Blinded by the light

Forum 2005-2010 (read only) Hardware Interfacing
1

Morning, all. I'm trying to roll my own electric meter reader, a la the Blue Line PowerCost Monitor (or the re-branded Black & Decker one). My meter flashes an IR LED for every one watt-hour of usage. I've succeeded in reading, logging and graphing the power consumption rate using an op-amp and a SFH205F photodiode. My problem seems to occur at around ten to 9 in the morning, when the sun peeks around the house and blinds the photodiode. I'm trying to find a solution to this problem.

This photodiode purports to have a daylight filter (and is indeed very dark plastic), but that filter doesn't seem to work very well. It also has a 60 degree viewing angle which I'm thinking may be a problem. What I believe to be happening is that the photodiode is getting saturated by the sunlight. This seems strange to me because although I can reproduce the effect on my workbench using a 60 watt lightbulb in very close proximity to the sensor, just pointing it at the sky actually seems to blind it even more! My first pass at the circuit was just the photodiode connected between ground and the inverting input of the op-amp in a current-to-voltage configuration. I've since added a 1uF capacitor in series with the photodiode to remove the DC component of the signal, so that I'm just amplifying the pulses. I'm able to confirm with an oscilloscope that I get a nice, clean, 5v, 10ms-wide pulse going to the Arduino, and 0v the rest of the time. I can see the width of the pulse decrease as I bring my 60W bulb close to the photodiode, until it eventually disappears as the bulb overwhelms the signal from my test IR LED.

Out at the meter, I've got a very ghetto arrangement with the photodiode duct-taped to the meter housing with about 3-4 inches of tape on either side of the sensor. I probably could just wrap the circumference of the entire glass dome of the meter with tape to try to get rid of any more light coming in (the side of the dome faces more-or-less due South), but this seems like I'm solving the wrong problem. The Blue Line device does shade the dome a bit, but much less than I'm doing.

So, I'm running out of ideas. From a signal-conditioning standpoint, I feel like I've reached the limit of the sensor. I might be able to isolate the pulse from the noise a little bit more using a signal diode, but I feel like the root cause of the problem is that the noise is overwhelming the signal. Rather than putting the photodiode right on the face of the meter, I'm thinking that I can try to raise it a half-inch and shield it (think hole drilled in wood with the photodiode recessed), which might prevent incident light from coming into play. Another possibility is replacing the photodiode with one with a narrower field of view. Have I missed anything obvious from a sensor and circuitry standpoint?

Thanks,
Brian

2

I feel like the root cause of the problem is that the noise is overwhelming the signal.

Yes basically you have.

just pointing it at the sky actually seems to blind it even more

It will do, the sky is very bright in IR much brighter than the bulb.
The problem with filters is that you also block out the light you are trying to receive. You could modulate the IR but then if the sensor is being saturated it is not going to get you anywhere.

I think your only hope is to peruse the mechanical shielding arrangements even more. If it only goes wrong at a certain time can you not arrange a curtain or something?

3

Here in the US, our electric meters are on the outside of the house. I could set up a curtain, I suppose, but that would look weird to the neighbors. ;D

Ok, on to the physical solution, then. I'm quite impressed with what I'm getting out of my signal filtering, so far; too bad it's not enough!

4

It sounds like you are on the right track at least.

My suggestions

  1. Shield it better
    or
  2. Go ahead and find a photodiode with a much smaller angle
    or
  3. If the photodiode is not actually being totally saturated then how about using two photo diodes? One that is much more directly in front of the IR LED and one next to it. Then they both would get approximately the same level of ambient IR but the one would get much more IR from the LED. So you could use your opamp to get the differential output. Depending on the noise level and the degree of saturation this might work fine.
5

Good ideas. I'll start with the shielding, and maybe try throwing another photodiode into the mix. Fortunately, I never buy just one of anything. :slight_smile: I like the idea of finding another photodiode with a smaller FOV, too.