What about a beam of high intensity ionising radiation aimed at the detector? Since we're dealing with lo intensity sources, this could force the RNG into a small part of its range. Beware of black helicopters carrying cyclotrons :-)
Thermal noise is a much less hazardous mechanism. Ionising radiation is a pain to deal with, it is something one generally attempts to get rid of in a fabrication process if possible. Environmental radiation is a strong polluting source. There are plenty of other quantum phenomena which can be tapped. Essentially all one needs to do is to build a very high gain amplifier "baddly". The main difficulty is removing bias. One can either go through filtering hohaa or use say MD5. The trully paranoid could Xor together an odd number of MD5 samples just in case there were bit pattern dependencies. Phill