Figure 1

Figure 1 Phase and amplitude control module block diagram.

Figure 2

Figure 2 Block diagram of a 1x8-way phase and amplitude control matrix (a). The size of the unit is 220 mm x 290 mm x 150 mm (b).

Figure 3

Figure 3 Amplitude (a) and phase (b) accuracy of the 1x8 matrix at 3.5 GHz, showing a sequence of individual amplitude and phase settings between 40 dB and 360 degrees.

Figure 4

Figure 4 Amplitude (a) and phase (b) accuracy of the 1x16 matrix at 28 GHz, showing a sequence of individual amplitude and phase settings between 40 dB and 360 degrees.

Figure 5

Figure 5 Typical phase response from 2 to 6 GHz for the 1.7 to 6 GHz matrix with 15 dB attenuation and 50 degrees phase.

Massive MIMO (mMIMO) phase and amplitude controlled networks are the key RF technology of 5G. However, available phase shifters and attenuators that provide phase and amplitude control are narrowband, with typical bandwidths of only 200 MHz and have control resolution no more than 7 bits for sub-6 GHz and 6 bits for the mmWave bands. So building test systems to cover the 5G bands requires a significant investment in narrowband equipment, with performance insufficient to meet high test accuracy requirements. This article describes wideband designs for controlling phase and amplitude, offering the widest bandwidth - 1.7 to 6 and 24 to 40 GHz - and highest resolution and accuracy commercially available. Two modules cover the 5G microwave and mmWave frequency bands with full 360 degree control and dynamic range to 50 dB and a resolution and accuracy of 0.1 dB and 1 degree sub 6 GHz and 0.2 dB and 2 degrees at mmWave. These phase and amplitude control modules serve as building blocks for beamforming, mMIMO, over-the-air (OTA) and multichannel signal simulator test systems. With their wideband coverage, resolution and accuracy, building test scenarios is simple and fast, reducing cost.

With 5G services being deployed commercially, the market’s demand for 5G test equipment is urgent. The most significant difference between 5G and 4G is the application of mMIMO and mmWave technology, using these capabilities to enable broadband, high capacity, high data rate transmission. 5G test applications include simulating the signal environment, assessing multichannel mMIMO, evaluating phased array OTA beamforming and automating production testing. As the test methods and systems for 4G are already refined, combining mMIMO and mmWave technology with existed test technologies is key to developing 5G test systems.

mMIMO requires precise and synchronous control of the phase and amplitude of multiple signals in the microwave or mmWave frequency bands. mMIMO channel emulators and OTA test products are available; however, they tend to have the following deficiencies:

Accuracy and coverage - Existing test systems use the same attenuators and phase shifters as used by 5G system suppliers, which do not meet the higher accuracy requirements for test equipment. Testing is only effective if the accuracy of the test equipment is higher than the accuracy of the system being tested. Otherwise, it is difficult to guarantee the validity of the test results. These digitally-controlled phase shifter and attenuator MMICs have smallest steps of 1.4 degrees and 0.25 dB to 6 GHz. While these step sizes provide theoretic accuracies of ±2.8 degrees and ± 0.5 dB attenuation, the products typically have poorer performance. Few digitally-controlled MMIC phase shifters and attenuators cover the 5G mmWave frequencies, i.e., from 24 to 40 GHz. Reported products have only 5.6 degrees and 0.5 dB step resolution, and their accuracies will be worse than the ± 11.2 degrees and ± 1 dB theoretical values - far from the accuracy required for 5G. Further, most of these are narrowband, typically with only 200 MHz bandwidth. Building test systems to cover the various 5G bands would require multiple narrowband systems, a significant investment.

Phase and amplitude not independent - When the attenuator changes attenuation, the phase will change. When the phase shifter adjusts phase, the attenuation will change. During operation, the temperature and signal amplitude will also affect the phase and attenuation, complicating system calibration and temperature compensation. High accuracy across wide bandwidths is difficult to achieve. Because of these challenges, most of the microwave MIMO channel simulators on the market only simulate fading and cannot simultaneously adjust phase or simulate signal routes. For mmWave, MIMO channel simulators have not yet been announced.

Difficulty scaling - The complexity of the test system makes it difficult to expand the number of channels and maintain the technical performance at a reasonable cost.

IDEAL PHASE-AMPLITUDE CONTROL

The ideal multichannel phase and amplitude control system should have the following characteristics:

High accuracy and stability - The system should have phase and amplitude accuracy sufficient to meet the requirements of future test systems, offering 5G equipment suppliers and operators a one-time investment with long-term benefit. The performance must be stable with time, temperature and input signal level, providing assurance of performance repeatability and reliability.

Broad bandwidth - A single unit should cover all the 5G microwave or mmWave bands. This enables a single system to test either the sub-6 GHz or mmWave bands, avoiding investment in additional equipment and reducing dedicated space for testing. Each unit should have wide instantaneous signal bandwidth with low distortion, to achieve high signal quality and data rates.

Simple to use for multiple applications - To maximize test efficiency, calibrating the equipment should be simple, and the system should be easy to operate and connect to other equipment for automating tests. It should be easy to synchronize across multiple channels and MIMO systems.

Each of these requirements is a difficult challenge, prompting a different approach to the design of the phase and amplitude controller. The approach pursued by Mitron was to first develop analog phase shifters and attenuators, then digitize the control to create a digitally-controlled analog phase shifter and attenuator. With an analog device as the foundation, the number of control bits can be chosen to achieve the desired control accuracy. For example, to get 5 degree phase accuracy, choose 8-bit phase control; for 60 dB dynamic range with 0.5 dB attenuation accuracy, choose 9-bit amplitude control.

To miniaturize the equipment, the analog phase shifter and attenuator were integrated into a single phase and amplitude control module, controlled with a USB or TTL interface (see Figure 1). The unit measures 203 mm x 88.9 mm x 21.59 mm. At any amplitude and phase combination, the phase control provides 1 degree minimum step and 2 degree absolute accuracy (typically 1 degree). The amplitude control provides 0.1 dB minimum step and 0.2 dB absolute accuracy. The overall dynamic range of the module is 360 degrees and 50 dB. To reduce the influence between the phase and amplitude adjustments, the design uses a “combination calibration method” to automatically test and record the various phase and amplitude combinations, creating a database accessed by the test software. For any phase and amplitude selected by the user, the corresponding control code is looked up by the software.

The individual phase and amplitude control modules are combined to create multichannel matrices for 5G test applications. Figure 2 shows the block diagram and photo of a 1x8 channel matrix covering 1.7 to 6 or 24 to 40 GHz. Both versions fit in the same housing, including the power supply and control circuitry, which measures 220 mm x 290 mm x 150 mm. The units are powered with 110 to 220 V AC and controlled via USB. A high speed synchronized trigger in every module enables the phase and amplitude of each channel to be set to its respective value at the same time. As each unit’s calibration is internal, no external equipment or vector network analyzer calibration is necessary. The units have very good repeatability and reliability, with tests showing the matrices maintain the same accuracy without recalibration after one year.

To demonstrate the capability of the phase and amplitude control matrices, Mitron built a 1x8 channel matrix covering 1.7 to 6 GHz and a 1x16 channel matrix covering 24 to 40 GHz. The two cover all current 5G sub-6 GHz and mmWave bands and can realize any arbitrary phase and amplitude combination with 2 degree phase and 0.2 dB amplitude accuracy. Since the designs use repeatable devices with fault-tolerant optimization and effective compensation and calibration, both units actually achieve 1 degree and 0.1 dB accuracy.

Figure 3a shows the amplitude accuracy of the 1x8 matrix at 3.5 GHz with various phase settings between 0 and 360 degrees and amplitude settings between 0 and 40 dB. Figure 3b shows the phase accuracy at 3.5 GHz with the same range of phase and amplitude settings. For the mmWave matrix, Figure 4a shows the amplitude accuracy at 28 GHz with various phase settings between 0 and 360 degrees and amplitude settings between 0 to 40 dB. Figure 4b shows the phase accuracy at 28 GHz with the same range of phase and amplitude settings. For the sub-6 GHz matrix, Figure 5 shows the phase flatness across a 4 GHz frequency span, with the phase set to 50 degrees and the amplitude at 15 dB. For the mmWave design, Figure 6 shows the phase flatness across a 4 GHz frequency span, again with the phase at 50 degrees and the amplitude 15 dB.

APPLICATIONS

Figure 6

Figure 6 Typical phase response from 26 to 30 GHz for the 24 to 40 GHz matrix with 0 dB attenuation and 50 degrees phase.

The utility of the phase and amplitude control matrices is illustrated by considering several applications.

Beamforming algorithms - To test beamforming, M signals with unique phase and amplitude values can be generated and combined without antennas or OTA transmission (see Figure 7). This setup is useful for testing and validating beamforming algorithms. Evaluation of the test system shows the matrix has a 1.6 percent error compared to the theoretical signal combination.

Phased array OTA - To test the antenna elements of a phased array, including the power amplifiers driving the elements, the phase and amplitude control matrix can be used to create a feed network, providing signals with precise phase and amplitude values to each feed (see Figure 8). The signals will have no more than 2 degrees and 0.2 dB error over the sub-6 GHz and mmWave bands. Using this system, the beam steering and frequency performance of the array can be evaluated. The system can also be used as a reference array for comparison with arrays with integrated front-end modules.

Base station links - Since the phase and amplitude control modules are reciprocal and can transfer signals in both directions, networks of matrices using phase and amplitude modules with multi-way power dividers and combiners can be configured to simulate various MIMO cases between base stations and users. Figure 9 illustrates two scenarios: four base stations and four users (4x4) and two base stations and eight users (2x8). Additional configurations, such as 4x16 and 16x64, are straightforward by combining modules like LEGO®blocks.

Other environments - This system can be used for any test application requiring multiple synchronized signals, each with unique phase and amplitude values. For example, it can emulate multichannel signal fading, routing or a Doppler shift to simulate a moving object, such as a vehicle or aircraft, generate the complicated electro-magnetic signal environments. Because of its wideband response, the matrix is also useful for supplying signals for testing carrier aggregation and adjacent channel interference.

Figure 7

Figure 7 Setup simulating an ideal beamformer, useful for evaluating beamforming algorithms.

 

Figure 8

Figure 8 Setup for antenna array and OTA testing.

Conclusion

Figure 9

Figure 9 The reciprocal matrices can simulate various base station and user scenarios, such as 4x4 and 2x8.

Wideband and accurate phase and amplitude control modules and integrated matrices simplify testing over the sub-6 GHz and mmWave 5G bands. Modules have been developed to cover 1.7 to 6 and 24 to 40 GHz, and the upper frequency can be extended to 43 GHz. With simple calibration, the modules are easily combined with power dividers and combiners to create custom test systems, well-suited for R&D labs and production lines. The inherent accuracy of the design approach ensures the matrices will support future generations of communications standards.