The ROSAT PSPC pulse height spectrum of the peculiar He-rich hot white dwarf KPD 0005+5106 provided a great surprise when first analysed by Fleming , Werner & Barstow ( 1993 ) .
It defied the best non-LTE modelling attempts in terms of photospheric emission from He-dominated atmospheres including C , N and O and was instead interpreted as the first evidence for a coronal plasma around a white dwarf .
We show here that a recent high resolution Chandra LETGS spectrum has more structure than expected from a thermal bremsstrahlung continuum and lacks the narrow lines of H-like and He-like C expected from a coronal plasma .
Moreover , a coronal model requires a total luminosity more than two orders of magnitude larger than that of the star itself .
Instead , the observed 20-80 Å flux is consistent with photospheric models containing trace amounts of heavier elements such as Fe .
The soft X-ray flux is highly sensitive to the adopted metal abundance and provides a metal abundance diagnostic .
The weak X-ray emission at 1 keV announced by O ’ Dwyer et al ( 2003 ) instead can not arise from the photosphere and requires alternative explanations .
We echo earlier speculation that such emission arises in a shocked wind .
Despite the presence of UV-optical O VIII lines from transitions between levels n = 7 -10 , no X-ray O VIII Ly \alpha flux is detected .
We show that O VIII Lyman photons can be trapped by resonant scattering within the emitting plasma and destroyed by photoelectric absorption .