Salas SSLV1.3 UltraBiB shunt regulator

I know c3 values are sensitive due to oscillation. Is wiring a low value film bypass cap only advisable on c2 or can this be done on c3 also? If wiring on c2, would this be in parallel to an elctrolytic and, if so, is the best way to wire it to the pins on the bottom side of the regulator?
 
Thanks. It seems that for loads <300mA MUR120 is faster than the MSRF860G but with slightly higher noise. For ultrabibs used for analog outputs stages or powering digital components (all with <300mA) would you expect there to be any discernable difference?
 
The 120 because small is certainly the faster, MSRF is TO-220 slower but also soft recovery so it's possibly a better compromise when for digital. Especially for clocks. The rf small value resistor does some RC EMI filtering with C1 in all cases though. I would be more inclined about deringing the Tx & diodes combination than finding the super suitable bridge diode type if building detailed.
 
Salas, I am populating Ubib board for a supply 24v./1,5A.
I am concerned M1 in it's limits using high speed Fairchild FQP3P20.
What can I expect if exceeded, a burn?
I have not in parts bin a IRF9630, but seen a pair IRF9530, maybe still useful in this application. Does it slows much the performance?
Jordi
 
9530 is even slower unfortunately. FQP can withstand if well sinked. Worse thing to happen is baked until failure. In CCS it usually breaks open and cuts the circuit on fail because M1 is in series. Not big trouble. Not guaranteed but usually fails open, not shorted. Don't forget to fuse in any case.