LED frequency detection

I'm part of a robotics competition in the spring, and I have a problem...

There is going to be a beacon suspended off the ground some inches and I need to build a detector to see it. The room the competition will be operating in will be flooded with ambient light, so we need to pay attention to the frequency that the white LED blinks at (something in the KHz region, not yet specified).

In our lab, we did have some success hooking up a photodiode to an o-scope and detecting the 120Hz of the the fluorescent tubes in the ceiling, as well as an LED we'd hooked up to a function generator (which was tuned to something in the KHz region).

My question is this though: Is there an IC (or other sensor) that can detect multiple frequencies (one at a time), and output something simple we can read with a dsPIC, such as some analog voltages, or some sort of serial connection?

You could use a PLL tone detector chip (NE567) to detect a specific frequency.

too bad you could not specify the frequency... there are nice IR remote ICs. I have used them for low speed com links ... they come in several frequencies around 40KHz

It's not that I can't specify the frequency, it's that it hasn't been specified by the people putting the competition together (yet).

Also, the blinking is done using white LEDs, so IR isn't really going to fly.

well that sucks. the simple way is to just do a high pass filter to block out line frequencies... it gets a little weird if there are tvs and monitors around

A better way is to take two readings , one with ambient light and one without ambient light , take the ADC Difference , then u get reading without ambient light .

ahmedragia21 said:

A better way is to take two readings , one with ambient light and one without ambient light , take the ADC Difference , then u get reading without ambient light .

Click to expand...

that does not work well since the background is constantly changing and you can not rely on them giving you a background reading in a contest.

it works much better to remove the DC background level and an AC carrier to discriminate background noise... it could even be used to allow communication with multiple recievers at once.

infrared

I have a problem here.
i want to detect infrared produce from fire. use photodiode or phototransistor? what type?
please give any referens.

white leds put out infrared so do red even just not as much as a infrared leds so you need a progarm that can pick up the blink of the led in khz I bet thay use 36khz to make it hard
on you but if you could think of a way to use CMOS active pixel sensor cell i bet it can pick up any binking led

gs3user007 said:

I'm part of a robotics competition in the spring, and I have a problem...

There is going to be a beacon suspended off the ground some inches and I need to build a detector to see it. The room the competition will be operating in will be flooded with ambient light, so we need to pay attention to the frequency that the white LED blinks at (something in the KHz region, not yet specified).

In our lab, we did have some success hooking up a photo-diode to an o-scope and detecting the 120Hz of the the fluorescent tubes in the ceiling, as well as an LED we'd hooked up to a function generator (which was tuned to something in the KHz region).

My question is this though: Is there an IC (or other sensor) that can detect multiple frequencies (one at a time), and output something simple we can read with a dsPIC, such as some analog voltages, or some sort of serial connection?

Click to expand...

That is a pretty challenging problem, but I guess that is one purpose of a competition isn't it?

Well I see two important requirements. First to be able to detect a 'modulated' light source at a specific frequency and second to be able to 'home in on it' to be able to steer your robot. I'm assuming below that you will never be told the specific frequency of the signal source prior to the contest, is that correct? If they don't provide that info then:

A. A tunable sharp voltage controlled audio bandpass filter. This tunable filter would have to be able to cover the possible range of the LED blinking rate. I would assume they will not be using a frequency in the 60-120hz range (normal room lighting would swamp the LED source) so your filter might have to range from say 1khz to 20khz, just a guess, we don't know the possible range right?. This could be made from ops amp circuits (a couple in series perhaps) and such circuits are common in audio equalizers and such. Search and study on active filters.

Your filter circuit would have to be calibrated such that you you know what frequency it's tuned to by the value of the DC voltage you output to it. Drive the tuning control voltage to the filter from a D/A from your micro-controller. The output of your tunable filter will need to be rectified and filtered such that if will provide a average DC voltage proportional to the amount of AC voltage at it's tuned frequency. You could start with sending a slow ramp voltage to the filter and monitor the output of the filter with a A/D input in your controller. When you see a sharp increase in the A/D input your program should then know the frequency of the modulated light signal. The filter would be driven by a AC (capacitor) coupled photo-transistor and possibly use a op amp to provide some voltage gain. The filter will show maximum output voltage when the frequency of the blinking white LED matches the tuned value of your bandpass filter.

B. Second task is your steering/scanning algorithm. I guess at first you only have to have the ability to have your robot slowly turn in a circle say 5-10 degrees at a time stop and do the filter ramp/detection step. Once a signal is detected your controller can just keep the filter set at that detected frequency and now you can steer toward the source reading the A/D signal as a colder/hotter signal source. You will have to experimate with the optics or mounting of your photo-transistors. You need some limit on it's degrees of sensitivity so as to give a good sharp signal a you turn toward and away from the source, but not so sharp that when you move a few degees and lose all the signal.

So that's my thoughts. If they could provide the specific frequency of the LED blink rate then the sensor electronics would me a lot simpler, just a fixed tuned active filter.

Good luck and keep us posted on your progress.

Lefty

i see.
but i mean,
can I detect fire with infrared detector?
how the circuit, if I use phototransistor?
thankyou

to_ said:

I have a problem here.
i want to detect infrared produce from fire. use photodiode or phototransistor? what type?
please give any referens.

Click to expand...

You really should start your own posting, asking your question rather then to break into this one. It might get better or at least some response