Marc Popek said:
Yannick,
If you want to take red light and say red light shfited by 420Mc and then
have both light sources fall onto an APD and then you like to see how much
430 MC you can extract for varying:
light levels (re and red shifted by 420mc) and the bias and the apd devcie..
no can do in spice....
Your best bet for an answer is to use a pencil and paper and come close to
the answer, assume an exponential transfer function and exime the mixing
prodcut's ampltiudes.....
I think hes using the technique I sugested for hetrodyne LIDAR some time
ago,
The light is modulated with one frequency and the bias to the APD is
modulated with another frequency,
the resulting electrical output of the APD is the product of the two
frequeuncies,
the mixing being done at the opto-electronic level not the optical or
electronic level,
the advantage is that the high frequency light modulation frequency is not
converted directly to an electronic signal therefore the APD capacitance
does not have any effect on this wich would severly limit the acheivable SNR
for low light, only on the beat frequency wich as it is much lower frquency
is much less affected by it.
It relies on the multiplication effect of the bias voltage, the gain of the
device rises with voltage very sharply at voltages just below vbr.
In a simple demonstration of this I used gain=1+(v/250)^k with k 20~200 wich
gives an arbitraraly sharp curve just over 250v.
I dont know how steep the curve is as it isnt ploted on the data sheet very
far but at some point just above vbr it tends to infinity as it latches on.
There seems to be no information on the frequency response of the gain to
the bias voltage, although the gain bandwidth product is given but this is
the response to the light signal, but my gues they are closely related, but
I had it working some time ago at 10mhz and just recently at 100mhz, im
looking to get try it up to 2ghz.
Colin =^.^=
