I made an object counter with this at the front end:
The IR parts are a Radio Shack pair, with no datasheets and the typical minimal information on ratings on the package.
I was driving the emitter at 100 mA, which is 2/3 of the rated "maximum continuous current" of 150 mA.
I ran it continuously for three months--about 2200 hours. I expected V.out to remain at the same value (when not blocked), but V.out increased from 2.2 volts at startup to 4.7 volts at the end of the period:
The current through the detector at the end of period is 68% of the current at the start.
This is apparently due to a decrease in IR emitter output. A new emitter pushes V.out back down to 2.2 v or lower.
Later, I found this article about IR emitter aging: http://www.vishay.com/docs/81117/81117.pdf, which shows that the output of the worst-performing IR emitter material degrades to 85% of original power after 4000 hours.
So, the question is:
Is my change in V.out over only 2200 hours consistent with normal IR aging (doesn't seem like it to me), or is something else going on?
PS: FWIW, I'm using 10 v because an adjacent component (not shown) needs it, and I have current to spare. Also, this sensor is at a remote location (75 ft from data logging equipment) and there are no "smart" components (e.g., microcontrollers) at that location. Also, I'm also logging the time that the beam is broken (which can be as short as a few milliseconds) in addition to counting objects.
150mA is the maximum, in that the emitter won't immediately burn up. But semiconductor junction's lifespan is shortened considerably with heat.
The data there really stinks, I agree. If you can get this to work with more like 20mA in the emitter, do so.
Why is the 1k resistor there? Without it, Vout should be around half a volt with the path not blocked. What are you doing to scale the signal to 5V when the beam is blocked?
Yes, it's that Radio Shack part number. But the specs vary depending on the package. The new one I bought says 100 mA max continuous, whereas the old one said 150 max.
Why 100 mA? Well, it seemed like a safe fraction of the maximum and I thought "more is better." I am running the new emitter with about 50 mA and it works fine. I'll see if even lower will work.
The 1k resistor is there because without it, I was getting negative V.out when the emitter was new and powerful. I didn't want to feed that into the op amp immediately downstream.
V.out feeds a crude op amp-based driver that feeds about 75 feet of twisted pair that goes to a resistor divider that takes the voltage down to TTL levels and my data logger.
If you get a "negative Vout" from that circuit under any conditions, there is a serious problem with your ground reference. Make sure all grounds are connected together with reasonably heavy wire, and if that doesn't work, use an optocoupler to isolate the two circuits completely.
A ground loop can happen if two power supplies are on different branches of an AC circuit, for example, and it can destroy equipment very quickly!
Hmmmm.... I got the neg V.out with the circuit at the top "on the bench" and nothing else attached. The two sides were tied together to ground in close proximity to each other. I thought it was very odd, but recalled reading that an isolated illuminated photodiode (I drew a phototransistor, but I just checked the new package - it shows a photodiode) generates a voltage, so wasn't too concerned. The 1k solved it, so I moved on!
