Mixers Terminate Intermod Distortion

Dec. 17, 2009
The two-tone-terminator (T3) line of diode mixers offers excellent suppression of IMD while providing high performance levels with both sine-wave and square-wave local-oscillator signals through 12 GHz.

Minimizing intermodulation distortion (IMD) in mixers can benefit many commercial and military RF/microwave systems. For that reason, Marki Microwave has developed its two-tone-terminator (T3) line of mixers for superior suppression of two-tone IMD. The T3 mixers cut conversion loss to typically only 7 dB from 10 MHz to 12 GHz, with input third-order intercept (IIP3) points that are typically 10 dB (and as much as 20 dB) higher than the local oscillator (LO) drive level. Conversion loss is constant for drive levels from +12 to +25 dBm, and the 1-dB compression point is high, only about 1 to 3 dB below the level of the LO drive. Also, the performance of a T3 mixer can be optimized via the use of a square-wave LO drive signal.

The T3 mixers employ internal feedback circuitry and commutating (switching) functionality to achieve near-ideal IMD behavior compared to conventional double-balanced diode mixers. When combined with square-wave LO drive, the linearity of the commutating mixer is additionally improved because the switching is allowed to occur nearly instantaneously. Traditional double-balanced mixers are optimized for use at LO levels over about a 3-dB range. The passive feedback circuitry of the T3 mixers reduces their sensitivity to LO drive level. Essentially, the switching level of a T3 mixer is adjusted automatically according to the LO drive, so that a T3 mixer can be used with LO drive of +15 and +25 dBm without affecting conversion-loss performance. As LO drive is increased, the 1-dB compression and IMD performance for a T3 mixer actually improves significantly.

In a commutating mixer, when the LO alone (without contributions from the RF signal) controls the mixer's diode conductance, the mixer's behavior approximates an ideal switching function.2 The faster the LO passes through the diode transition region, the more linear the mixer becomes. In a commutating mixer, the role of the LO signal is to turn the mixer diode fully on or fully off. The time required to turn the diode on and off is associated with the rise time of the LO waveform. A common sine wave LO drive waveform has a rise time on the order of 30 percent of the total period (e.g., 300 ps for a 1 GHz sine wave LO). By using a square-wave LO signal with faster rise/fall time, IMD can be greatly reduced in the diode transition region.

In evaluating the performance of a T3 mixer compared to a traditional double-balanced mixer, conversion loss is a good place to start. Good conversion-loss performance tends to imply good performance in other mixer parameters, such as isolation and 1-dB compression. When a mixer's conversion-loss performance is poor, the other mixer metrics are often inferior as well.

Figure 1 compares the conversion loss of a T3 mixer with that of an M1 double-balanced mixer, with a focus on the popular 2-to-4-GHz band. Conversion loss was measured for constant IF downconversion, with RF and LO signal sources swept with a constant 15-MHz offset frequency. Each mixer was measured with both a sine wave and a square-wave LOthe sine wave LO was generated by a frequency synthesizer and the square-wave was generated by a model AP-0010EZP single-positive-bias square-wave amplifier from Marki Microwave. The amplifier's output can be tuned by changing the single positive bias from about +4 to +8 VDC.

When evaluated with a +9-dBm sine wave LO signal, the conversion loss of the double-balanced mixer (based on a low-barrier-height diode quad) is less than that of the T3 mixer due to its highly efficient, low loss design. Even when the double-balanced mixer is overdriven with a +19-dBm sine wave, the conversion loss is an outstanding 5.5 dB. With a square-wave LO applied, however, the conversion loss of the double-balanced mixer suffers. For the highest square-wave drive measured (+19 dBm), the conversion loss for the double-balanced mixer increased by as much as 2 dB. In comparison, the T3 mixer has much less sensitivity to LO drive. For all drive levels, whether with sine-wave or square-wave LO sources, variations in conversion loss for the T3 mixer are less than 1 dB.

A mixer's 1-dB compression point relates to how well the LO signal is able to control the conductance of the mixer's diodes, even in the presence of a strong RF signal.1 Mixer compression is evidenced by an increase in the conversion loss. When the compression points of T3 and double-balanced mixers were compared as a function of LO drive, for both sine wave and square-wave LO signals, the two mixers demonstrated much different behavior (Fig. 2).

For the double-balanced mixer, the 1-dB compression was found to be very weakly dependent on LO drive. As the LO level to the double-balanced mixer was swept from +10 to +20 dBm, the 1-dB compression point improved by, at best, 2 dB. It is generally understood that the 1-dB compression point of a double-balanced mixer will be 5 to 7 dB less than the highest recommended LO drive level for that mixer. For the double-balanced mixer with low-barrier-height diodes used in this comparison, +10 dBm is the maximum recommended LO drive level, implying that the 1-dB compression point for that mixer will be between +3 and +5 dBm, confirmed by test results.

For the T3 mixer, the 1-dB compression point improved as the LO drive level was increased. For a sine-wave LO drive, the T3 compression point improved about 0.7 dB for every 1 dB increase in LO drive. When a square-wave LO was applied, the impact on 1-dB compression was even more striking. For the T3 mixer, use of a square-wave LO raised the 1-dB compression point by 3 to 4 dB compared to the results for a sine-wave LO. The improvement in compression-point performance was so dramatic with the square-wave LO, that the dynamic range of the measurement equipment was limited in its capability of fully evaluating the compression performance of the T3 mixer for drive levels above about +18 dBm.

Double-balanced mixers are designed to cancel even-even, even-odd, and odd-even single-tone IMD products, although the level of cancellation depends upon the balance of the mixer circuitry.2 But the T3 mixer is actually designed to prevent the generation of single-tone spurious products as well as suppress them. Cancellation using circuit balance adds an extra 25 to 35 dB mixer- and frequency-dependent spurious suppression.

Unlike single-tone IMD, multitone IMD is solely attributed to the nonlinear characteristics of the mixer diodescircuit balance does not decrease multitone spurious levels. To reduce multitone IMD levels and achieve good input third-order intermodulation (IIP3) performance, a mixer can be designed with high-barrier-height diodes or with commutating diodes. Figure 3 shows how the IIP3 performance of a T3 mixer compares to that of a double-balanced mixer with low-barrier-height diodes. The performance of the T3 mixer clearly improves by increasing the LO drive, while the IIP3 performance of the double-balanced mixer remains flat around +10 to +12 dBm for all LO drives.

When a square-wave LO is applied, the T3 outperforms the double-balanced mixer in IIP3 performance by a significant margin. In general, the IIP3 of the T3 mixer is at least 10 dB above the LO drive level to the mixer. For extremely high LO square-wave drives of +25 dBm, the IIP3 performance of the T3 mixer exceeds +35 dBm. For the double-balanced mixer, the IIP3 tends to be about 1 to 2 dB above the maximum recommended LO drive level. For the double-balanced mixer in this study, the maximum recommended drive is +10 dBm, so that the IIP3 for that mixer is about +11 dBm.

Continue to page 2

Page Title

To give double-balanced mixers their due, the IIP3 performance levels of several different double-balanced mixers with different diode barrier heights (or diode forward voltage, Vf) were evaluated for comparison. Traditionally, the diode barrier potential is used to control 1-dB compression, two-tone performance, and the amount of LO drive required. The higher the diode barrier height, the more LO drive is required. Prior to the T3, high IIP3 mixers (i.e., IIP3 of more than +20 dBm) required high-level diodes because there was no other convenient way to prevent the RF-driven intermodulation in the transition region of the diode. The T3 mixer has removed these constraints for achieving IIP3 performance, outperforming all diode-based double-balanced mixer options currently available (Fig. 4).

To evaluate the IIP3 performance of a T3 mixer as a function of square-wave rise time, a Microwave Wavefade lowpass filter3 from Marki Microwave was used to modify the rise times of pulsed signals from the model AP-0010EZP amplifier. The Wavefade filter is a quasi-infinite-element Bessel-shaped filter with flat group delay performance. By connecting Wavefade filters to the AP-0010EZP amplifier with cutoff frequencies ranging from 2.8 to 17.5 GHz, it was possible to control the rise time of the LO waveform to the T3 mixer. Unfiltered, the AP-0010EZP amplifier yields a rise time of 18 ps at 3 GHz. With a Wavefade filter with 7.47-GHz cutoff frequency connected to the amplifier, the rise time increases to 48 ps. When the Wavefade filter cutoff is reduced to 4.65 GHz, the rise time increases to 80 ps.

Figure 5 shows the two-tone IIP3 performance of a T3 mixer as a function of LO square-wave rise time. Unfiltered, the fast-rise-time square-wave LO drive yields two-tone performance exceeding +32.5 dBm. As the LO drive is more heavily lowpass filtered, the IIP3 level decreases. As a limit, as the square wave's higher-order harmonics are removed by the lowpass filter, the square wave is reduced to a sine wave. In Fig. 5, a 3-GHz sine wave corresponds to a rise time of about 100 ps and a two-tone intercept of about +22.5 dBm. Therefore, the IIP3 penalty for not using a square-wave LO with a T3 mixer can be as high as 10 dB. But even with a suboptimal sine wave, the performance of the T3 mixer exceeds that of a double-balanced mixer by as much as 5 to 7 dB. Marki Microwave, 215 Vineyard Court, Morgan Hill, CA 95037; (408) 778-4200, FAX: (408) 778-4300, e-mail: [email protected], Internet: www.markimicrowave.com.


1. Ferenc Marki and Christopher F. Marki, "Mixer Basics Primer," Application Note, Marki Microwave, Morgan Hill, CA, www.markimicrowave.com/menus/appnotes/mixer_basics_primer.pdf.

2. Bert C. Henderson, "Predicting Intermodulation Suppression in Double-Balanced Mixers," Watkins-Johnson Co. Technical Notes, Vol. 10, No. 4, July/August 1983.

3 Christopher F. Marki, Wavefade Filters Product Overview, Application Note, Marki Microwave, Morgan Hill, CA, www.markimicrowave.com/menus/appnotes/ wavefade_product_overview.pdf.

Sponsored Recommendations

Getting Started with Python for VNA Automation

April 19, 2024
The video goes through the steps for starting to use Python and SCPI commands to automate Copper Mountain Technologies VNAs. The process of downloading and installing Python IDC...

Can I Use the VNA Software Without an Instrument?

April 19, 2024
Our VNA software application offers a demo mode feature, which does not require a physical VNA to use. Demo mode is easy to access and allows you to simulate the use of various...

Introduction to Copper Mountain Technologies' Multiport VNA

April 19, 2024
Modern RF applications are constantly evolving and demand increasingly sophisticated test instrumentation, perfect for a multiport VNA.

Automating Vector Network Analyzer Measurements

April 19, 2024
Copper Mountain Technology VNAs can be automated by using either of two interfaces: a COM (also known as ActiveX) interface, or a TCP (Transmission Control Protocol) socket interface...