12

I have setup a BPW-21 photodiode as shown below:

enter image description here

The photo-diode is activated by an oscillating laser beam. I expected to get a clean +5V to 0V transition at point A when the laser beam falls on the photodiode and a 0V to +5V transition when the laser moves off the photodiode. However what I actually get on the oscilloscope are multiple 0V to +5V transitions lasting for few hundred microseconds before settling to the expected voltages. Some example traces are below:

enter image description here
enter image description here

My question: Why is the voltage at point-A "bouncing"? What is happening in the photo-diode to make the voltage bounce between + to +5V before settling on the expected value? Any ideas

Abhishek

George
  • 1,396
  • 4
  • 16
  • 33
Abhishek
  • 409
  • 5
  • 11
  • 2
    Try blocking the laser beam with an opaque object and check the plot. *Moving the laser off the photodiode* sounds like it can involve all sorts of vibrations. – Dmitry Grigoryev Mar 30 '17 at 10:22

4 Answers4

13

Semiconductor laser effect is described by two coupled partial differential equations of carrier density and photon density, the rate equations.

The solution of these equations result in a non linear current-intensity relationship causing relaxation oscillation when the diode is turned on.

See here or following image: laser diode relaxation oscillation
(image source: p. 45 of this document)

And what you see is exactly this oscillation near the rising edge of the signal.

Curd
  • 16,043
  • 34
  • 43
  • Good add. Though I could not tell if "oscillating laser beam" meant physical or electrical. – Trevor_G Mar 29 '17 at 18:39
  • I meant physical ... actually the laser beam is reflected off a mirror mounted on a torsion pendulum – Abhishek Mar 29 '17 at 19:16
  • @Abhishek: ??? Isn't electrical also physical? BTW By "oscillating laser beam" I understand that its intensity is modulated electrically (and it doesn't matter if it is reflected off a mirror or not). If it is operated continuously relaxation oscillation can not be an explanation of what you see on the scope. – Curd Mar 29 '17 at 19:41
  • @Abhishek: maybe your laser beam is also electrically modulated? – Curd Mar 29 '17 at 20:16
11

It could be motion. Probably at the laser source, but it could also be at the photodiode end, even a fan somewhere can cause sensitivity.

However, if you have no aperture, or the wrong aperture, it is also possible to get stray laser paths to the sensor as the laser crosses the metal shroud of the sensor or reflects around inside the arrangement.

If you have any other kind of optical window lasers can bounce around inside those too.

enter image description here

Over sensitivity of the detector circuit will also cause you grief. For best results you want the detector circuit to give you somewhere in the 80-90% swing when fully exposed, not swamped. This will give you sufficient tolerance for it to work over a variety of devices and power conditions while still giving you sufficient signal range to use appropriate hysteresis.

GENERAL COMMENTS:

Often people think they need to use pinpoint lasers for position detection because they think lasers are wonderful. The truth is, unless you want to position something at a distance, to <1mm accuracy, using a laser can actually cause you more grief than using a less columnar light source.

With lasers it is important to align both ends. With a simple light source and appropriately appertured receiver, you only need to position the receiver accurately.

Lasers tend to ricochet. There are occasions when the laser can actually bounce around the object you are tying to measure and still end up on the sensor. Worse they can actually bounce around within your sensor.

enter image description here

If the laser and receiver are meters apart you can have thermal issues. Relative motion between them due to thermal expansion of whatever they are attached to can cause the laser to miss the target entirely. In fact keeping both ends mechanically coupled is an issue in general.

In many occasions I have found it prudent to actually defocus the laser so it arrives as a quarter sized spot at the receiver end. The aperture on the detector was accurate enough for the task at hand, but the alignment and vibration issues went away.

Trevor_G
  • 46,364
  • 8
  • 68
  • 151
  • Dear Trevor ... to speak the truth, this is also what I suspect...in the last paragraph you have mentioned "aperture of the detector was accurate enough", do you mean that you put a separate aperture (like the one here http://www.e-holmarc.com/product/variable-round-aperture-ssid-mssid-series-25/) or do you mean that the diameter of the window on the photo-diode package was enough for the job at hand? – Abhishek Mar 29 '17 at 19:13
  • @Abhishek, an aperture can be a slit or a small hole, smaller than the sensor, fixed at some distance in front of the sensor itself such that light coming at the sensor will only hit the diode when the hole in the aperture lines up with the light source. It does not need to be a fancy device like that. A simple 1mm or1/2mm drill hole in a thin plate is sufficient. The further in front of the sensor the more accurate it is, but you trade off a little sensitivity with distance. – Trevor_G Mar 29 '17 at 19:35
  • In the tens of microseconds between those pulses, light would travel for kilometers. The mechanism you propose would explain bouncing at a different, much faster, timescale. (unless I grossly miscalculated something - or the OP has very thick windows :-) ) – Sredni Vashtar Mar 29 '17 at 21:18
  • 1
    @Sredni Vashtar no you are missing the point. As the laser physically moves onto the sensor the extra paths can cause modulation in the laser brightness due to such fringing type effects. The speed of the sensor modulation is dictated by the physical speed of the source / modulator. – Trevor_G Mar 29 '17 at 21:21
  • @Trevor, oh, that's interesting. So basically, that would be the 'beat frequency' - much smaller than the 'carrier'? – Sredni Vashtar Mar 29 '17 at 21:25
  • @SredniVashtar you could think of it that way as you come into alignment yes. – Trevor_G Mar 29 '17 at 21:27
-1

There is mechanical movement. That likely will have vibrations.

The photodiode output ---- what you have shown ---- is rail-to-rail because the Laser is so intense. Place a filter between the Laser and PD, and get a better view of the arriving energy.

analogsystemsrf
  • 33,703
  • 2
  • 18
  • 46
  • The moving laser is mounted on an independent suspension (i.e. suspended by a string) and oscillating with a time period of about 10 seconds so mechanical vibrations can't be the problem. The photodiode itself is not moving .. and even-if its rail to rail why is it "bouncing"? .. there are no mechanical contacts to vibrate – Abhishek Mar 29 '17 at 16:25
  • Vertical vibration? Brownian motion? Or I suspect ringing in the TIA interface analog circuit. Again.............insert a 1-stop or 2-stop optical attenuator, and look for changes. – analogsystemsrf Mar 29 '17 at 16:33
  • TIA? what is that? Point A is connected to one of the inputs of LM339 ... and what is a 1-stop or 2-stop optical attenuator?...um excuse me for my ignorance :-) – Abhishek Mar 29 '17 at 16:36
  • Get a piece of dark glass, and hold in front of the PhotoDiode. Also, show the complete LM339 circuit; include VDD connection, GND connection, local VDD capacitor, any load on LM339 output. TIA is trans_impedance_amplifier, a popular interface for a photodiode. – analogsystemsrf Mar 29 '17 at 16:47
  • 1
    It looks like interference patterns changing with angle of deflection from the edge. How well defined is the aperture for edge deflections? – Tony Stewart EE75 Mar 29 '17 at 16:48
  • okay, but why should it affect the response?..other than changing the incident laser beam intensity – Abhishek Mar 29 '17 at 16:48
  • 1
    This *IS* laser light, with scintillation! – analogsystemsrf Mar 29 '17 at 16:50
  • Dear Tony ...the laser beam is a conventional pointer, maybe 2mm in dia and the BPW-21 comes in a TO-39 package, looks like there is no aperature which can cause diffractions or interferences..however can you please elaborate what you have in mind? If the laser beam is not falling perpendicularly on the photodiode, there may be reflections/interferences cause, but it should be the same for each cycle, its different – Abhishek Mar 29 '17 at 16:52
  • analogsystemsrf ... um isn't ionising radiation a necessity for scintillation? – Abhishek Mar 29 '17 at 16:56
  • make a smaller aperture over the sensor and see how it changes. The fringe patterns will change with slightest different paths each time. Then try a two stage aperture. The farther away from sensor the less aperture interference. Look at any SLR camera and see how the black stepped cone is designed to block light around the edge. of the lens in front of aperture shutter – Tony Stewart EE75 Mar 29 '17 at 16:56
  • analogsystemsrf and Tony ... thanks I will try them out – Abhishek Mar 29 '17 at 16:58
  • Just a side note.... Apertures are important for blocking reflections. I recall making a 5mm IR LED could be blocked with a 1mm moving wire detector across a 1m path. Both sides were recessed to get this 1mm path resolution and a black heatshrink was used around emitter. The path used PCM with builtin AGC for each path with multiple emitter detector pairs and a unique PCM code for each path. Moving a resistor wire at any hand speed across middle of a 1m path blocking the IR beam was easily detected. The deep hole recess made the aperture beam width narrow yet alignment not that critical. – Tony Stewart EE75 Mar 29 '17 at 17:30
  • 1
    @Abhishek I suspect he's using "scintillation" as a layperson's term for [speckle pattern](https://en.wikipedia.org/wiki/Speckle_pattern) – Random832 Mar 30 '17 at 05:01
-1

I suspect that the laser beam might be smaller than the area of the PD. If this is the case, then as the beam moves across the area, some parts of it can conduct and then not conduct causing the apparent "bouncing" until enough of the PD is activated to turn the diode on. On its way out, the effect is repeated until the beam is off all the areas of the PD. This can be verified by holding the beam steady and using something to interrupt its path, instead of moving it across the PD.

Guill
  • 2,430
  • 10
  • 6