The LMH5485-SP offers the advantages of a fully differential amplifier (FDA) design, with the trimmed input offset voltage of a precision op amp. The FDA is an extremely flexible device that provides a purely differential output signal centered on a settable output common-mode level. The primary options revolve around the choices of single-ended or differential inputs, AC-coupled or DC-coupled signal paths, gain targets, and resistor Value selections. Differential sources can certainly be supported and are often simpler to both implement and analyze.

Because most lab equipment is single-ended, the characterization circuits typically operate with a single-ended, matched, 50 Ω input termination to a differential output at the FDA output pins. That output is then translated back to single-ended through a variety of baluns (or transformers) depending on the test and frequency range. DC-coupled, step-response testing uses two 50 Ω scope inputs with trace math. The starting point for any single-ended-to-differential, AC-coupled characterization plot is shown in Figure 8-1.

Figure 8-1 AC-Coupled, Single-Ended Source to a Differential Gain of a 2 V/V Test Circuit

Figure 8-1 shows how most characterization plots fix the R_f (R_f1 = R_f2) value at 402 Ω. This element value is completely flexible in application, but the 402 Ω provides a good compromise for the parasitic issues linked to this value, specifically:

• Added output loading. The FDA appears like an inverting op amp design with both feedback resistors as an added load across the outputs (approximate total differential load in Figure 8-1 is 500 Ω || 804 Ω = 308 Ω).
• Noise contributions because of the resistor values. The resistors contribute both a 4kTR term and provide gain for the input current noise.
• Parasitic feedback pole at the input summing nodes. This pole created by the feedback R value and the
  1.25-pF0.9-pF differential input capacitance (as well as any board layout parasitic) introduces a zero in the noise gain, decreasing the phase margin in most situations. This effect must be managed for best frequency response flatness or step response overshoot. The 402 Ω value selected does degrade the phase margin slightly over a lower value, but does not decrease the loading significantly from the nominal 500 Ω value across the output pins.

The frequency domain characterization curves start with the selections of Figure 8-1. Then, various elements are modified to show their impact over a range of design targets, specifically:

• Gain setting is changed by adjusting R_t and the 2 – R_g elements (holding a 50 Ω input match).
• Output loading, including both resistive and capacitive load testing.
• Power-supply settings. Most often, a single +5 V test uses a ±2.5 V supply, and a +3 V test uses ±1.5 V supplies.
• The disable control pin is tied to Vs+ for any active channel test.

Because most network and spectrum analyzers are a single-ended input, the output network on the LMH5485-SP characterization tests typically show the desired load connected through a balun to a single-ended, 50 Ω load, while presenting a 50 Ω source from the balun output back into the balun. For instance, Figure 8-2 shows a wideband MA/Com balun used for Figure 8-1. This network shows a 500 Ω differential load to the LMH5485-SP, but an AC-coupled, 50 Ω source to the network analyzer. Distortion testing typically uses a lower-frequency, DC-isolated balun (such as the TT1-6T) that is rotated 90° from the wider band interface of Figure 8-2.

This approach allows a higher differential load, but with a wideband 50 Ω output match at the cost of considerable signal-path insertion loss. This loss is acceptable for characterization, and is normalized out to show the characterization curves.

Figure 8-3 shows the circuit used as a starting point for time-domain or DC-coupled testing.

In this case, the input is DC-coupled, showing a 50 Ω input match to the source, gain of 5 V/V to a differential output, again driving a nominal 500 Ω load. Using a single supply, the Vocm control input can either be floated (defaulting to mid-supply) or be driven within the allowed range for the Vocm loop (see the headroom limits on Vocm in the Electrical Characteristics: VS+ −VS = 5 V tables). To use this circuit for step-response measurements, load each of the two outputs with a 250 Ω network, translating to a 50 Ω source impedance driving into two 50 Ω scope inputs. Then, difference the scope inputs to generate the step responses.  Figure 8-4 shows the output interface circuit. This grounded interface pulls a DC load current from the output Vocm voltage for single-supply operation. Running this test with balanced bipolar power supplies eliminates this DC load current and gives similar waveform results.