[SOLVED] Noise in photodiode amplifier circuit

Hello,

I've got a problem with noise in my photodiode transimpedance amplifier circuit. I've looked around on this forum and on the web but couldn't find something useful so here I go...

I need to detect a moving ball passing through a collimated IR laser beam. My photodiode circuit should detect the reflection of the ball. The reflected pulse width varies between 5 and 50us and is of low power (2mW laser reflected 1 meter away...). My circuit is the following :

**broken link removed**

The only difference is I'm reverse-biasing my photodiode (Vishay bpv22nf) with -15V to reduce diode capacitance.

Rf is 820kΩ, Cf is approx 0.5pF, Cdiode is approx 20pF. I'm using a OPA606 as transimpedance amplifier on a dual -15V/15V supply. I'm also using bypass capacitors where needed. Everything is mounted on a standard breadboard.

Now to the problem : I'm getting some significant noise with V_noise_rms ≈ 8mV and frequency of 100KHz when ambient light shines on the photodiode (indoor, normal conditions...). This noise goes away if I cover the photodiode. And I can't filter it because it is in the frequency range in need to detect...

The problem does not come from the amplifier oscillating : the feedback capacitance Cf is high enough. Could Johnson noise in my photodiode / feedback resistor create such a high noise voltage at such a frequency ? Do you have some ideas ?

I would be very grateful if somebody that worked on a similar problem could shine some light in my direction :wink:

Mike

PS : I'll attach some scope screenshots tomorrow !

I don't know how you are getting 100MHz out of the OPA606! What is happening to the DC level out of the opamp?

Keith

Yep sorry I meant 100Khz The DC level varies with ambient light. There is a 50Hz component due to the light, but I can filter that out wihout problems. I don't have high frequency noise (I've got a 150Mhz scope, doesn't show any other noise apart from this ~100Khz noise)

Mike

EDIT : Just to make things clear, the circuit works, I can detect the reflected light, but if the reflection is far away and short in time it can get lost in the noise.

Johnson noise would be broadband and random-looking. It would not look anything like a periodic 100kHz signal. The fact that it comes and goes when you cover the photodiode suggests that it is present in the ambient light. But that is highly unlikely, unless you have some pretty strange ambient light sources. My guess is that when the photodiode is uncovered, it shifts the DC operating point and enables an oscillation that is suppressed when the photodiode is covered. How periodic does this noise appear? Have you checked the bias supply? Could the 100kHz be coming from that? What if you temporarily replaced your bias supply with a 9-volt battery? What if you temporarily increase the feedback capacitance? Sure, that would kill your bandwidth, but it would interesting to see how much more feedback capacitance you need to kill the "noise".

It is not really periodic, more like random pulses that are around 10us in width... I checked all my power supply lines, there no such noise in there. I couldn't check the current coming from my photodiode because it is really small.

Tomorrow I'll try to replace the OPA606 with some other op-amp to see if it changes anything, and play with other Rf values to see if the noise scales with it....

Mike

What is your ambient light source? You can get 50-100kHz in fluorescent lights.

Keith

Yes there are fluorescent tubes in the ceiling ! That could explain the noise I'm seeing !

First thing tomorrow I'll do the test.

Mike

Hi,

Sorry for the late reply. I found the problem ! Thanks Keith for pointing me in the right direction : it wasn't the fluorescent lights (almost no influence) but the computer and oscilloscope monitors ! They generate a rather large ~100Khz signal in my circuit ! If I turn the screen off and hide the oscilloscope everything works fine

Mike

Be careful of LED torches (flashlights) as well!

Keith