SDAA090 November   2025 INA234

 

  1.   1
  2.   Abstract
  3.   Trademarks
  4. 1Introduction
  5. 2Current Sense Amplifier Based on Chopping Techniques
  6. 3Aliasing in CCCSA
  7. 4Input Filtering
  8. 5Summary
  9. 6References

Input Filtering

An output filter can be effective if there is enough frequency separation between the modulated tones and the desired signal; otherwise, the output filter is not very effective.

Due to factors such as device variation, input signal variation and temperature effect, it is safe to assume that the modulated tones will show up at any frequencies. For this reason, an input filter can be adopted to minimized the risk of aliasing. Figure 4-1 shows a low-pass filter for anti-aliasing, as recommended in the INA190 datasheet. The filter is implemented with two resistors, RF, and one capacitor, CF. This filtering scheme can be adopted for INA186 and INA191 when application requires. These devices are based on the same capacitively coupled chopping architecture.

 Low-pass Filter for
          Anti-aliasing Figure 4-1 Low-pass Filter for Anti-aliasing

Figure 4-2 shows the effect of such an input filter. In this experiment, the filter is only added to the input of Unit 2, with RF = 100Ω and CF = 0.1uF.

The filter corner frequency is approximately 8kHz, providing enough attenuation for the 800kHz input noise. As a general rule, the corner frequency is selected to be the lowest possible allowed by measurement speed requirement. Also, according to Nyquist theorem, the corner frequency needs to be below half the chopping frequency. For the INA190 and family, specified signal bandwidth is well below the chopping frequency.

As Figure 4-2 shows, with anti-aliasing filter in place, Unit 2 output becomes very clean. FFT confirms that the harmonics are effectively removed.

 Input Filtering on Unit 2 Figure 4-2 Input Filtering on Unit 2

Digital power monitors, if based on the same capacitively coupled chopping techniques, can suffer from similar aliasing effect. Similar to the analog output counterparts, the symptom includes elevated offset, or slow varying offset that is uncharacteristic of random noise. The same input filtering scheme can be employed for anti-aliasing purposes. For example, the INA236 data sheet has a section dedicated to input filtering. It is worth noting that aliasing effect is distinguished from the intrinsic device noise, which is not affected by input filtering.