[Laser] OPT-101 vs PIN diode

[F1AVYopto at aol.com]

Dans un e-mail daté du 14/12/2006 19:05:28 Paris, Madrid, KY1K at verizon.net  a 
écrit :

>Rather than 'improving' the bandwidth without 
>adding the noise  component of a large value 
>resistor, I'd think that 'controlling' the  bandwidth would be most useful.

Art
 
You are fully right.
"Controlling" is better because all the feed back  systems avoid partially 
the capacitances but reduce the open loop gains and flat  the open loop 
bandwidth.
 
>I'm curious though....the FET has a very large 
>(100 megohms  +/-) input impedance, wouldn't the 
>presence of this large value resistor  built into 
>the FET contribute a noise itself, that would be impossible  to negate?

The basic idea of the PGP design is the use of the  biggest Rload as possible 
for the photodiode.
In the PGP front end  this Rload is composed by the internal FET gate 
resistor (FET gate leakage) in  parallel with the internal photodiode resistor.
With a good FET at low Vg  bias, R gate can be much above 10 Gohms.
The internal Rshunt of a good  photodiode near the photovoltaic mode can 
reach more than 10 Gohms.
The  Johnson thermal noise from a resistor goes up like R^1/2
The useful signal  from the photocurrent goes up like R.
So, the S/N improves like the square  root of R !!!
With the 1/2piRC formula, even the smallest photodiodes  capacitance should 
give a no usable bandwidth.
But the experiments do not  confirm this point.
The sensitivity in the 20H to 3 KHz BDW seems ultimate  and the low rates 
pulses response stays very good...
Regard  
Yves