The summary in Table 2-1 is based on the results of multiple simulations where the ALM2403-Q1 was used for the buffer amplifiers and an OPA210 acted as the error amplifier. A complex load of 20Ω || 820pF was utilized. To compare the bandwidth, both circuits were constructed such that the trace resistances (10mΩ), ballast resistances (1Ω), and all other component values were matched between the two channels. A stability analysis was performed on each and the feedback capacitances were selected (using standard component values) for stability. The -3dB bandwidth of the circuits (measured for an AC input to the error amplifier) were then recorded. Note that the qualitative comparison below assumes the same R_ballast value is used for each circuit.

For this load example, the conventional buffer arrangement was found to result in approximately 21x higher -3dB bandwidth versus the parallel improved Howland pump arrangement (with gain R_fX / R_inX = 0.1), despite using the same C_F and C_fx values. Correspondingly, the settling time of the parallel improved Howland pump circuit for a 50mV input step was significantly longer.

However, it was noted during further tests that as the load capacitance increased tenfold to 8.2nF, the parallel improved Howland pump circuit remained stable without requiring any modifications, whereas the conventional buffer arrangement exhibited a ringing response and poor stability unless the ballast resistances were significantly increased to 10Ω. In fact, the same unaltered parallel improved Howland pump circuit was able to drive as much as 50nF of load capacitance and remain stable. This implies that under certain load conditions the parallel improved Howland pump circuit can actually prove more adaptable and resilient to capacitive loading effects than the conventional buffer. As stated earlier, however, this is highly dependent on the specific circumstances of the application and the amplifiers used.