Celebrating Twenty Years of Digital Photonic Alignment

A cost-effective, modular alignment engine is updated

The first digital gradient search technique was developed two decades ago to allow fiber optic devices to be efficiently aligned using the micropositioning devices of the day, which were low in speed, resolution and synchronization capabilities compared to today's piezoelectric nanopositioners.  Systems based on this technology were the earliest of a decade-long wave of offerings targeted at industrial automated alignment applications. 

As the telecom boom crested, we were approached by a leading industrial player to provide an especially cost-effective, robust and flexible alignment platform for coarse/fine alignment of photonic devices such as waveguides and laser diodes.  The desire was for a simple stack of stages based on our NanoCube XYZ nanopositioning stage, which provides 100 microns of travel in three orthogonal axes with 2nm resolution.  The customer specified that the software was to be modular, open-source and based on LabVIEW.

We reviewed this and similar applications, noting challenges like fiber-through-tube package designs and irregular coupling cross-sections.  We decided that comprehensive application coverage together with highly time-efficient throughput was possible with a sequence of two operations:
  1. A space-efficient double-spiral-scan, using motorized long-travel stages, for first-light capture and rough optimization, followed by
  2. An extremely fast raster scan with synchronous data acquisition to compile the transverse coupling cross-section and identify the global maximum.
An advantage of the raster scan approach versus established gradient search techniques was its insensitivity to local maxima in the coupling-cross-section; this option had been unavailable a decade earlier due to the limitations of the motion devices of that day.  Put plainly: since our piezo devices are so fast, why not collect lots of data to localize the global maximum directly rather than inferring the vector to it from the limited data older architectures could provide?

We called the result CyberAligner, and a YouTube video of it in action is still viewable (see below).

Recently this architecture enjoyed many advancements:
  • An upgrade to LabVIEW 2010,
  • Leveraging of the latest I/O capabilities provided by today's multifunction analog/digital hardware, including incredibly fast USB units,
  • Motion code based on PI's GCS General Command Set, allowing any type of motorized stages to be used for first-light seek and coarse alignment: cost-effective stepper-motor, robust DC servo-motor, stiff and stable NEXACT®, swift PILine®... and whatever will come next.
The update virtually doubles CyberAligner's speed from the already blazingly-fast version shown in the video, now allowing a full-field scan-and-align on the order of 250 milliseconds.  The coupling cross-section data can be saved to a local or network drive, providing valuable process and device diagnostics in production.   All-USB configurations are featured, cabling is simplified, and multiple workstations can be run off of one PC.  Source-code, compiled and .dll versions of the modular workstation software are offered.

Today, as the twentieth anniversary of the first digital aligners dawns, the photonics industry is reawakening after years in the doldrums after the telecom bubble popped.  With CyberAligner as part of PI's broad toolkit of solutions including the popular F-206 HexAlign six-degree-of-freedom hexapod alignment microrobot, we look forward to meeting a new generation of device and applications challenges.
PI’s solution to fast SiP alignment automation leverages fab-proven controls and mechanisms as shown in configuration for planar test.

Updated Version for Silicon Photonics
For even higher performance applications such as industrial silicon photonics alignment sub-systems, a new system is now available.  Read more on Fast Optical Alignment for SiP Production

No comments:

Post a Comment