PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
June 5, 20260 citationsOpen Access

GaN and GaN/AlN-based monolithic microwave integrated circuits for 5G applications

View Full Paper
SHSanaul Haque

Key Points

  • This research aims to enhance linearity and develop compact RF front-end modules for 5G applications using GaN and GaN/AlN technologies.
  • Designed an LNA with improved linearity for 5G FR1 bands and a half-duplex RF front end for FR2 bands.
  • Compared performance of RF components fabricated in GaN and GaN/AlN processes at FBH.
  • Integrated RF SPST switch for compactness while maintaining RF performance in half-duplex systems.
  • Achieved a 6 dB improvement of OIP3 at 5 GHz without compromising gain and power consumption.
  • Developed a module with a 24% smaller chip area due to RF SPST switch integration.
  • GaN/AlN LNA showed competitive performance but with a slightly higher noise figure than GaN LNA.

Abstract

The scope of this thesis is two-fold. One is to design an LNA with improved linearity in 5G FR1 bands, and a robust, compact, and multifunctional half-duplex RF front end in 5G FR2 bands. Another is to compare the performance of RF components such as LNA and PA fabricated in the newly developed GaN/AlN HEMT process and the established GaN HEMT process at FBH. Linearity enhancement using derivative superposition (DS) technique is demonstrated for the first time in GaN-based LNA MMIC. In the design process, the limitation of DC-based bias selection at high frequency is highlighted and a novel RF-based approach to implement the DS technique is presented. An excellent 6 dB improvement of OIP3 is achieved in the two-tone measurements around 5 GHz, without compromising gain and power consumption, with a slightly degraded noise figure and reasonable form factor. In this thesis, two Ka-band half-duplex RF front-end modules are designed and characterized. One module is designed with an LNA, a PA and an RF SPDT switch. To realize a more compact front-end module, an RF SPST switch is integrated with both LNA and PA. An innovative approach to integrate the RF SPST switch in LNA is presented, while the RF switch integration in PA is relatively straightforward but very effective in realizing a compact front-end module. The two front-end modules exhibit identical RF performance. However, the latter module occupies a significant 24% less chip area due to the integrated RF SPST switch in both LNA and PA. In half-duplex systems, the transmit and receive path are not in operation simultaneously. The power supply to the idle component can, therefore, be turned off to reduce power consumption. This is addressed by integrating a DC supply switch with all Ka-band LNAs and PAs. An innovative idea of RF switch-less receiver is presented using an LNA with only DC supply switch. The LNA provides a high input reflection coefficient (-3 dB at 30 GHz) and a high port-to-port isolation (20 dB at 30 GHz) in transmit mode without degrading RF performance in receive mode. The RF switch-less receiver, upon realization, will significantly improve receiver gain, noise figure, sensitivity, and module form factor. Ka-band LNAs are also characterized for input stress. LNAs either as stand-alone or in front-end modules survive more than 34 dBm of input stress at 30 GHz, demonstrating highly rugged LNA or receiver in GaN and GaN/AlN HEMT process. GaN/AlN-based LNA exhibits very competitive small-signal performance, but slightly higher noise figure compared to that of GaN-based LNA. The noise figure is expected to improve with the maturity of the technology. On the other hand, GaN/AlN-based PA cannot outperform the GaN-based PA in terms of large-signal performance, indicating the presence of trap-related dispersion in the GaN/AlN HEMT process as well. Improved and optimized epitaxial layers in the GaN/AlN HEMT process are required to fully benefit from this exciting technology.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Sanaul Haque (2026) studied this question.

synapsesocial.com/papers/6a2268f9763171746d5478eehttps://doi.org/10.26127/btuopen-7363
Ask AI
Helpful
Bookmark
Share
View Full Paper