Linear and Efficient RF Transmitters with Digital Pre-Distortion and Supply Modulation

Wireless communication systems require the translation of an information-bearing signal at higher frequencies (such as radio- or mm-wave frequencies) to allow propagation through the wireless medium (the channel). This translation is typically performed in transmitters and receivers that along with the channel form the communication system. In the transmitter, a power amplifier (PA) is used to boost the power of the signal to a level sufficient to overcome the channel's attenuation and arrive with sufficient signal strength at the receiver. Today's best PAs are capable of 60-70% efficiency when used at their maximum output power. This means that about 1/3 of the power is wasted into heat only for the purpose of amplifying at such higher frequencies. Efficiency decreases when reduced output powers are required.

Modern communications standards such as 5G generate signals which present a large power variation over time (this is also described by the peak-to-average power ratio or PAPR) and this causes the PA to operate even more inefficiently with values down to 10-20%, instead of the aforementioned 60-70%. Wasting almost 90% of the DC power into heat causes additional demands on the energy supply network which may lead to an increase in carbon emissions. Higher DC power dissipations result in reduced transmitter performance (e.g. less output power and so less coverage), reduced battery lifetime, in additional weight, cost, and size because of the heatsinks and necessary cooling hardware. Heat dissipation causes the electronics within the PA to operate at higher temperatures which are known to degrade the component's reliability (ageing) and change their electrical behaviour.

The goal of this project is to radically improve the RF PA efficiency by using a technique called supply modulation (SM). Unlike the 1952's envelope-tracking (ET) method, SM uses a very high-efficiency modulator to generate a number of voltage levels (Vmin, ..., Vmax) that are applied to the drain of the PA. When the RF output power in the PA is high, the PA is supplied with the maximum voltage level and so it operates at maximum efficiency. Vice-versa, when the PA output power is low, a lower voltage level is supplied to the PA drain. This change in the supply results in an efficiency improvement usually in the range of 20-30% (and so in a SM-PA efficiency of 30-50%), but most importantly, it typically reduces the DC power consumption by ~50% for the same output power.

Achieving wider and wider bandwidths for high link capacities requires this SM-PA to commutate very rapidly as a consequence of a wideband signal. The current state-of-art bandwidth is ~100MHz for the SM-PA. Achieving 1GHz bandwidth, as required in multi-band and mm-wave PAs, is thus the target of this project. To achieve this, new circuit topologies combined with high figure-of-merit semiconductor technologies will be explored, with the unavoidable hardware imperfections compensated through signal processing techniques such as digital pre-distortion (DPD).

The SMPA specifications and top-level design parameters will be agreed between the University of Bristol (UoB)'s team and the project's partners to ensure relevance for industrial applications. This SM-PA is firstly simulated in the SM part, then in the PA, and then co-simulated together as a complete sub-system. The fabricated prototype is then characterized in terms of linearity, efficiency, and power with the latest communication standards. The SM circuit can also be combined with existing PAs as an 'efficiency upgrade'. Results of this theoretical and experimental activity are presented at conferences and published in journals by the UoB team. Public engagement and industry impact is also ensured by the presence of an advisory board.

In summary, this project is an adventurous research programme that will re-define next-generation RF transmitters amplifiers and so contribute to UK's leadership in wireless technologies.