The mass defect can’t be positive. If it was, that would mean the assembled atom would have more total mass-energy than its constituent particles, which would mean that it could (and therefore would) fly apart spontaneously.
What’s happening is that the mass of your typical nuclide is the mass of a neutral atom of that isotope. The mass of a proton and a neutron combined is 1.007825 Da (for a neutral atom, the number has to be the same, so we can group them together); and the mass of a neutron is 1.008665 Da. That means for your typical neutral atom with mass number A, the mass of the constituent particles will about 0.8% greater than A daltons. But because of the mass defect, the total mass of the bound atom is less than the total mass of its consistuents; and for many nuclides, the mass defect is big enough that the total mass of the bound nucleus is less than A daltons.
The example that I know of is Bismuth-209, which had its half-life estimated to be about 4.6 \times 10^{19} years but was not actually observed to decay until 2003, when its half-life was found to be closer to 2.0 \times 10^{19} years.