The virtual DC tube method

Physically, having a dc tube between the source tube and object under test is inconvenient. Also the join between source tube and dc tube will not be perfectly smooth, so a small reflection will be present at the start of the input impulse response. As discussed in section 5.8 the result is that only half of the available two milliseconds in the input impulse response data is used in the dc offset calculation. As an alternative we present a new method described as the virtual dc tube method [62]. The effect of the dc tube is simulated by starting recording the reflections from the object under test 2ms earlier, using a digital filter (see section 5.4.2) to add the losses that would have occurred if the sound had travelled across a dc tube.

In effect, the last 40cm of the source tube has been turned into a virtual dc tube, perfectly joined onto the source tube. Notice how about 2ms into the dc tube method reflections in figure 5.11 there was a small reflection from the join between the source tube and the dc tube. The object reflections measured using the virtual dc tube method from figure 7.1 show that the problem has been avoided entirely.

Figure 7.1: Object reflections using virtual dc tube

A bore reconstruction achieved using the virtual dc tube method is shown in figure 7.2. The reconstruction is much the same as that achieved using the dc tube method (see figure 5.12) except that the small error of about 0.1mm in the average value of the radius of the last cylindrical section is absent in the new method. Because the frequency bandwidth of the measurement is not altered, the oscillations at the changes of cross-section and at the open end are of the same size irrespective of whether the virtual dc tube method is used.

Figure 7.2: Reconstruction of test object consisting of stepped cylinders