From a theoretical perspective, the more antennas a transmitter/receiver is equipped with, the better the performance will be in terms of data rate and/or link reliability. Massive multiple-input/multiple-output (MIMO) technology, with its many-antenna arrays, offers the potential for a 1000-fold increase in network capacity when used alongside dense small cell deployments

[1, 2]. Our recent five-part blog series discusses the new theoretical techniques and concepts that have been proposed in many-antenna base station designs.

In Part 5, the series looks at the need for the practical prototyping and implementation of these techniques and concepts. To benefit from the so-called “massive effect” however, the number of antennas must typically meet or exceed a certain quantity (approximately 40 elements). This is where Nutaq’s TitanMIMO-4S comes into the picture. The TitanMIMO Series provides a robust platform that addresses the need to rapidly and accurately model, simulate, test, and deploy advanced next-generation massive MIMO wireless protocols.

Based on industry standards (MicroTCA, AMC, and FMC), the TitanMIMO is a large-scale 100×100 MIMO test bed that combines a highly efficient backplane, bandwidth, low system latency, and extensive FPGA processing power. Complementing the impressive hardware is Nutaq’s software tools, which can reduce the overall development time by 30–60%. Nutaq supports both model-based design and GNU Radio, and includes a complete QAM64 OFDM reference design for the development of 4G and 5G applications.

TitanMIMO-4S Massive MIMO 100×100 Testbed

Key benefits

• Easily migrate designs from test vector validation to real-time implementation
• Reduce the overall development cycle by 30–60% using the included software tools
• Accelerate final product deployment by prototyping with industry standards

Key features

• 100×100 RF transceivers and beyond
• Data throughput allows scalability towards 5G (see TitanMIMO roadmap)
• MicroTCA.4 RTMs miniSAS high speed Aurora 4x interfaces allows high throughput data aggregation for the central processing FPGA unit
• Backplane supplies high-speed, low-latency PCIe interconnectivity for control
• Based on industry standards (MicroTCA.4, AMC, FMC)
• Supports full-bandwidth synchronous record/playback and real-time streaming interfaces
• Timestamp support for accurate time-based control
• GPS-disciplined (PPS) synchronization for frequency accuracy
• Simultaneous auto-calibration of all radios
• Xilinx Virtex-6 field-programmable gate arrays (FPGAs)
• Radios: 28 MHz bandwidth, tuning range from 300 MHz to 3.8 GHz
• Rapid prototyping using model-based design tools (GNU Radio, System Generator for DSP)
• QAM64 2×2 MIMO orthogonal frequency-division multiplexing (OFDM) reference design

Hardware overview

The TitanMIMO-4S supports 100×100 MIMO with three 12-slot MicroTCA.4 chassis, with each slot containing a 4×4 radio node (Perseus 611X with Radio420X FMC). An embedded CPU (Quad-core i7 CPU blade) enables each autonomous 4×4 MIMO subsystem to receive time-based commands via Nutaq’s application programming interfaces (APIs).

This functionality provides many features useful for massive MIMO development:

• 100x RF channel synchronous calibration and programming
• 100x RF channel synchronous record & playback
• 100x RF channel synchronous CPU streaming (half-duplex or full-duplex)

TitanMIMO-4S 100x100 block diagram

TitanMIMO-4S 100×100 block diagram

Radio overview

Stacked in a 2×2 MIMO configuration, and based on the agile Lime Microsystems LMS6002D RF-IC, Nutaq’s Radio420x FPGA mezzanine card (FMC) radio module targets a wide range of broadband and narrowband waveforms between 300 MHz and 3.8 GHz.

Each RF node offers the following features:

• Stackable to 2×2 to reduce deployment costs and footprint
• Highly shielded, field-ready design
• Tunable from 300 MHz to 3.8 GHz
• Dynamically selectable bandwidth between 1.5 MHz and 28 MHz
• QAM64-capable for 4G and 5G performance requirements
• Auto-calibration saves time and improves performance (IQ imbalance and DC-offset adjustment).
• Low-jitter phase lock loop (PLL) and 30.72 MHz reference clock
• Accepts external reference clocks from FPGA carrier card or front panel connector
• Internal or external GPS disciplined alignment

Radio420X block diagram

Accelerated waveform development

Nutaq’s model-based design approach accelerates massive MIMO waveform development and deployment by up to 60%. The TitanMIMO-4S supports mixed PC-FPGA designs and offers a complete integration of both GNU Radio and Xilinx System Generator for DSP.

These sophisticated model-based design and automatic code generation tools can eliminate the need for a traditional multi-disciplinary team of algorithm engineers, FPGA engineers, and PC engineers. Instead, the radio waveform can be deployed directly by the algorithm developer, avoiding errors that come from interface interaction and code-translation.

The TitanMIMO-4S includes a 2×2 QAM64 OFDM reference design, implemented in both Simulink (FPGA) and GNU Radio (host) to help you quickly get started with massive MIMO development.

GNU Radio host implementation

References

[1] M. Ahmed Ouameur, “Massive MIMO – Part 4: Massive MIMO and small cells, the next generation network,” http://www.nutaq.com/blog/massive-mimo-%E2%80%93-part-4-massive-mimo-and-small-cells-next-generation-network[2] Qualcomm, The 1000x Data Challenge, http://www.qualcomm.com/solutions/wireless-networks/technologies/1000x-data