Building a Nanomeasuring Machine Using LabVIEW and NI PXI
"The new system architecture has greatly increased raw processing system performance in the new devices. The modular hardware platform helps us easily and quickly adapt the signal- and data-processing unit to meet new challenges."
- Felix Balzer,
Ilmenau University of Technology
Developing a nanomeasuring machine (NPMM-200) with a measuring volume of 200 x 200 x 25 mm³ that must perform signal acquisition and preprocessing at a high frequency with extremely low latency for subnanometer resolution, high-precision moving stage control, and nanometer uncertainty.
Using three modular, scalable NI PXI systems and 11 field-programmable gate array (FPGA) modules to complete the signal- and data-processing unit and easily integrate innovative metrological and control concepts.
Felix Balzer - Ilmenau University of Technology
Johannes Klöckner - Ilmenau University of Technology
Stephan Zschäck - Ilmenau University of Technology
Brandon Percle - Ilmenau University of Technology
The ongoing trend of component miniaturization in several high-tech industries and laboratories worldwide has significantly increased the need for dimensional micro- and nanometrology. There is a strong demand for multiple-scale measurements, which occur when objects have outer dimensions in the millimeter range, but contain features reaching atomic dimensions and associated tolerances in the nanometer range. The Collaborative Research Centre 622 (SFB 622), funded by the German Research Foundation, has been working to develop scientific and technological fundamentals to design and construct nanomeasuring machines since 2002.
We hope to achieve traceable 3D metrology at subnanometer accuracy over a several-hundred-millimeter range by 2020. To reach this goal, the SFB 622 is developing a new nanopositioning and measuring machine called the NPMM-200. Due to groundbreaking nanopositioning technology research results over the past decade, we significantly increased the measuring volume up to 200 x 200 x 25 mm³. We used seven interferometers with a 0.08 nm resolution for length measurement and six degree-of-freedom (DOF) stage control.
Additionally, we used two autocollimators to initialize the interferometric angular measurement. We applied novel fiber-coupled confocal sensors to nullify the interferometric length measurement. Several analog and digital inputs are available to use various tactile and nontactile probe systems to make high-precision microstructure measurements.
Why We Chose NI Technology
We needed a new signal- and data-processing unit system architecture to meet high-performance requirements, especially for system frequency, closed-loop control, and the management of large amounts of data. To achieve our ambitious goal of realizing a measurement uncertainty below 30 nm for a specific measurement task, we needed the following:
- Parallel synchronized data acquisition with 666 2/3 kHz
- In-cycle measurement value pre- and post-processing at 83 1/3 kHz
- Closed-loop control at 8 1/3 kHz
- A low-latency data transfer
The flexible, modular NI PXI hardware platform met all system requirements. It features several PXI modules with signal and triggering capabilities, such as FPGA modules with analog and digital interfaces. We can also use these modules for processing. In addition, we appreciated the hardware-independent graphical dataflow-oriented programming and the seamless LabVIEW and NI hardware integration.
The NPMM-200 Signal- and Data-Processing Unit
The signal- and data-processing unit (SDPU) uses three NI PXI chassis with a real-time controller and 11 FPGA modules. The system consists of three subsystems: the data acquisition system (DAS), the sequence control system (SCS), and the controller system (CS). The DAS conducts signal acquisition and preprocessing at a frequency of 666 2/3 kHz for 39 analog inputs simultaneously. Measurement postprocessing, such as complex computations and data correction, occurs at a frequency of 83 1/3 kHz. The SCS and the CS perform the control and system administration at a frequency of 8 1/3 kHz.
The measurement data from the DAS is transmitted to the SCS, and setpoint and other control inputs are transmitted to the CS. The CS then outputs the actuator signals to the mechanical systems. The DAS has the most demanding real-time requirements, due to a data acquisition rate of 666 2/3 kHz and the 12 µs measurement cycle time, including the in-cycle data pre- and post-processing. The pre- and post-processing include complex algorithms using floating-point arithmetic with single and double precision.
According to the low latency, in-cycle computation, and cycle time requirements, the computation can only be realized within the FPGA modules. Therefore, we developed a high-performance ﬂoating-point processor (LISARD) for complex algorithm computation within an FPGA module.
We could not use the available standard PXI communication methods to exchange data between the FPGA modules. Instead, communication occurs over a specially designed communications interface that uses digital inputs and outputs. One digital line is used for the clock signal, and up to 32 digital lines are used for data. This communication interface is used with different data widths (4- to 32- bit). The achievable maximum data rate approaches 1.28 Gbit/s.
The measurement processing is divided into three phases:
- Acquiring eight A/D converter channels per FPGA module directly from the process
- Implementing measurement preprocessing, raw value correction, and data fusion
- Correcting the combined measurement values, postprocessing the measurement, and transferring it to the sequence and control systems
By using a cascaded structure with a pipeline mechanism, we can meet the overall timing requirements. Each single phase has a computation time lower than 12 µs. The DAS processes 39 analogous signals with a 16-bit resolution and a frequency of 666 2/3 kHz in parallel. The star trigger bus triggers data acquisition by using the NI PXI-6653 timing and synchronization module. According to the available analogous inputs, we use five NI PXI-7853R FPGA modules for data acquisition. These modules also provide signal preprocessing and measurement value generation.
The DAS can further be divided into two subsystems: the positioning and probe systems. A total of 23 analogous signals, mainly position and angle information, are processed within the positioning subsystem. After some minor algorithmic corrections, raw measurement value generation occurs at a frequency of 83 1/3 kHz. The positioning system uses three data acquisition modules and one NI PXI-7813R module to merge the raw measurement values. Additionally, we use one NI PXI-7853R module and one PXI-7854R module for memory-intensive postprocessing.
DAS output is a structure of measurement values containing three double-precision values and 20 single-precision values, with a minimum data rate of 8 1/3 kHz for further processing in the downstream subsystems, and a maximum data rate of 83 1/3 kHz for storage.
The SCS performs overall system sequence control, which also includes a trajectory generator. The CS includes a powerful closed-loop controller consisting of three Kalman filter and six proportional integral derivative controllers. The complete-control algorithms use floating-point arithmetic. We use two NI PXI-7853R FPGA modules to compute the algorithms and interface the process. The maximum control cycle rate is 115 kHz.
NI Technology Overview
Our application uses an NI PXI-8108 embedded controller for administration tasks. We control timing and synchronization using NI PXI-6653 and NI PXI-6651 modules. For interfacing, in-cycle computation, and communication, we use an NI PXI-7813R module for data fusion, an NI PXI-7853R module for process interface and signal processing, and an NI PXI-7854R module for additional signal postprocessing.
To develop the current stable version, we used NI LabVIEW 2010 software, the LabVIEW FPGA Module, and the LabVIEW Real-Time Module. Development greatly depended on LabVIEW training and our NI Volume License Agreement.
We can meet next-generation NPMM device computational challenges using NI technology and self-developed communication methods with computational concepts. The modular and scalable system architecture provides a high clock rate for real-time, low-latency signal and data processing for generating position, angle, and probe sensor data sets with very high accuracy and precision. The NPMM-200 is one of the most accurate nanomeasuring machines in the world, with 80 pm resolution in a measuring volume of 200 x 200 x 25 mm³.
The new system architecture has greatly increased raw processing system performance in the new devices. The modular hardware platform helps us easily and quickly adapt the signal- and data-processing unit to meet new challenges.
Bookmark and Share
Explore the NI Developer Community
Discover and collaborate on the latest example code and tutorials with a worldwide community of engineers and scientists.
Who is National Instruments?
National Instruments provides a graphical system design platform for test, control, and embedded design applications that is transforming the way engineers and scientists design, prototype, and deploy systems.