Many RF systems require distribution of an input signal from an antenna to feed multiple individual receivers. System designers use multi-couplers to accomplish this while multiplexing different paths into different bandwidths, maintaining the desired signal power levels, and circumventing unwanted effects depending on the specific system.

Imagine the challenges of designing and building a worthy 16-way receiver multicoupler in the 1990s.  For those of us interested in history, or perhaps old enough to remember designing in the 1990s, some of the best silicon bipolar transistor MMICs were supplied by Avantek (acquired by HP in 1991) and the most popular design software was a program called Touchstone, supplied by EEsof (acquired by HP in 1993). Reactive splitters had long been the prevailing signal distribution component for receiver multicouplers because of their low loss, broadband and rugged nature.1  These quarter-wavelength reactive devices could be designed to split HF, VHF and UHF frequencies evenly and contiguously to the output ports with very low intermodulation distortion due to their high power handling capability.  Many of the MMIC amplifier devices available in that day and age exhibited too great a noise figure and lower linearity than required for most receiver applications. Consequently, discrete transistors were often designed into the multicoupler to meet a given combination of specifications.

About three decades since that time, a customer approached Mini-Circuits with a project to redesign just such a 16-way receiver multicoupler for a SATCOM application originally produced in the ‘90s. The system was still in service, but many of the components in the assembly were obsolete and the workforce no longer knew how to build the clunky, magnetic splitter that underpinned the legacy design.

Moreover, vast improvements in RF components had been achieved in Mini-Circuits’ product line and the industry at large. MMIC amplifiers that exhibit 1 dB or less noise figure and IP3 levels greater than +30 dBm are now commonplace. Matched amplifier pairs integrated on the same die and combined with baluns in push-pull configurations can routinely achieve IP2 levels of +60 dBm or greater across more than a decade of bandwidth. These innovations had taken place at Mini-Circuits across so many product lines that the receiver multicoupler redesign effort that ensued needed only focus almost exclusively on Mini-Circuits’ components.

Design Specifications

The design specifications for the multi-coupler redesign effort were as follows:

  • Frequency Range:  50~1500 MHz
  • Gain:  5 ± 1 dB
  • Noise figure: 10 dB max
  • OP1dB: +5 dBm min
  • OIP3: +20 dBm min
  • OIP2: +45 dBm min

The specifications are challenging, given that the input signal is split 16 ways, and perhaps the first parameter that jumps out at a designer is the relatively high OIP2 requirement. This implies that the amplifiers utilized throughout the design are most likely the push-pull configuration. One simple rule-of-thumb says that that OIP2 for an amplifier-based system is typically 10 dB greater than OIP3. In this case, OIP2 is a whopping 25 dB greater than OIP3. Consequently, the need for even order harmonic-cancelling, OIP2-enhancing push-pull amplifiers is a certainty. This conclusion is a good starting point since it narrows down the selection of amplifiers for the design.

Component Selection

Component selection for this redesign began with a priori knowledge of return rates and root cause for the predecessor (1990s vintage) line of multicouplers. The most common return was for a damaged input LNA due to the user having exceeded the maximum input power, which was not surprising as the predecessor design had not incorporated a limiter. Vowing to not let this shortcoming drive the return rate for the new design, a specialty ultra-high-linearity limiter was selected for the front end which realizes 0.35 dB of insertion loss interpolated at 1.5 GHz. The limiter is shown on the lefthand side of the block diagram of Figure 1.

Block diagram of a 16-way receiver multi-coupler.

Figure 1: General block diagram of a 16-way receiver multicoupler designed with Mini-Circuits parts.

The LNA for the new multicoupler is selected from Mini-Circuits’ line of dual matched amplifiers. The PHA-11+ has a noise figure that linearly interpolates to 2.03 dB at 1.5 GHz. As shown on the data sheet, when operating as a push-pull pair utilizing input balun TCM2-33WX+ and output balun TCM2-43X+, the PHA-11+ delivers an OIP3 of 39.7 dBm and an OIP2 of 60.8 dBm, also at 1.5 GHz. These performance levels start the lineup off on a path to success by incorporating plenty of margin before the 16-way power split.

The splitter presents an interesting choice; resistive, reactive, or Wilkinson. Resistive splitters are very broadband, of course, extending down to DC, but they exhibit loss on the order of 21 dB, which appears untenable for these tight design constraints. A reactive splitter exhibits low loss (less than 3 dB over theoretical) and low amplitude and phase imbalance over broad bandwidth. However, these designs can be bulky and costly, especially if an attempt is made to custom-design one. A Wilkinson splitter strikes a balance between the other two choices. Its primary shortcoming – low power handling – is overcome by the fact that it is driven by a push-pull LNA (PHA-11+) with an OP1dB of approximately 24 dBm, or 0.25W.

While it is possible to design a custom multi-section Wilkinson power divider to cover the 50 MHz to 1.5 GHz band, in this redesign effort, thanks to the availability of samples from the local rep, the Mini-Circuits Z16PD-252-S+ was found to be an ideal fit. Over the band of interest, the splitter exhibits less than 4 dB over theoretical loss at 1.5 GHz and a power handling of 1W, which more than accommodates the 24 dBm input signal from the LNA. Once the splitter is added to the cascade, the gain of each of the outputs becomes too low, and additional amplification is needed on each channel.  Keep in mind that any gain stages must have premium OIP3 and OIP2 to meet the system requirement, so the customer once again turned to Mini-Circuits’ line of dual matched amplifiers.

From a linearity perspective, the MPGA-105+ dual-matched pair is impressive, boasting greater than 34 dBm OIP3 and over 55 dBm OIP2. The gain has a slight roll-off from 50 MHz to 1.5 GHz and the design is physically quite compact, measuring only 5 x 6 x 0.9mm. space economy is important at this stage of the cascade as this amplifier must be repeated 16 times. Following each of these amplifiers is the commensurate fixed slope equalizer to flatten the gain roll-off. Mini-Circuits offers a wide selection of MMIC fixed slope equalizers with different frequency ranges and slope values in 2x2mm QFN and bare die, but in this case, the equalizer was realized utilizing discrete parts, a three-resistor tee attenuator with capacitors in parallel with the series resistors and an inductor in series with the shunt resistor. In this case specifically, a 9.6 dB three-resistor tee attenuator with 11 pF capacitors and an 82 nH inductor directly on the customer’s board. 

The final component in the cascade is a simple fixed attenuator again realized utilizing discrete chip resistors. Chip resistors provide for a broad selection of values in order to set the attenuator level to center the gain in the 5 ± 1 dB specification window. The actual attenuation needed was 1.5 dB. For customers less inclined to engineer this block for reasons of repeatability, board layout or sourcing complexity, Mini-Circuits offers MMIC fixed attenuators for a vast range of requirements.

The Final Word

One final word about component selection. The design described in this application note is presented as if it rolled off a log. In reality the parts selected for the design were ultimately chosen because they survived the rigors of cascade analysis where many parts were evaluated. The effort to achieve the delicate balance that brings all system parameters into specification compliance usually takes some trial and error through several stages of simulation, prototyping and design validation.

For readers interested in this analysis subsequent application notes will walk through the analysis of Gain, NF, P1dB, IP3, and IP2 used in the selection process for this design.

Components Referenced in This Article

%d bloggers like this: