HVA Wideband Voltage Amplifier

Select the narrowest model that still passes the highest frequency and edge rate the measurement must preserve. The 10 MHz variants have a 35 ns rise/fall time and suit slower pulses, detector outputs and transient measurements that do not require hundreds of megahertz of bandwidth. The 200 MHz models reduce the rise/fall time to 1.8 ns, while the HVA-500M-20-B provides a 500 MHz upper cut-off frequency and a 750 ps rise/fall time. Choosing 500 MHz when 10 MHz is sufficient exposes the measurement chain to more wideband noise. In practice, compare the required rise time, signal harmonics and instrument bandwidth, then retain suitable margin without automatically choosing the widest model.

Choose a B-input model for impedance-matched 50 Ω systems and an F-input model when the signal source must not be heavily loaded. The B variants provide lower voltage-noise densities, including 0.9 nV/√Hz for the 10 MHz model at 60 dB and 1.2 nV/√Hz for the 200 MHz model at 40 dB. The F variants have a 1 MΩ input and much lower bias currents of 2 pA or 10 pA, but their voltage-noise densities are higher at 4.7 and 4.5 nV/√Hz respectively. They also carry a 200 V transient rating when the transient originates from a 200 pF source. The correct choice therefore depends on source impedance, allowable loading, bias-current sensitivity and required noise performance.

Use DC coupling when the absolute signal level, pulse baseline or low-frequency content must remain intact. The 10 and 200 MHz models can also be switched to AC coupling, with a 1 kHz lower cut-off frequency on B-input versions and a 1 Hz lower cut-off on F-input versions. AC coupling can remove a steady DC component, but it may introduce baseline movement or pulse undershoot when the signal contains long pulses or changing digital patterns. The 500 MHz model is DC coupled only, which supports sub-nanosecond edge measurements without an AC-coupling high-pass response. The adjustable output offset should be used to position the baseline while retaining sufficient output headroom for the expected signal excursion.

Yes, the amplifiers have a 50 Ω output and are intended to operate with a 50 Ω load for their specified performance. The 10 MHz models provide a linear output range of ±3.5 V into 50 Ω, while the 200 and 500 MHz models provide ±1 V. Maximum output current is 100 mA for the 10 and 500 MHz versions and 60 mA for the 200 MHz versions. A properly terminated 50 Ω coaxial signal path also reduces reflections when measuring fast edges and short pulses. The oscilloscope may use its internal 50 Ω termination or an appropriate external terminator, but the resulting load and signal amplitude should be checked before connection.

The approximate input limit can be estimated by dividing the linear output range by the selected voltage gain. For a 10 MHz model, the nominal limits are about ±35 mV at 40 dB gain and ±3.5 mV at 60 dB gain, based on its ±3.5 V output range. The 200 MHz variants correspond to approximately ±100 mV at 20 dB and ±10 mV at 40 dB because their output range is ±1 V. The 500 MHz model also corresponds to roughly ±100 mV at its fixed 20 dB gain. These calculated values are ideal limits, so practical designs should leave headroom for input offset, DC signal components, noise, overshoot and output-offset adjustment.

Each model requires a regulated bipolar supply of ±15 V through a 3-pin power connection. Pin 1 carries +15 V, pin 2 carries −15 V and pin 3 is ground, so polarity and grounding should be checked before power is applied. Typical current is stated as ±70 mA for the 10 and 200 MHz models and ±40 mA for the 500 MHz model, while a minimum supply capability of ±150 mA is recommended. The ±20 V absolute maximum rating is a protection limit rather than an alternative operating voltage. A low-noise supply and short, properly routed connections can help prevent power-rail interference from entering a small-signal measurement.

Input voltage-noise density is useful for comparing front ends, but it should not be considered separately from bandwidth, source impedance and integrated noise. At their higher gain settings, the 10 and 200 MHz B-input models specify 0.9 nV/√Hz and 1.2 nV/√Hz respectively, whereas the high-impedance F versions specify 4.7 and 4.5 nV/√Hz. Integrated input noise increases from 20 µV peak-to-peak for the 10 MHz B model at 60 dB to 150 µV for the 200 MHz B model at 40 dB and 0.5 mV for the 500 MHz model. This illustrates why unused bandwidth can raise the total noise entering the measurement. Model selection should therefore balance bandwidth, gain, source loading and the smallest signal that must be resolved.