Pulsed drive of white LEDs?

Use a pulsed drive, for sure, but also an IR filtered photodiode. Otherwise you get blown out of the water by the shot noise of detected daylight and the 40 kHz-1 MHz junk from electronic ballast fluorescents. It's mildly inconvenient not to be able to see what you're doing, but a webcam can see 850 nm just fine.

Or do it in the visible, if you don't need much range. It'll be cooler looking, for sure, but it'll crap out much sooner with distance.

Cheers

Phil Hobbs

Cheers

Phil Hobbs

I didn't get your point at first, but I do now. You correctly pointed out that it's only the line of sight between source and receiver that matters - well, up to a point. Some care will still be needed to limit the angle of detection for two reasons. One is that reflections could cause problems. The other reason is that there will be two virtual lines in parallel working independently, only 7 inches apart over a distance of several feet. Cross detection is unacceptable.

I suppose I could pulse the two beams at different frequencies and filter the signals at the receivers, but that adds complexity and could result in unforeseen problems..

Correct. It's why I rejected the laser option, at least for the upcoming race. The only sensor with an area large enough to ease alignment would be an LDR, but it's too slow. Anyway, I doubt if I can get one an inch wide or larger in a short time.

That's an idea. I'd intended to place the four emitters as unidirectional transmitters between the two racers, battery powered without any electrical connection to the receiving units. Running a cable to the middle of the track from the side, even behind the starting line, might be a problem. But it's worth thinking about.

I'm busy on a project right now, so I can't try it myself right now, but I'm contemplating how to view the photocell-controlled street light across the street from me and use it to control my yard lights.

(I'm down in a notch, between three hills, so ambient light is a poor controller. But the street light is about thirty feet up with a photo sensor with a better sky-view than I have ;-) ...Jim Thompson

One thing I'm pretty sure of is that there won't be any fluorescent ballasts spewing EMI. The venue is an abandoned airfield and the starting line will be quite some distance from the old terminal building. There won't be any electrical supply at that point even if there's one in the building, unless we string a long power cable. This is why I decided to power everything with batteries, including the christmas tree lights.

Thanks for the webcam idea. I still haven't completetely abandoned the IR route yet.

I see. Sunlight will be the issue then--it really is very bright, even with Jim T's idea of a feedback loop to get rid of the DC photocurrent. That's much easier to talk about than to get right, because of the noise issue.

The basic problem is that your detected electrical power will go as

You have to design the feedback loop carefully too, in order to prevent its noise from dominating.

Cheers

Phil Hobbs

I made a trial batch of street light controllers for my state's electricity board nearly 20 years ago. Control was from distribution points rather than at each lamp post. The problem was that practically the whole state, including the capital, is made up of small jagged hills. Daylight varies from point to point and the controllers had to be set individually. They worked fine when set properly, but the engineers left the setup to unskilled workers who had no idea what they were doing. That turned the experiment into a failure.

Would not red laser pointers work.

Now when you have a fairly intense (street) light with AC modulation on it.

I'm planning on a collimating tube as well, and directly behind the street light is hill, so I don't have to confront a direct AZ sun ;-)

...Jim Thompson

This could be an important factor. The persistence of the phosphor(s) used could seriously affect the waveform of the optical output.

I haven't seen much data on this, as I suspect that most of the engineering done for white LEDs is aimed at their use for illumination, where persistence isn't a factor. But it shouldn't be too difficult to measure. Drive a white LED with a square wave and observe its output using a phototransistor and a 'scope.

Two common ways to sense IR in the presence of ambient light are the TV remote receivers (three-terminal ICs in the receiver include some ~40 kHz filter and preamp so only the remote signal gets through), and receivers based on the LM567 tone decoder (which is specified for up to 500 kHz operation with logic-level detector output).

You can limit the view with a simple shield. I used "Omnidirectional" just to try to get the point across.

Even simpler, shut down one transmitter when the opposite receiver is in use. Alternate at some sufficiently high rate.

A laser pointer would be simple to point accurately (kinda the purpose of the thing ;). Get the interest of all the neighbor cats too. ;-)

Limiting factor is the die, not the phosphor. I've seen 200 nanosecond rise times on white leds with standard yellow phosphors.

Steve

Ambient rejection servo. I was looking for LTs app note on this, but Hickman must do:

Steve

Fall times too?

Cheers

Phil Hobbs

Luxeon claims 200 nS up and 200 nS down for half watt dies.

I only measured up time for a strobe application, but the Nichia and Luxeon parts drive very nicely off 50 ohm pulse generators with a

Steve

To avoid cross detection, you can do this:

emitter1------->detector1

detector2

Thanks for the info - now I know phosphor persistence is pretty small, as I suspected.

This data point appears to me to indicate that phosphor persistence counts for something, since unphosphored LEDs usually have rated rise/fall times of 30 nanoseconds.

- Don Klipstein ( snipped-for-privacy@misty.com)

Emergency vehicles already signal at intersections. The receivers sit high on the wires or poles across an intersection and are IR-based. They use optical baffling to narrow the acceptance angle downward towards oncoming traffic at the right spot and optical wavelength filtering to narrow the accepted IR color bandwidth, as well. They use electronic receivers that are tuned with extremely narrow notches so that only a very precise flicker rate of the emitter is accepted. The sun doesn't yield much flicker in that band (circa 15Hz, if I recall.)

I haven't been reading the thread and I apologize if I've duplicated something already said. My thought though is for you to use a tube, baffling and possibly some optics, a narrow-band wavelength filter (thin film?), and a receiver/emitter pair that transmit/receive in an extremely narrow bandpass. The combination appears to work, elsewhere. I believe I've also seen similar things designed to count traffic, replacing those hoses that lay across the road. Maybe there's no time to screw with this stuff, though.

Jon