increased PDFCalculator.rstep gives very different PDF

This stanza gives much different PDF values although they should not depend on r-sampling.  Find out what is going wrong with qmax cutoff and/or rstep interpolation.
```python
pc = PDFCalculator(qmax=11, rstep=0.01)
ni = loadStructure('nickel.cif')
r0, g0 = pc(ni, rstep=0.01, qmax=11)
r1, g1 = pc(ni, rstep=numpy.pi / 11, qmax=11)
```