[Components] Cut Loss In Low-Voltage, Wideband PIN Attenuators Some simple techniques, such as the addition of a ferrite bead, can be used to reduce the minimum attenuation of a PIN diode attenuation for broadband, high-frequency applications. Chin-Leong Lim | ED Online ID #18648 | April 2008 Variable attenuators based on the pi (p) configuration can be designed with frequency coverage spanning multiple octaves, from the high-frequency (HF) range to the lower microwave region. As a result, they are suitable for numerous wideband applications, such as in satellite television and cable-television (CATV) systems. The use of PIN diodes in these attenuators enables them to be realized as low-cost components and compact alternatives to mechanical attenuators. Traditionally, designers of attenuator circuits have had the luxury of utilizing a wide control voltage range (e.g. 0 to 15 V) to achieve a wide span of attenuation values. The industry trend toward lower operating voltages and lower power dissipation imposes severe limitation on the range of voltages available to drive the PIN attenuator. Reducing the maximum voltage available for controlling the adjustable attenuator unfortunately degrades the minimum attenuation limit. This paper describes various circuit tricks that can be used to lower the attenuator’s minimum attenuation while maintaining a low maximum voltage requirement. The degree of improvement afforded by these techniques makes the PIN attenuator a serious contender to replacing the costly mechanical attenuator, which is prized for its low minimum attenuation. In CATV distribution, designers aim to reduce installation costs by minimizing the use of higher-power trunking/line amplifiers. Being able to reduce the minimum loss of the attenuator is highly desirable as it allows designers to meet system gain target while using a lower power and cheaper amplifier. Adjustable attenuators are useful in a variety of communication applications, such as power leveling and automatic gain control in receivers and transmitters. PIN diodes, which behave like currentcontrolled resistors at RF frequencies, are ideal for forming the resistive elements in an electronically controlled attenuator. Although transistor-based devices such as FETs and MESFETs can also fulfill the role of the variable resistor in attenuators, PIN diodes offer several advantages: better linearity[1], smaller size and no need for a negative-polarity control voltage. The PIN diode’s lower distortion performance makes it a better choice when the attenuator is expected to function in the presence of high power or multiple-tone signals or both. The most basic form of attenuator can be made by connecting a PIN diode either in series or in shunt to the transmission path. However, these simple configurations attenuate by reflecting the signal back to the source and, therefore, have poor return loss (RL) at high attenuation values. If the RL must remain low over the entire attenuation range, then an alternative attenuator family that reduces signal amplitude by absorption is required. Absorptive attenuators, which are a subset of constant impedance attenuators, commonly are implemented as p or bridged-T resistor networks[2]. At HF to the lower microwave region, the p attenuator is arguably the most popular configuration because of its attractive combination of compact size, multi-octave frequency range and large attenuation range. The PIN p attenuator’s 0.3 to 3000 MHz operating range can easily fulfill the bandwidth requirement of CATV/ SATV networks and set-top boxes. It is also useful in wireless and cellular communication infrastructure equipment in the very-high-frequency (VHF) and lower microwave regions, especially where long wavelengths preclude attenuators having distributed components such as the hybrid coupled and resistive line attenuators. In the past, high linearity PIN p attenuators used a wide control voltage range (e.g. 0 to 15 V) to achieve a good dynamic range (the difference between minimum and maximum attenuation[3]. However, the trend toward miniaturization and batterypowered devices requires attenuators that can be operated over a much lower control range (e.g. 0 to 5 V) while maintaining similar dynamic range. In a typical PIN attenuator, the upper end of the control voltage range corresponds to the minimum attenuation setting. As a result, the low-voltage PIN attenuator may have higher minimum insertion loss compared to one using a higher voltage. Higher power trunk amplifiers cost more in CATV distribution. As a result, CATV system designers are interested in adjustable attenuators with the lowest possible minimum insertion loss. However, new CATV systems tend to operate at a lower voltage and may not be able to supply the 15 V maximum control voltage required by the PIN attenuator. Mechanical attenuators formed of a set of fixed-value attenuators connected and disconnected as needed by a bank of relays would have the lowest minimum attenuation. However, the cost and the size of mechanical attenuator make it unattractive for consumer products and so it is generally found only in test equipment. When low minimum attenuation is desired in a low-voltage PIN attenuator, one possible solution is to replace the thick bulk PIN with a thin bulk PIN. The latter has lower equivalent series resistance at a given bias current, and consequently can achieve low minimum attenuation at a maximum control voltage of 5 V. Unfortunately, the thin bulk PIN may not be an ideal choice for some applications as its linearity is poorer than its thick bulk cousin. It has been shown that the minimum attenuation can be reduced in a wideband attenuator by increasing the current flow through the PIN diodes in the series arm*√*. While maintaining the control voltage (Vc) at a maximum value of 5 V, the bias current can be increased by reducing the value of the resistor in series with the control voltage (R3). In an attenuator without this minimum loss requirement, R3 is normally dimensioned to be four to six times larger than the system’s characteristic impedance, Zo, so that it can also provide RF blocking for the control voltage. However, after lowering the value of R3, an RF choke, L1, should be added in series with R3 for the purpose of RF blocking. Without L1, the RF loss will increase due to the lower value of R3. If the attenuator is meant to be used in a multi-octave application like CATV, a ferrite bead inductor will perform better than a conventional ceramicbased chip inductor. Ferrites are nonconductive ceramics that are used to increase the effective impedance of inductors. High permeability ferrite is usually chosen for the bead material so that the hysteresis losses will increase rapidly with frequency. Below the bead’s self-resonance frequency, it is usually modeled as an inductor in series with a frequency-dependent resistor[5,6]. The ferrite bead is commonly found in power amplifiers (PAs) where it is used to decouple RF from the supply line. It aids PA stability by dissipating rather than reflecting the unwanted RF energy back to the amplifier. However, it can be fruitfully transplanted into attenuator applications because its choking action spans a wider frequency range than a conventional chip inductor. Very wideband ferrite bead inductors may need to be modeled as multiple sections of RLC in order to account for the multiple resonances that manifest as a broad reactance peak (Fig. 3)*x*. The ferrite bead’s wideband characteristic is clearly shown in the manufacturer-supplied impedancefrequency graph. Since ferrite beads come with a variety of impedance characteristics due to the use of different core materials, the designer should choose the one that best matches the intended operating frequency range. For evaluating the effect of resistor R3 on the minimum attenuation, a PIN attenuator was constructed using Avago Technologies’ HSMP- 3816 thick bulk diode[8]. An initial value of 330 O was chosen for R3. As this initial resistance value is much larger than the system Zo (typically 50 or 75 O), no inductor or ferrite bead was required for choking the control voltage. Subsequently, the resistor R3 was replaced with values of 82 O and, finally, 22 O. At the smaller values of R3, a Murata BLM18RK102SN1 ferrite bead was used to provide additional choking. Figure 4 shows the minimum attenuation dropping from greater than 5 dB to approximately 3.1 dB. This result is inclusive of the losses in the FR-4-based printed-circuit board (PCB) and edgelaunched SMA connectors. The minimum loss can be further improved by either changing to a lower-loss PCB substrate or shortening the microstrip traces between the RF connectors and the attenuator. Associated Figure 1 Continue on Page 2
| |||||||||||||||||||||||||||||||||||||||||||||||||
Resources

RSS














