Quoting Harold Hallikainen <harold@hallikainen.org>:

The idea with using the transimpedance amplifier is to avoid voltage
changes across the PD capacitance.

Phil Hobbs has some excellent tips on high performance PD circuitry,
both in his superb book, and there's a quite good paper available
online if you just search for it.

Last one I did, I used an op-amp with a 1.5 or 1.8GHz GBW, if memory serves.


> I'm working on an application where we're sending frequency shift keyed
> data on an RF carrier on infrared. There are other carriers on the IR
> carrying FM audio. I'm demodulating the RF carrier using an SA639. I
> upconvert from my RF frequency to a 10.7MHz IF, use ceramic filters at
> 10.7MHz, then a ceramic quadrature detector. The trick is the interface
> between the photodiode and the SA639.
>
> Right now, I'm using a preamp that's a current to voltage converter based
> on an LT application note. It uses a JFET source follower in front of an
> op amp. This is then configured as a current to voltage converter with the
> photodiode reverse biased about 1V. The current to voltage converter
> drives an AC coupled op amp amplifier, then an LC network into the SA639.
> The LC provides some RF tuning and voltage gain. The op amps have a GBW
> product of about 100MHz.
>
> The performance of this circuit is adequate, but I'd like to do better and
> reduce parts count.
>
> Most photodiode interface circuits I've seen assume the diode is a current
> source and run it in to a current to voltage converter. It SEEMS, though,
> that we'd get maximum power out of the photodiode (allowing us to get over
> noise in the rest of the system) if we instead matched the photodiode
> impedance. Looking at the datasheet, I divided the open circuit voltage by
> the short photo current (with a specified reverse bias) to come up with
> the Norton resistance (the resistance across an ideal current source). I
> also used the rated capacity at the reverse voltage I'm running to come up
> with a model of the photodiode being a current source with 20pF across it
> and 400k across it. It SEEMS that we'd get the most power out of this if
> we matched this impedance.
>
> I've run some SPICE simulations based on this model and modeled the input
> of the SA639 as an 800 ohm resistor with 3.5pF across it. I then, in my
> model, use a "tapped capacitor tank circuit" to provide the reverse bias
> to the photodiode and impedance match the photodiode to the SA639. My
> experimental approach on this was to try adjusting the capacitor across
> the SA639 input, then adjust the other capacitor of the tapped capacitor
> pair until the peak in the voltage across the SA639 input was at the right
> frequency.
>
> We're about to test this against the previously discussed current to
> voltage converter approach. However, in all my reading about photodiodes,
> I have NEVER seen it suggested that we attempt to match the output
> resistance of the photodiode, which is what I'm trying to do here. So...
> anyone on the list have photodiode experience and would care to comment?
>
> THANKS!
>
> Harold
>
>
> --
> FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
> opportunities available!
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>



Best regards,
Spehro Pefhany
-- 
"it's the network..."                          "The Journey is the reward"
s...@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist