In recent decades, and continuing into the foreseeable future, the level of connectivity between devices has been, and will remain, on an exponential growth trajectory. Within this increasingly interconnected landscape, the Internet of Things (IoT) has emerged as a transformative paradigm, enabling the seamless integration of billions of devices across diverse application domains—ranging from industrial automation and smart infrastructure to healthcare and consumer electronics. By leveraging cloud computing and pervasive wireless communication, the IoT is not only redefining how data is acquired and processed, but also unlocking unprecedented opportunities for the deployment of intelligent, contextaware systems. This technological shift, however, imposes stringent requirements on the design of wireless interfaces, calling for high-performance, area- and energy-efficient, and cost-effective transmitter architectures capable of sustaining long battery life while maintaining reliable connectivity. In this context, digital polar transmitter architectures have attracted growing interest due to their inherently higher power efficiency compared to conventional Cartesian implementations. By decomposing the signal into amplitude and phase components, digital polar transmitters enable the use of highly efficient, digitally controlled, nonlinear power amplifiers—thereby improving overall system efficiency and facilitating agile reconfiguration across different modulation schemes. The thesis presents the design of an inverse Class-D power amplifier (PA) tailored for Bluetooth Enhanced Data-Rate (EDR) digital polar transmitters. The targeted digital phase-locked loop (DPLL) based transmitter architecture is thoroughly analyzed using behavioral modeling. A novel signal-processing method is introduced to relax the tuning range requirements of the DPLL, thereby enhancing system robustness and performance. Key system parameters necessary to meet Bluetooth specifications are extracted from the behavioral model, providing design insights and validation. The power amplifier is implemented in a 28 nm CMOS process and operates with a supply voltage of only 0.6V. The design includes the driver circuitry and an on-chip matching network. An integrated inside-transformer notch enables selective third-harmonic filtering, achieving harmonic suppression of –40 dBc. Post-layout simulations demonstrate a peak output power of 9.9dBm with a corresponding drain efficiency of 30%. The PA delivers an average output power and efficiency of 4.9 dBm and 22.1%, respectively. Overall, the stage performance is comparable to state-of-the-art designs. The AM-PM and AM-AM nonlinearities of the PA are evaluated to assess their impact on modulation performance metrics such as error vector magnitude (EVM), in-band, and out-of-band emissions. While the AM-PM distortion is negligible, the AM-AM nonlinearity requires dedicated digital correction techniques. Accordingly, digital predistortion is investigated to compensate for this effect. Results show that after correcting the AM-AM nonlinearity, the system meets the spectral and accuracy requirements imposed by the standard, achieving an RMS EVM of approximately 2%.
Negli ultimi anni, e nel futuro prossimo, il livello di connettività tra dispositivi ha osservato, e continuerà ad osservare, una crescita esponenziale. In questo scenario sempre più interconnesso, l’Internet of Things (IoT) si è affermato come un paradigma rivoluzionario, consentendo l’integrazione di miliardi di dispositivi in una vasta gamma di applicazioni, che spaziano dall’automazione industriale e dalle infrastrutture intelligenti fino ai settori della sanità e dell’elettronica di consumo. Grazie all’utilizzo combinato del cloud computing e della comunicazione wireless, l’IoT non sta solo rivoluzionando il modo in cui i dati vengono acquisiti ed elaborati, ma sta anche aprendo le porte a opportunità senza precedenti per l’implementazione di sistemi intelligenti e adattivi. Questa rivoluzione tecnologica, tuttavia, impone requisiti stringenti nella progettazione delle interfacce wireless, che devono garantire alte prestazione, efficienza energetica e di area, e bassi costi, mantenendo allo stesso tempo connettività affidabile e lunga durata della batteria. In questo contesto, le architetture di trasmettitori digitali polari hanno suscitato crescente interesse grazie alla loro superiore efficienza energetica rispetto alle controparti cartesiane tradizionali. Scomponendo il segnale nelle rispettive componenti di ampiezza e fase, i trasmettitori digitali polari consentono l’utilizzo di amplificatori di potenza non lineari, altamente efficienti e digitalmente controllabili—migliorando in questo modo l’efficienza complessiva del sistema e facilitando la riconfigurazione flessibile per diversi schemi di modulazione. La tesi presenta la progettazione di un amplificatori di potenza (PA) a classe D inversa, ottimizzato per trasmettitori digitali polari Bluetooth Enhanced Data-Rate (EDR). L’architettura di trasmissione, basata su un digital phase-locked loop (DPLL) è stata analizzata in dettaglio a seguito di una modellizzazione di sistema. Viene introdotta una nuova tecnica di elaborazione del segnale volta a rilassare di requisiti di tuning range del DPLL, migliorando così la robustezza e le prestazioni del sistema. A seguito della modellizzazione, sono stati estratti i parametri chiave necessari a soddisfare le specifiche del Bluetooth, fornendo indicazioni progettuali utili alla validazione. L’amplificatore di potenza è stato implementato in tecnologia CMOS a 28 nm e opera con una tensione di alimentazione pari a soli 0.6V . Il progetto include i circuiti di pilotaggio e una reta di adattamento integrata on-chip. E’ inoltre presente un notch integrato all’interno del trasformatore, che consente il filtraggio selettivo della terza armonica, ottenendo una soppressione armonica di −40dBc. Le simulazioni post-layout evidenziano una potenza di uscita di picco pari a 9.9dBm, con un efficienza di drain del 30%. L’amplificatore fornisce una potenza media di 4.9dBm e un efficienza media del 22.1%. Complessivamente, le prestazione dello stadio risultano in linea con l’attuale stato dell’arte. Le non linearità AM-PM e AM-AM dell’amplificatore vengono valutate al fine di analizzare il loro impatto su metriche di performance come l’error vector magnitude (EVM), le emissioni in banda e fuori banda. Mentre la distorsione AM-PM risulta trascurabile, la non linearità AM-AM richiede tecniche di correzione digitale dedicate. A tal proposito, viene analizzata l’applicazione di una predistorsione digitale (DPD) del segnale di ampiezza per compensare questa non idealità. I risultati mostrano che, dopo la correzione della non linearità AM-AM, il sistema soddisfa pienamente i requisiti in termini di spettro e accuratezza, imposti dallo standard, con un EVM RMS pari a circa il 2%.
Design of a low-voltage inverse Class-D power amplifier for enhanced data-rate bluetooth digital polar transmitters in 28nm CMOS technology
Frisone, Sara
2024/2025
Abstract
In recent decades, and continuing into the foreseeable future, the level of connectivity between devices has been, and will remain, on an exponential growth trajectory. Within this increasingly interconnected landscape, the Internet of Things (IoT) has emerged as a transformative paradigm, enabling the seamless integration of billions of devices across diverse application domains—ranging from industrial automation and smart infrastructure to healthcare and consumer electronics. By leveraging cloud computing and pervasive wireless communication, the IoT is not only redefining how data is acquired and processed, but also unlocking unprecedented opportunities for the deployment of intelligent, contextaware systems. This technological shift, however, imposes stringent requirements on the design of wireless interfaces, calling for high-performance, area- and energy-efficient, and cost-effective transmitter architectures capable of sustaining long battery life while maintaining reliable connectivity. In this context, digital polar transmitter architectures have attracted growing interest due to their inherently higher power efficiency compared to conventional Cartesian implementations. By decomposing the signal into amplitude and phase components, digital polar transmitters enable the use of highly efficient, digitally controlled, nonlinear power amplifiers—thereby improving overall system efficiency and facilitating agile reconfiguration across different modulation schemes. The thesis presents the design of an inverse Class-D power amplifier (PA) tailored for Bluetooth Enhanced Data-Rate (EDR) digital polar transmitters. The targeted digital phase-locked loop (DPLL) based transmitter architecture is thoroughly analyzed using behavioral modeling. A novel signal-processing method is introduced to relax the tuning range requirements of the DPLL, thereby enhancing system robustness and performance. Key system parameters necessary to meet Bluetooth specifications are extracted from the behavioral model, providing design insights and validation. The power amplifier is implemented in a 28 nm CMOS process and operates with a supply voltage of only 0.6V. The design includes the driver circuitry and an on-chip matching network. An integrated inside-transformer notch enables selective third-harmonic filtering, achieving harmonic suppression of –40 dBc. Post-layout simulations demonstrate a peak output power of 9.9dBm with a corresponding drain efficiency of 30%. The PA delivers an average output power and efficiency of 4.9 dBm and 22.1%, respectively. Overall, the stage performance is comparable to state-of-the-art designs. The AM-PM and AM-AM nonlinearities of the PA are evaluated to assess their impact on modulation performance metrics such as error vector magnitude (EVM), in-band, and out-of-band emissions. While the AM-PM distortion is negligible, the AM-AM nonlinearity requires dedicated digital correction techniques. Accordingly, digital predistortion is investigated to compensate for this effect. Results show that after correcting the AM-AM nonlinearity, the system meets the spectral and accuracy requirements imposed by the standard, achieving an RMS EVM of approximately 2%.File | Dimensione | Formato | |
---|---|---|---|
2025_07_Frisone_Executive_Summary.pdf
non accessibile
Descrizione: Executive Summary
Dimensione
2.85 MB
Formato
Adobe PDF
|
2.85 MB | Adobe PDF | Visualizza/Apri |
2025_07_Frisone_Tesi.pdf
non accessibile
Descrizione: Testo della Tesi
Dimensione
56.92 MB
Formato
Adobe PDF
|
56.92 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/240550