How to Build an Automated 3-Axis Fiber Alignment System for $13K

Aligning optical fibers with non-motorized positioning stages can be tedious and requires exceptional precision. Single-mode fiber cores are often just 1–9 μm in diameter, with a Mode Field Diameter (MFD) of 3–10 μm, making precise alignment particularly challenging. Researchers are bound to spend significant time adjusting thumbscrews, checking the power meter, and repeating the process across multiple axes. Even after achieving alignment, vibration or temperature changes can shift the fibers and require the process to be repeated.

Automating this process can eliminate most of this manual work and keep errors introduced by manual adjustment low. However, automated fiber alignment systems can seem expensive and complicated to build. We wanted to see whether a practical research system could be built using readily available components and straightforward software.

The result is a 3-axis automated fiber alignment system with approximately $13k USD in motion hardware. This guide shows how we built it, the components we used, and how to automate the alignment process with a pre-developed Python script that can be adapted for your own needs.

Automated Fiber Alignment System Requirements

The goal is to create a practical reference design that any research lab could use as a reference to build and adapt for their experiments. We focused on three priorities:

• Sub-micron positioning: Choosing devices with highly precise positioning capabilities is needed for effective single-mode fiber alignment, along with enough travel range for initial coarse alignment.
• Cost-efficient motion: Selecting stages that meet the required positioning performance without paying for capabilities beyond what’s needed.
• Minimal integration effort: Utilizing stages that have a straightforward hardware setup and providing a sample script that researchers can adapt rather than developing an automation system from scratch.

These priorities guided our selection of the motion hardware, controller, and software used in the system.

The Solution: Automated Single Mode Fiber Alignment for $13k

The final system combines three motorized direct-drive linear stages, an optical power meter, and an open-source Python script to automate single-mode fiber alignment. The complete motion system costs approximately $13k USD and is built entirely from modular, off-the-shelf stages.

Key Results

• 20 seconds throughput: Reduces average alignment time from approximately 15 minutes of manual trial-and-error to around 20 seconds, depending on the initial alignment and search parameters.
• Repeatable Results: Every fiber is aligned using the same algorithmic workflow, eliminating operator-dependent variability.
• An affordable system anyone can copy and use as a starting point: Built entirely from modular off-the-shelf components ($13k) with a ready-to-run Python example (free!).

How the Automated Alignment System Works

The system uses three motorized linear stages to position one fiber relative to another. Rather than requiring an operator to manually adjust each axis while watching the power meter, the software continuously monitors transmitted optical power and iteratively adjusts the stage positions until it finds the position that maximizes coupling efficiency.

The overall workflow consists of five steps:

1. Mount and roughly align the fibers
Secure the input and output fibers in their mounts and bring the fiber end faces into approximate alignment. The automated routine performs the fine alignment, so only a coarse starting position is required.

2. Start the alignment routine
Launch the provided python script, which communicates with the motion controller and optical power meter.

3. Find first light
The stages perform a coarse spiral search until optical power is detected, indicating that the fiber cores are close enough for light to couple.

4. Optimize the alignment
Once first light is found, the alignment routine progressively refines the position while continuously monitoring optical power. The search converges on the position that maximizes transmitted power.

5. Complete the alignment
When the measured power reaches its maximum, the routine stops and the stages remain at the optimal position. If coupling efficiency later decreases because of drift caused by temperature changes or vibration, the script can be run again to automatically restore alignment.

Building the System: Step by Step Instructions

Motion Hardware List

To meet our system requirements, we selected three Zaber DMA direct-drive linear stages driven by an X-MCC3 controller. Together, they provide the positioning precision, and travel range needed for automated single-mode fiber alignment.

• 10 nm minimum incremental move to provide the sub-micron positioning required for coupling single-mode fibers.

Note: For less demanding multi-mode fiber alignment, a stack of three LSM025A-T4A stages would likely provide adequate performance at an even lower price point.

• 15 ms typical move and settle time to enable the rapid execution of search algorithms by minimizing the time spent between successive measurement points.

• 50 nm repeatability to ensure the system can reliably return to previously identified peak-coupling positions for consistent, repeatable alignments.

Figure 1. Diagram illustrating key motion hardware components.

Pro Tip: Zaber multi-axis systems are composed of individual modular stages, allowing labs to purchase axes incrementally rather than as a single large system. This flexibility can make it easier to fit within annual equipment budgets or institutional purchasing workflows while still arriving at the same final system. If you no longer need this system, you can take it apart and use the individual stages for other projects.

Optical Components List

Here’s an overview of what we ended up choosing for our system:

Note: The optical components shown here were selected for a simple demonstration system. Choose components appropriate for your wavelength, fiber type, and other application requirements.

Figure 2. Diagram illustrating key optical components.

Hardware Assembly Instructions

Building the alignment platform takes less than an hour and requires no specialized tools. The assembly process is divided into four straightforward steps:

Step 1: Assemble the 3-Axis Motion System

1. X Axis: Mount the first DMA stage to the optical breadboard using an AP279 adaptor plate.

2. Y Axis: Secure the second DMA stage orthogonally on top of the base stage using a second AP279 plate to complete the XY motion plane.

3. Z Axis: Fasten the AB280 angle bracket to the Y stage, then mount the third DMA stage vertically to the bracket.

4. Cabling & Controls: Connect each stage to the X-MCC3 controller using the MC10L060 power/data cables, and plug the controller into your computer via the U-DC06 USB cable.

Full assembly instructions can be found in the multi-axis DMA manual here.

Step 2: Mount the Optomechanics and Fiber Holders

1. Attach the AB106 bracket to the vertical stage, then secure the second FC ferrule clamp to serve as the motorized fiber holder.

2. Thread an optical post into the breadboard directly opposite the motion system with a FC ferrule clamp ferrule clamp attached to hold the target fiber.

3. Mount the power meter to the breadboard (optional).

Step 3: Integrate Optical Power Meter Feedback

1. Connect the PM101 power meter to the X-MCC3 controller via the analog inputs.

Note: Connecting the power meter to the controller's analog input instead of directly to the computer allows the controller to sample the analog signal at precise positions while the stages are moving. This enables streamed scans, where stage position and optical power are monitored simultaneously by the controller without stopping at each measurement point.

2. Attach the power sensor to the power meter.

Step 4: Connect the Laser Source & Mount the Fibers

1. Secure the fiber optic patch cables into the FC ferrule clamp mounts on each stage.

2. Attach the pigtailed laser diode to the input fiber using the FC mating sleeve.

3. Plug the laser power supply into a standard wall outlet.

Software Setup Instructions

Before running the alignment routine, we need to configure the motion system in Zaber Launcher. This setup process involves connecting the hardware, homing the stages, and tuning the servo controller for the payload carried by each axis. Once completed, the system is ready to run the Python alignment script.

Step 1: Download Zaber Launcher

Start by downloading our free Zaber Launcher software from here. This application is primarily used for initial device setup, testing individual movements, and low-code or no-code control.

Step 2: Set up Device Connection

1. Connect the 3-Axis motion system to the computer via USB.

2. Add a New Connection in the Zaber Launcher application, as shown in the tutorial here. Zaber Launcher automatically detects connected devices, so each DMA stage plus the controller will appear in the device list without any manual configuration.

Step 3: Servo Tuning

The DMA uses direct-drive linear servo motors, which require servo tuning to optimize the control loop for the installed payload. Proper tuning ensures stable, accurate stage motion and positioning performance. Open the Servo Tuning tab in Zaber Launcher and enter the total payload carried by each stage, including any mounted stages, adapters, fiber mounts, and optical components. The controller uses this information to automatically calculate the appropriate control loop gains for the system. Click here to learn more about servo tuning.

For our setup, we used the following payload values:

* X-Axis Payload: 1.52 kg

* Y-Axis Payload: 0.87 kg

* Z-Axis Payload: 0.02 kg

Warning: The system will vibrate once plugged in. Remove any fragile components from the stages if necessary.

Step 4: Device Homing

Open the Basic Movement App in Zaber Launcher and press the Home button to home all three stages. This process clears the WR (Warning) flag and sets the device’s reference position.

Alignment Algorithm

Background

The alignment routine is divided into two stages: a first-light search and a fine optimization. The first-light search scans the fibers until sufficient optical power is detected, indicating that the fiber cores are close enough for coupling. Once first light is found, an optimization algorithm iteratively adjusts the stage positions to maximize the measured optical power and achieve peak coupling efficiency.

For the first-light search, we developed a spiral scan routine that leverages coordinated multi-axis motion and our controller's synchronized analog inputs. Rather than stopping at each measurement point, the stages move continuously while optical power is monitored in real time. This streamed approach eliminates move-and-settle delays, enabling first light to be detected much faster than with a point-by-point spiral or raster scan. Click here for more information on our streamed spiral scan algorithm and other first-light search routines.

Figure 3: Visualization of the alignment algorithm. The initial search locates first light by scanning the search area, after which a hill-climb optimization routine iteratively moves the fiber toward the peak coupling efficiency.

Running the Alignment Routine

The complete source code along with instructions, alternative search routines, and additional configuration options, is available in the Fiber Alignment Example GitHub repository. If alignment is lost due to vibration, temperature fluctuations, or other disturbances, simply rerun the alignment script to automatically relocate and optimize the fiber position.

Conclusion

By combining modular motorized motion hardware with an open-source Python alignment routine, researchers can build an automated fiber alignment system that delivers fast, repeatable, and high-precision alignment for approximately $13k USD. Because the system is modular and fully programmable, it can be readily adapted to a wide range of fiber alignment applications and experimental setups.

If you're interested in automating fiber alignment in your lab, contact our Applications Engineering team. We'd be happy to help.