Search

Applications such as plasma generation require the generation and delivery of radio-frequency (rf) power into widely-varying loads while simultaneously demanding high accuracy and speed in controlling the output power across a wide range of power levels. Attaining high efficiency and performance across all operating conditions while meeting these system requirements is challenging, especially at high frequencies (10s of MHz) and power levels (1000s of Watts and above). This paper evaluates different architectures that directly address these challenges and enable efficient high-frequency operation over a wide range of output power levels and load impedances with the capability of fast output power control (e.g., within a few microseconds). We review techniques for achieving fast output power control and evaluate their suitability in efficient rf systems demanding accurate and fast control of output power. Two dc-to-rf system architectures utilizing the discussed power control techniques are presented to illustrate both the achievable performance benefits as well as their robustness to load impedance variation, and are compared using time-domain simulations. The results indicate the ability of the proposed architectures to maintain high efficiency (> 90%) across a very wide range of output power levels (e.g., over a factor of up to 85x) while being robust to load impedance variations.

Industrial radio frequency (rf) power applications, such as plasma generation, require high-frequency rf power over a wide dynamic power range and across variable load impedances. It is desired in these applications to maintain high efficiency and fast dynamic response. This paper introduces a scalable power amplifier (PA) architecture and control approach suitable for such applications. This approach, which we refer to as Multi-Inverter Discrete Backoff (MIDB), losslessly combines the outputs of paralleled switched-mode PAs, and modulates the number of active PAs to provide discrete steps in rf output voltage. It further employs outphasing among sub-groups of PAs for rapid and continuous output power control over a wide range. In doing so, the architecture can maintain high efficiency and fast rf power control across a very wide backoff range. A device selection and loss optimization method for MIDB architectures is discussed for plasma generation applications. We address the use of GaN FET-based, ZVS class-D PA units, and consider dynamic R[subscript ds],on effects and C[subscript oss] losses typical of GaN FETs.

Plasma generation systems represent a particularly challenging load for radio-frequency power amplifiers owing to the combination of high operating frequency (e.g., 13.56 MHz) and highly variable load parameters. We introduce a dynamic matching system for inductively coupled plasma (ICP) generation that losslessly maintains near-constant driving point impedance (for low reflected power) across the entire plasma operating range. This new system utilizes a resistance compression network (RCN), an impedance transformation stage, and a specially configured set of plasma drive coils to achieve rapid adjustment to plasma load variations. As compared to conventional matching techniques for plasma systems, the proposed approach has the benefit of relatively low cost and fast response, and does not require any moving components. We describe suitable coil geometries for the proposed system, and treat the design of the RCN and matching stages, including design options and tradeoffs. A prototype system is implemented and its operation is demonstrated with low pressure ICP discharges with O $_{2}$ , C $_{4}$ F $_{8}$ , and SF $_{{6}}$ gases at 13.56 MHz and over the entire plasma operating range of up to 250 W.

Plasma generation systems represent a particularly challenging load for radio-frequency power amplifiers owing to the combination of high operating frequency (e.g., 13.56 MHz) and highly variable load parameters. We introduce a dynamic matching system for Inductively Coupled Plasma (ICP) generation that losslessly maintains near-constant driving point impedance (minimal reflected power) across the entire plasma operating range. This new system utilizes a Resistance Compression Network (RCN), an impedance transformation stage, and a specially-configured set of plasma drive coils to achieve rapid adjustment to plasma load variations. As compared to conventional matching techniques for plasma systems, the proposed approach has the benefit of relatively low cost and fast response, and does not require any moving components. We describe suitable coil geometries for the proposed system, and treat the design of the RCN and matching stages, including design options and tradeoffs. A prototype system is implemented and its operation is demonstrated with low pressure ICP discharges with O 2 , C 4 F 8 , and SF 6 gases at 13.56 MHz and over the entire plasma operating range of up to 250 W.