help needed in IR demodulation circuit

OK, I'll bite.

When you say the two IR sources have a different frequency, do you mean they are modulated with a different frequency, or that their light output is a different frequency (usually given as wavelength)?

Will both be transmitting simultaneously? If they are, then you may see a far more complex input signal.

 

The problem is relatively simple to solve if only one device is transmitting.

You simply look for a series of pulses with a specfic period.

so for a signal like this:

____-____-____-____-____ (20% duty cycle)

you would look at the period (___-) and see if this repeats a number of times (you want to ignore noise) and then you can say that the signal is present. Essentially you track the time between positive (or negative) edges.

Note that this allows your duty cycle to differ without affecting the detection. So this signal would also be detected:

__---__---__---__--- (60% duty cycle)

These may appear to be different lengths because your browser bat show _ and - as a different length. Cut and paste them into something with a fixed pitch font (notepad?) and it will become clearer.

The duty cycle can sometimes seem to change as you move the transmitter and receiver closer or further away, so eliminating this as a part of how you detect the signal is a good thing.

Imagine we have a second signal. It has a slightly shorter period

___-___-___-___-

If we can detect that duty cycle as another signal, then we can trigger some other output when it is received.

There are some things to resolve before going any further. The transmitter frequency (or the receiver clock) may drift over time, temperature, and/or battery voltage. Thus the duty cycle may seem to drift. Your code should look for a sensible range of duty cycles, not a specific single one.

Now the next problem.

imagine they are transmitting together

____-____-____-____-____ (signal A)
___-___-___-___- ___-___ (signal B)

the receiver will see this:

___--__-_-_-__--___-___--_

This is far more difficult to decode.

Before we discuss that, do you understand all of the above?

 

Sorry, in that paragraph, I should have said FREQUENCY, but period is a suitable alternative. Well spotted

Here is where I lie and say I put in that deliberate mistake to see if you understood

 

The next part involves you explaining more stuff.

1) are you making the transmitter(s) as well as the receiver?

2) is range an issue (if so, what range do you require?)?

3) what research into detector circuits have you done?

4) Is there a purpose for this? Just for fun, or something more serious?

 

OK, here is where you need to explain things clearly because I can't see why you need a single detector. 2 detectors and 2 sources seems like the usual way. And then you don't need anything really fancy to decode the signals.

If you only have 1 detector then it would seem logical that the beams are not parallel, and therefore the speed will be quite hard to calculate.

 

No it's not clear at all.

The issue I see is that any decoding of signals will slow the response time. If you have to deal with multiple interacting signals then that delay could be variable. Both of these point to sources of potentially significant errors in your timing.

I would only be using a single sensor if there were absolutely no other choice. It complicates things horribly.

Why don't you start by explaining the actual problem you're trying to solve, then how you intend to solve it, and then how this dual receiver fits in.

These facts are important or you may end up with a perfectly working but totally useless device.

A simple statement of what you were doing at the beginning could have been

You have avoided answering the questions I have asked about WHY you're doing this, so I'm going to presume it's a school project.

 

Sorry about that. I was dealing back and forth with your thread and another one. It wasn't you that failed to answer

I'm going to need to know as much as possible. I presume that latency will be an issue? Will the transmitters / receiver be positioned in a fixed location?

Which beams will be broken? In what order? three beams? so which ones will be used for the timing? How long will the beams be blocked for? Where is this being done (in a lab, outside)?

My thinking at this point is that yo have the 2 IR beams going to the same receiver sending different length pulses (at a slightly different frequency). As long as you detect pulses of the correct length you know they're not blocked. When the pulses are missing for some period (at least 3 or 4 times the delay between pulses) then you can be certain the beam is blocked. As soon as you've seen several within a small time frame you can assume the beam is unblocked again. This goes against everything I said above, but will probably allow you to have a smaller and potentially close to fixed latency. It will be more difficult to set up, but I'm assuming it's in a lab environment.