[Laser] Is speed of photodiodes inversely proportional to sensitivity?

[Chris L]

   Dear laser group,

   Of  recent  years, I've chiefly used photodiode types BPW34, BPX65 and
   IPL10040DW.

   We  had  the  BPW34  and  IPL10040DW  operating behind fresnels in two
   simultaneously  operating  receivers on Mount Barrow (Tasmania) during
   our  104  mile  contact  on  19 February. The BPW34 is a cheap-looking
   square  photodiode  of  about  7 sq mm area with plastic encapsulation
   - an utterly unimpressive-looking device selling here for about US$2 -
   and  I  think  it  has  been available for about the last 20 years. By
   contrast, the  IPL10040DW is a fancy looking gold-plated device, metal
   TO5 body with an elegant tiny glass window and sensitive area 4 sq mm,
   selling here for about US$20.

   The interesting thing is that when one looks into the BPW34 one sees a
   square,   completely   black  silicon  sensitive  surface,  while  the
   IPL10040DW  reflects  light from its sensitive surface somewhat like a
   mirror.  The  BPW34  is  rated  at  approx 100 nS rise time, while the
   IPL10040  is  much faster - around 9 nS rise time. The latter can also
   take  a  much higher reverse voltage to speed the device up and reduce
   junction  capacitance,  so  it would be the ideal choice for video etc
   while the BPW34 seems a better prospect for baseband audio modulation.

   The  surprise on 19 February is that the el-cheapo BPW34 gave us about
   (by  ear  estimation)  10  dB  better  sig/noise  than  the  expensive
   IPL10040DW.  In  both cases, the optical system, alignment and pre-amp
   were  identical.  Has  anyone  an  explanation?  Are  fast photodiodes
   processed with a thinner, less sensitive silicone layer?

   I  actually  bought  the  IPL10040DW  because of the third lead on the
   package,  the  lead enabling its metal body to be earthed. We do a lot
   of  our DX tests on mountains with road access having messy sources of
   rf  (or  even  public  TV  broadcast  transmitters)  on their summits.
   Operating a high impedance front end at high amplification levels near
   these  rf  hash  sources,  especially  TV  transmitters, is a bit of a
   problem...  as  you  might  imagine!  Earthing everything in sight and
   bypassing  all  amp  power and output leads with .001 ceramics is also
   desirable.

   We  could not use the BPX64 on our fresnel receivers, even though they
   were  faster  than  either of the above photodiodes. Unfortunately the
   BPX65  has a sensitive surface recessed deep inside the metal package.
   This  limits  its angle of view to a small portion of the central part
   of  our  f1  fresnels, so the BPX65 is optically incompatible with our
   application.

   Anyway,  can anyone tell me the reason for the sensitivity difference?
   I  cannot  see much in the specs to account for this - or maybe I just
   have a bum IPL10040DW !!! Unfortunately, our logging tape on the night
   of  19  Feb was taken from the receiver with the IPL photodiode, so it
   gives nowhere near a good impression of the excellent results achieved
   on  the  night  -  so I guess we're going to have to do the whole damn
   thing again to get a better log tape! (Not that I'm complaining - it's
   all good fun!)

   Has  anyone  else  found  these significant differences in sensitivity
   between different photodiodes?

   Any ideas, you photodiode boffins out there?

   All the best,

   Chris Long (Melbourne Australia).