Photocurrent amplification ?

[acp32]

I desperately require some circuitry advice before I go mad!
I have built some interface circuits for silicon photodiodes (quadrant detector) which seem to work on a knife edge.
I have tried different op amps with low input bias currents, all have not seemed immuned from problems. In the latest circuit I'm trying to debug I'm using the AD711 op amp.
The first stage is made up of the photodiode across the inputs to the opamp with a 1M feedback resistor to convert nA to mV. This seems to work fine the voltage goes up and down with light intensity. However, I need to amplify in a second stage to get the mVs signal up to Vs. I do this with a bog standard inverting amplifier. However, the second stage does not seem to deliver the gain. If I lower the gain to 2 we start seeing the expected voltage. I have similar non-ideal behaviour when I tried a non-inverting amplifier. Note that when I replace the first stage with a small voltage source the second gain stage works fine.
Does anyone know what op amp limitations are restricting the gain functionality in the second stage?

 

[IRstuff]

You don't say what impedance you're trying to drive with the AD711, which is the most obvious potential problem. The AD711 is spec'd for a max of 25mA output current.

TTFN

 

[nbucska]

mV to V is 1000 times. Offset ( Vtg/current) may overdrive
your 2nd stage !!

Is the output close to rail ?


<nbucska@pcperipherals.com>

 

[acp32]

Thanks for comments. Here is a bit more detail.
The second stage provides a gain of -160, with R1=4k7 & R2=750k. During testing I didn't have the second stage driving anything.
I have made some modifications which have helped the circuit. Firstly, the photodiode was not across the OPAs inputs. The cathode of the photodiode was at 5V, with anode connected to -ve input. While the +ve input was at ground.
This was to have the photodiode more heavily reverse biased.
Secondly the detector is on the end of a lead, the length of which seems to make a difference to whether the circuit works.
If anyone can explain the reasons behind these observations I would be very grateful.

 

[IRstuff]

oscillation, perhaps? Can you stick an oscilloscope on it?

TTFN