spot with the output fiber to capture the light completely and to realize the desired filtering effects.

In the dispersion direction (y-axis in our implementation) the movement of the output spot is the result of asymmetry in the system. The second pass is farther off the axis of the lens than the first pass. This asymmetry is necessary to avoid picking up light from the first pass with the photodetector. The movement in the y-axis is predicted by theory and is consistent from unit to unit.

In the nondispersion direction (x-axis), the movement of the output spot is the result of manufacturing tolerances that cause the lines of the diffraction grating not to be perfectly parallel to the axis of rotation of the diffraction grating (see Fig. 1). Parallelism and perpendicularity of critical parts in the grating rotator assembly are aligned to very close tolerances. Even so, tens of micrometers of beam movement still result. If the lines of the grating are not perfectly parallel to the axis of rotation, the lines of the grating will precess* about the axis of rotation as the grating rotates, causing the output spot to move in the x-axis. For a given monochromator this movement is very repeatable and tracking is possible with a precise positionlag device.

Tracking the Output Spot

Once it was determined that a micropositioning device was necessary to track the output spot, other advantages of having such a device were envisioned. One advantage is noise and stray light cancellation. Most systems that attempt to do noise cancellation chop the optical signal with an apettymire that alternately passes or blocks the light to the detector. When the light to the detector is blocked it is possible to measure the electrical noise of the detection system which can be subtracted from the reading obtained when the aperture passes light to the detector.

When the aperture passes light to the detector, the output of the detection system represents signal stray light electrical noise. Subtracting electrical noise leaves signal stray light.

With a fast and accurate micropositioner at the output of the monochromator it is possible to perform another kind of optical chopping to remove both electrical noise and stray light. If we assume that the stray light is relatively uniform in the region around the output spot as it might be in the case of scatter from optical components and diffuse reflections from the inside of the monochromator cavity, physically moving the output fiber laterally away from the output beam would allow a measurement to be made of stray light electrical noise. Alternately moving the output fiber into and out of the output beam allows the stray light electrical noise term to be subtracted from the signal stray light electrical noise term leaving only the signal value. The digital signal processor described in the article on page 68 controls the micropositioner and performs the subtraction of stray light and electrical noise from the measurement. This mode is activated automatically in the HP 71450A and 71451A analyzers when the user requests a very sensitive setting that results in a sweep time greater than 40 seconds.

Having an electrically actuated micropositioner at the output of the monochromator also eliminates the need for the user to make manual adjustments to the second-pass aperture (output fiber) relative to the first-pass aperture (slit) to maintain signal symmetry or to adjust the optical output of the HP 71451A for fiber-in/fiber-out measurements.* Most optical spectrum analyzers that have double monochromatots require the user to adjust the optical output for maximum signal strength at a given wavelength with manual micrometers. With the HP optical spectrum analyzers, output coupling is automatically maintained over the entire wavelength range. If the instrument is dropped or experiences significant changes in temperature, there is an AUTO ALIGN key on the front panel of the instrument that the user can push to initiate an alignment routine to ensure that optimum output coupling is reestablished.