Qorvo In-depth: GaN targets 3.5GHz and moves towards 5G
The demand for wireless network infrastructure is expected to continue to grow rapidly. In order to adapt and meet these demands, wireless communication infrastructure continues to develop. Network operators are committed to continuous improvement in many aspects such as cost, capacity, coverage, user experience quality, reliability, functionality, interoperability, spectrum efficiency, mobility, etc.
That doesn’t mean the demand is negligible. Cisco reports that mobile traffic grew 69% in 2014, reaching 2.5 exabytes per month. Another way to put that level of data transfer rates is that in 2014, mobile traffic was nearly 30 times the global Internet traffic in 2000. It is predicted that by 2019, average monthly data rates will increase 10 times today to 24 exabytes per month. The growth in data traffic is driven by both traditional mobile users and the expected growth in all types of data connections, the so-called Internet of Things (IoT). The need for additional network capacity seems insatiable, and as LTE continues to gain popularity around the world, network operators are already planning for the next major evolution in mobile networks: 5G.
5G is the fifth generation of mobile infrastructure networks, incorporating many network technology advances and expectations. Like previous generations, 5G is expected to further improve spectrum efficiency, support more users, provide higher data rates, and improve the user experience. Exactly how network operators will create a 5G experience is still unknown, but it is clear that all network operators have a common need for more bandwidth to meet the growing network demand. Utilizing additional spectrum is a major requirement and focus of next-generation network systems. Numerous R&D projects cover all frequency bands: low-frequency white space TV signal bands; unlicensed bands at 2.4 and 5 GHz; existing point-to-point and point-to-multipoint bands at 15, 28-30, 50, 60, and 71-86 GHz (E-band); and 3.5 GHz. Each band has its own advantages and disadvantages, and as heterogeneous networks continue to extend, next-generation network solutions will likely cover all of them.
The 3.5 GHz band provides a realistic solution to meet the growing spectrum demand without network operators having to wait for 5G solutions. For hardware manufacturers, the design platform required for 3.5 GHz solutions is very similar to existing traditional cellular bands compared to higher frequency alternatives.
3.5 GHz offers a total of 200 MHz of spectrum in the 3.4-3.6 GHz range; it is available in most parts of the world and is recognized as a potential globally harmonized band for TDD. Japan is at the forefront of the use of 3.5 GHz technology for mobile infrastructure, and recent reports indicate that field trials in China are increasing. Europe has long allocated fixed broadband bands; mobile infrastructure is expected to dominate future use. The United States faces more difficulties in harmonizing bands with the rest of the world because some of the bands are used for radar. However, the FCC recently opened up 100 MHz of spectrum for commercial use, the latest popular "innovation band." Given these spectrum allocations, 3.5 GHz will play a key role in future network expansion, both for carrier aggregation and on a standalone basis.
These new trends will continue with a common theme of increasing network density. Increasing network density uses a layered approach, where multiple access layers are installed to increase capacity in high traffic areas. Existing base stations are available with a variety of power levels, often varying by manufacturer, depending on the indoor or outdoor nature of the coverage area:
• Femtocell, less than 0.25 W;
• Picocell, 0.25-0.5 W;
• Microcell, 1-5 W;
• Metro cellular, 5-10 W; and
• Traditional macro cells, more than 10 W.
With a variety of power levels, operators have great flexibility to build smaller, denser, and higher-capacity coverage areas in their networks. Initially, 3.5 GHz was considered ideal only for small cell base stations, but now it is expected to be deployed at all power levels, providing network operators with a full-tier solution.
In response to the growing demand for outdoor 3.5GHz solutions, Qorvo has developed new GaN amplifier products and Doherty power amplifier reference designs with an eye on 1, 2 and 20W average output power at the antenna reference plane. In the future, 10W metro cellular and 40W macro cellular solutions will be developed. GaN is an ideal choice for this frequency band because it has high gain, high power density and high efficiency, which far outperforms competing technologies in performance. Such designs use Qorvo's 100mm (4-inch) wafer silicon carbide-based gallium nitride process with a gate of 0.25µm. Qorvo's rich process options provide 28-32V operating voltage solutions for small cell applications and 48V operating voltage solutions for macro cell applications. The cost of 100mm GaN has dropped significantly in the past few years, and the planned 150mm (6-inch) wafer transition solution will further reduce costs.
0.25 µm GaN has higher gain and operating frequency than GaN processes with wider gate widths such as 0.5 µm. To maintain high efficiency in the power amplifier product (including the driver and pre-driver stages), the gain of the final Doherty amplifier (an excellent choice for the 3.5 GHz band) needs to be as high as possible. The high power density of GaN compared to GaAs or silicon allows for lower Cds, thereby increasing bandwidth. The low Cds and higher intrinsic impedance of the device allow for the creation of internal package matching networks suitable for high video bandwidth applications. The video bandwidth in the 3.4-3.6 GHz band must be high as 100 MHz signal bandwidth is planned and 200 MHz is under discussion.
At 3.5GHz, the insertion loss between the power amplifier (PA) and the antenna (including circulator, board losses, and filtering) is estimated to be 2dB. Therefore, if the average radiated power at the antenna is 20W, the Doherty power amplifier reference plane will require 32W of power. The required peak power of the Doherty power amplifier is a function of the modulated carrier peak-to-average ratio. For the downlink LTE signal of a macrocell base station, the peak-to-average ratio is generally around 7dB after peak clipping. An additional 1dB of digital pre-distortion (DPD) margin is added to the power amplifier specification to compensate for temperature performance and device-to-device variations. Therefore, when the average power at the antenna is 20W, the peak power must be 200W, which is 8dB higher than 32W.
Figure 1. A symmetrical Doherty power amplifier achieves 2 W average output power in Band 42. The design uses two TQP0103 GaN transistors, both in a 3 × 4 mm plastic QFN package.
To demonstrate the performance of GaN, we developed a symmetrical Doherty power amplifier reference design for Band 42 (3.4 - 3.6 GHz). The output power at the antenna is 2W, and both the carrier and peaking amplifiers use TQP0103 GaN transistors (see Figure 1). At 8 dB back-off, the power amplifier can deliver more than 20W of peak power with an efficiency of more than 44%. Gain and efficiency as a function of output power are shown in Figure 2.
Figure 2. Gain and efficiency of a symmetrical Doherty power amplifier as a function of output power using a single-carrier, 64 DPCH, 10 dB PAR WCDMA test signal at 3.5 GHz.
The solution for 1 W average output power at the antenna end uses an asymmetric Doherty power amplifier, whose carier amplifier uses TQP0102 GaN transistors and peaking amplifier uses TQP0103. The power ratio of the peaking amplifier to the carrier amplifier is 2:1. Under 8 dB back-off conditions, the efficiency of the asymmetric Doherty power amplifier is higher than that of the symmetric Doherty power amplifier. At the amplifier reference plane, this reference amplifier design will achieve an efficiency of more than 50%. At the same back-off power, this will increase the efficiency by more than 6 points compared to the symmetric Doherty power amplifier.
When designing an asymmetric Doherty PA, special care must be taken to ensure that the AM-AM and AM-PM responses of the Doherty PA are smooth and monotonic, which is a requirement when used with DPD. The gain and phase responses must be monotonic during the transition from the peaking amplifier switching from off to on and load modulating the carrier amplifier to peak power. Achieving the proper gain and phase response is not easy because two different devices are used, each with different gain and phase responses. Also, the operating modes are different, with the carrier amplifier in class AB and the peaking amplifier biased in class C. Experiments have shown that an asymmetric GaN Doherty design can achieve better than -60 dBc ACPR performance with a 10 MHz signal bandwidth using a third-party DPD system (see Figure 3a). With a 20 MHz LTE signal, DPD can achieve better than -57 dBc ACPR performance over a wide range of back-off power levels (see Figure 3b). The range extends from deep back-off, where only the carrier amplifier is active, to the transition when the peaking amplifier is on and load modulating the carrier amplifier. DPD will improve linearity as long as the sum of the average power plus the peak-to-average ratio of the modulated carrier is less than the saturation power of the Doherty amplifier. DPD systems cannot compensate for nonlinearity at powers exceeding the saturation power capability of the Doherty amplifier. When attempting to compensate, the ACPR performance degrades rapidly, with the slope becoming almost vertical. Figure 3 shows a 6.5dB PAR when backed off to less than 6.5dB.
Figure 3 ACPR vs. output power (P3dB or less) of an asymmetric Doherty power amplifier, where (a) is a 10 MHz, 6.5 dB PAR WCDMA signal and (b) is a 20 MHz, 6.5 dB PAR LTE signal.
The asymmetric Doherty architecture is also used in the 20W average power (antenna end) reference design. The amplifier uses the QDP3600 as the carrier amplifier and the T1G4012036-FS as the peak amplifier, both of which operate at 48V. The load is designed to achieve 70W and 140W of carrier and peak amplifier power, minus the combined P3dB value of 200W at the Doherty power amplifier output power. The design is based on the load-pull measurements of the transistor. The load-pull contours of the carrier amplifier at 3.5GHz are shown in Figure 4. The red contour shows the peak power (P3dB) capability of the device. The black circle is the maximum power contour under 3:1 VSWR conditions, showing the carrier amplifier efficiency matching conditions that can be achieved by the device in the Doherty configuration. The green and blue contours show the gain and drain efficiency of the Doherty amplifier at the target average power of 45dBm, respectively. On the left side of the 3:1 VSWR circle, it can be seen that 60% drain efficiency and 19.6dB gain are achievable when the carrier amplifier efficiency is matched. When the peaking amplifier is turned on, it modulates the carrier amplifier load to the center of the 3:1 VSWR circle, with a load condition of 48.9dBm P3dB. When the peaking amplifier is matched to a power ratio of 2:1, the expected design performance is 14dB gain and 55% efficiency at 45dBm average output power.
Figure 4. Load-pull output power, efficiency, and gain contours for the QDP3600 carrier amplifier transistor used in an asymmetric 20W Doherty amplifier.
This efficiency level is significantly higher than silicon LDMOS, the current power amplifier technology for the cellular infrastructure market. The difference between GaN and LDMOS increases with frequency. The latest 3.5 GHz devices currently under development by LDMOS are expected to have an efficiency of 37.5% when using Doherty power amplifiers, based on data sheet specifications.
There is much talk and excitement about the future of 5G networks. While the 3.5 GHz band has not yet reached its full potential as a globally harmonized band, there are many opportunities in 3.5 GHz to meet the near-term needs for mobile infrastructure bandwidth while 5G deployments take time. New spectrum development plans are growing rapidly for both small cell and macro cell deployments, where GaN is well positioned to meet the needs for high power, high efficiency, and wide video bandwidth.
Long press to identify the QR code in the picture to follow~
Featured Posts
京公网安备 11010802033920号