[sdiy] photodiode sfh213

[Czech Martin]

hi Bert,
don't worry, @ 12MHz most op amps won't work anyway ;->.
AFAIK the photo diode delivers current dependent on light
intensity. I gues you want to operate it in forward direction.
So you need a current amplifier.
This calls for a npn transistor base (transistor in common emitter circuit).
This will convert base current variation into collector voltage variation.
Because of Miller (d)effect this could be difficult @ 12 MHz.
So maybe a simple common emitter circuit won't do, you need a cascode.
All resistors must be as low as possible, and you should run the
thing with considerable bias current (otherwise parasitic capacitance
will kill your bandwidth).

I do not know what you want exactly, so a differential input stage might
be necessary to suppress input stage variation due to temperature,
voltage etc. Of course you don't need that for a narrow band application.

So what about a 2 npn cascode  amplifier, followed by an emitter follower
(capacitive coupling)?

How much amplification do you need? Perhaps one stage won't do for
more then 10x or so.

If your signal is narrow band, you may want to make the circuit sensitive
to only this frequency to kill noise. Then the broadband cascode is wrong.
You better use a tuned common emmitter stage, followed by a capacitively
or transformer coupled emitter/source follower for low impedance output.

This means that the emiter has some resistance to ground (degenerated emitter
feedback). This will later settle the amplification.
The collector resistor is replaced by an LC paralellel resonant circuit.
This looks strange, because the collctor is shorted to Vdd (for DC).
It can not move. It can only move if the LC tank will get high
impedance, ie. when you hit the resonant frequency.

A practical value would be 1nF, 175.9nH. The inductance is so low that you
do not need any core, air coil! So you solder your tranny, take some insulated
copper wire (start with 10 turns) and biuld the circuit. Next you stimulate
it with your  lab oscillator and look for the resonance. Too low?
Add capacitance. Keep in mind that your oscilloscope will add 10pF or so.

(I'm actually building such a circuit, but for 77.5kHz, which calls for
coils with feritte core, but this does not change the principle).

You may need to adapt the bandwidth via parallel resistor.
You may also want to tune the circuit via variable capacitor (trimmer).
The emitter resistor has to be chosen for the desired amplification.
Too much amplification, and you have built an oscillator!
Base biasing and emitter resistor make the DC bias current.
Let Ic be 1mA. This will also set a limit for supply voltage.
arround 10V (power).

Now you have amplification in the first stage. The problem is that
it has high impedance. You can use a FET to decouple it. JFET source
follower, with source resistor. Capacitor from LC tank to gate.
drain to Vdd. The source resistor depends on threshold voltage.
You need to experiment, because this differs a lot for different
Fets. Use versions with lower Uth (BF244A), cause the resistor should
be < 500Ohm.

That will make a decent narrow bandwidth amplifier.

This is written in haste, but I hope you got some idea about
the circuit and all the dependencies that will logically guide
you to the right component values once the "topology" is chosen.
I hate this t-word.


m.c.

-----Original Message-----
From: Bert Schiettecatte [mailto:bert.schiettecatte at esat.kuleuven.ac.be]
Sent: Mittwoch, 7. Mai 2003 17:13
To: synth-diy at dropmix.xs4all.nl
Subject: [sdiy] photodiode sfh213


hi all,

i have been experimenting with photodiodes lately and I'm 
trying to build a circuit that will give me a voltage output 
signal using the SFH213 photodiode. I don't want to use opamps 
or other expensive components for amplification, and the signal 
I'm looking for is a 12Mhz sine. I have no clue about what circuit
to build around this photodiode, what supply voltage to use, or 
what amplifier to build... I would appreciate all suggestions or 
pointers to resources!

bert