Using Multi-Configuration Operands to control parameters in a single configuration system

Also available in 日本語

This article illustrates how to use the multi-configuration editor to optimize, tolerance, and pickup values that cannot be accessed in any other editor.

Authored By Akash Arora

download example (.zip)

Introduction

OpticStudio supports a multiple configuration (MC) capability, allowing a system to be modelled in several different states. The Multi-Configuration Editor (MCE) defines the system parameters that differ from one configuration to the next. This allows users to model systems such as zoom lenses, scanning mirrors, and thermal (multiple temperature) systems where components move or change.

The MCE does have uses outside of MC systems. It can be used to greatly expand the range of parameters that can be allocated as variables for optimization, defined as tolerances for perturbation analysis, and linked to other values via solves. This article will explore examples of each of these three uses.

Multi-Configuration operands as variables

Optimization is the process whereby a set of variable parameters is systematically changed in order to achieve some target performance. Any value defined in the Lens Data Editor (LDE) or Non-Sequential Component Editor (NSCE) can be a variable. There are many system and surface parameters that aren’t defined directly in these editors, but rather in dialog boxes, leading one to conclude that they cannot be allocated as variables. Some examples are system apodization factor, system wavelengths, and surface aperture sizes.

However, in addition to the previously mentioned editors, any value that can be defined in the MCE can also be allocated as a variable. Thus, even if we aren’t modeling a MC system, we can still utilize the MCE to define variables that wouldn’t otherwise be possible. There is a complete list of MC Operands in the Multi-Configuration chapter of the OpticStudio Help Files. Some MC Operands control values that are already listed in one of the editors (i.e. NPOS – NSC object position) and some control values that wouldn’t be reasonable as variables (i.e. AFOC – afocal image space), but many control parameters that one may want to optimize and couldn’t any other way. Let’s look at an example.

Open the “Conic Interconnect” sample file, downloadable from this article. This system has been optimized to achieve maximum coupling efficiency for a 0.15 NA Gaussian input beam. Currently the apodization factor is 1, meaning the 1/e² intensity point lies at the edge of the entrance pupil.

Let’s say we wanted to determine the ideal Gaussian source fiber mode to maximize coupling efficiency. This could be done by optimizing the apodization factor for maximum coupling efficiency. The APDF MC Operand allows us parametric control of this value. Open the MCE (Setup Ribbon...MC Editor), double click on the Type cell, and select APDF from the drop-down list (or enter “APDF” into the Type cell). Allocate the parameter as a variable just as you would for a value in any other editor.

With this variable added, as is often the case, we will want to add some boundary constraints to the Merit Function to ensure that the optimizer doesn’t bring us to an unrealistic solution space. Add an MCOG (Multi-Config Operand Greater) constraint of 0, and MCOL (Multi-Config Operand Less) constraint of +5, making sure to set their weights to a non-zero value:

From the Optimize Ribbon...Optimize! tool, and run the local optimizer (DLS) with an automatic number of cycles. OpticStudio should give a value of ~1.75 for the apodization factor. Again, this variable couldn’t be optimized any other way. Note that the fiber coupling calculation only considers the apodization factor when the ignore source switch is enabled.

This same method can be utilized to have otherwise unobtainable values defined as variables in the universal plot. In this example, let’s plot the value of the FICL operand versus the apodization factor. Open a 1D universal plot (Analyze Ribbon...Universal Plot...Universal Plot 1D...New) and define the settings as follows.

We can clearly see a peak in coupling efficiency around 1.75, as expected, and the efficiency drops off to either side. Note that an apodization factor equal to zero is uniform. It is apparent that a uniform source fiber mode gives much worse coupling efficiency because the fiber coupling calculation assumes a Gaussian receiving fiber mode.

Multi-Configuration operands as tolerances

MC Operands can also be used to tolerance values for which there aren’t any supporting Tolerance Operands. The TMCO Tolerance Operand allows users to tolerance any value that can be defined in the MCE. Let’s look at an example. For information on tolerancing, please refer to the Knowledgebase Article "How to perform a sequential tolerance analysis".

Open the “Beam expander” sample file linked in the downloads section of this article. This system is a 3x laser beam expander that has been optimized for collimated output at 632.8 nm. Imagine the laser cavity has some instability that can alter output wavelength. What sensitivity does the beam expander have to input wavelength?

Without the MCE, there would be no way to tolerance the system wavelength. Note that there is a TWAV Tolerance Operand, however, this is only used for converting fringe tolerances to physical dimensions. We can use the WAVE MC Operand to determine how performance changes with wavelength. Define a single WAVE operand in the MCE.

Now open the Tolerance Data Editor (TDE) via the Tolerance Ribbon...Tolerance Data Editor, and define a single TMCO operand with the settings shown below. For this example, we will only look at the sensitivity of the design to wavelength. We will assume a ±50 nm wavelength range.

Run a Sensitivity Analysis with RMS wavefront as the criterion, sampling 5, and no compensators. As expected, the performance changes significantly. The nominal design had essentially zero wavefront error at 632.8 nm. The wavefront error is 0.39 waves at 582.8 nm and 0.26 waves at 682.8 nm.

This method can also be used with a Monte Carlo or inverse sensitivity analysis.

Multi-Configuration operands and solves

Another important use for MC Operands is to link values via solves. The MCE expands the range of values that can be linked using solves. Once again, let’s see an example.

Open the sample file “Schmidt-Cassegrain Telescope” linked in the downloads section of this article. This file models a two-mirror telescope with a corrector lens in front. In Sequential mode, we simulate the obscuration of the secondary mirror with a separate surface after the corrector lens. Ideally the size of the obscuration should be the same as the secondary mirror. We can set this manually, but if we plan on optimizing or tolerancing, using a solve is a much better solution.

Define two APMX (maximum aperture) operands in the MCE to control the maximum aperture on surfaces 3 and 5. Surface 3 has a circular obscuration, so the maximum aperture defines the obscuration size. Surface 5 has a circular aperture, so the maximum aperture defines the aperture size.

Defining both in the MCE allows us to apply a pickup solve linking the two values. Set the solve on the second operand as shown below.

The two values are now linked and will remain so if the system is modified or optimized.

KA-01670