Example 08 zero calibration does not affect displayed Amps value

從 Example08_DiagnosticTool 中的 doZeroCalibration 和 z/r 命令處理常式開始；比較本機的 zeroOffset 路徑與 mySensor.getCurrentA(amps)。確認函式庫的校正 API 使用一致，並確認歸零或重設後，顯示的 Amps 和 raw 欄位反映校正結果。

uint32_t corrected =
    (rawIrms > zeroOffset) ? (rawIrms - zeroOffset) : 0;

In Example 08 (`Example08_DiagnosticTool`), the `z` command calculates a local
`zeroOffset` from raw IRMS readings. That offset is applied to the displayed
`Corrected` and `Averaged` raw values as above.

However, the displayed current in amps is obtained independently using:

mySensor.getCurrentA(amps);

Consequently, zero calibration does not affect the Amps field. For example,
after zeroing, Corrected may be near zero while Amps continues to show the
original no-load value.

Also, linear subtraction of two RMS magnitudes is not generally the correct
noise-floor correction. The library already provides autoCalibrateA(), which
performs the correction in the squared/RMS domain.

Would you consider changing the z command to use the library calibration?

For example:
void doZeroCalibration()
{
    const uint16_t NUM_SAMPLES = 50;

    Serial.println(
        F("\nZero calibration - keep the CT connected with no current flowing."));

    if (mySensor.autoCalibrateA(NUM_SAMPLES) != ksfTkErrOk)
    {
        Serial.println(F("Zero calibration failed!"));
        return;
    }

    resetBuffer();
    Serial.println(F("Channel A zero calibration complete.\n"));
}

The r command could correspondingly call:

mySensor.clearCalibration();

This would make the zero operation affect getCurrentA(amps) directly and
would avoid having two independent calibration paths in the example.