It looks interesting. I would like to hear what you find out.
Frequency multipliers other than 2, if my understanding is correct, need tight band-pass filters to get rid of the unwanted harmonics. The frequency multiplication could be done in a PLL but then that would be difficult with the wide input frequency range of 1MHz to 100MHz.
EDIT: realise I read the thing too quickly - the factor is 260 not 48. This seems unlikely to be frequency multipliers as it contains the factor 13 - a 13 times multiplier would require filtering of the much higher content at 3, 5, 7, 9 and 11 times and I can't see this working over the range 1 - 100MHz.
For simply multiplying the phase noise (frequency deviation) then I would have thought that a PLL approach would be cleaner as in :
http://www.wriley.com/The%20PicoMult%20Frequency%20Error%20Multiplier.pdfbut that approach only works for a specific frequency input.
So I am puzzled and hope you can get some further information from the designer.
EDIT: my best guess is that a frequency synthesizer is used with a programmable PLL but a means of knowing the input frequency would be needed.