We have a device that counts dust particles in a sample of air. It works by shining a laser through the airstream, bouncing off a reflector and into a photodiode. When a piece of dust interrupts the laser beam, the photodiode voltage spikes and a comparator detects it and registers a tick for the controller to count.
We noticed something odd on the oscilloscope today, that when a particularly big particle goes by, causing a large voltage spike (>800mV or so), when the voltage comes back down it will dip below the baseline voltage, shoot up again, and then settle. It almost looks like a control system response but there's nothing controlling the photodiode response apart from a fixed-gain amplifier.
The "rebound" oscillation gets larger as the initial voltage spike gets larger, large enough that the comparator will sometimes count the rebound voltage as another particle. We were able to smooth that out with software controls, but I'd like to understand what is causing this behavior to begin with. I unfortunately don't have the specific model of photodiode handy, but I'm not an electrical engineer and I was curious if this sort of behavior is typical of photodiodes.
What's the time scale? Micro or nanoseconds? Sometimes "ringing" or overshoot, etc., can be caused by the capacitance/resonance of the 'scope probe so it may not be "real" or it may only exist when the probe is connected.
Do you mean full scale 2.5µs, or do you mean 2.5µs/division? Be clear with all measurements like this, any ambiguity will cause misunderstanding.
What is the dark baseline for the photodiode? Presumably there is some leakage from the laser beam into the photodiode with no dust particles around. Anything between the diode and laser beam could cause a dip in that light leakage.
You need to show the full circuit, the datasheet link for the components involved (photodiode and laser, "fixed gain amplifier") and the 'scope probing set-up (x1 or x10 or 50 ohm ?)
Basically all the details matter, there are many possible causes for something like this, electronic or otherwise, some of which will be problems that need fixing (like lack of decoupling for instance).
Have you characterized the photodiode response to an LED or laser pulse?
