I fixed the capacitor issue. The circuit uses 2  10 uF electrolytic caps in series to create a 5 uF cap.  Should 5 uF ceramic caps work the same? They would be a lot neater.

I didn't get all the receivers to work, but probably won't bother debugging as the tsop 1138 worked pretty well from a wide range of angles.

I tried 5 IR receivers in the same circuit to see which worked best. Using a circuit like in the application notes except it uses a .1 microfared capacitor, the tsop2438 and tsop1138 receivers work fine. But the tsop34838, tsop38238 and tsop4838 don't seem to work at all.

The recommended cap is 4.7 microfareds. Could the smaller cap be the problem? Should all receivers work with the same circuit?

I also noticed that the tsop1138 gets inserted backwards to how i would expect. Could inserting the receiver backwards fry it?

The receivers that don't work have a larger operating voltage: 2.5-5.5V. The working receivers operating voltage is 4.5-5.5, They are being powered by a USB powered Ardweeny's gnd and + pins. Could the different range be an issue? I assumed the wider range was an advantage.

I was thinking of modulated IR as they are the easiest to work with. The IR receivers cost about $1. IR LEDs are like $.20.

Pololu has a beacon that uses 4 receivers and 4 LEDs. This makes sense as the $5 in parts is cheaper then a motor to rotate them.

I guess as the bot changes direction slightly the receivers indicate if the chaser is still in pursuit. I don't see how the IR alone can give distance, but maybe distance is not necessary to chase. Could also make the bot kind of pan back and forth when it is hunting for a signal.

I was also reading about the Rhomba IR invisible wall. It acts as an invisible wall to prevent the vacuum from falling down stairs. This could be just an LED in an opaque tube (or snout) to keep the light beam narrow.

This solution does not seem very satisfying, but I guess it could work.

Anyone try to make a bot that can chase another bot?

I was thinking one bot would transmit infrared (a beacon) and the other would receive it. IR remotes work well for 10-20'. While I've only got that too work from 6' using a receiver like a tsop2438, that is for another thread.

The question here is how to find the direction of the transmitter. IR receivers either respond or don't respond. I was hoping a photodiode or IR detector like http://www.sparkfun.com/commerce/product_info.php?products_id=241 could provide an analog value that would indicate signal strength. The sender would broadcast regularly. But when I tried it with a photodiode, it's maximum range was about 6 inches. Another thread on Arduino Forum indicates the Sparkfun IR detector also has a short range.

Any ideas for a beacon-detector pair that would work from 6-20'? Again, the  detector needs an analog signal that indicates 'hotter' or 'colder' as the receiver changes direction or gets closer.

I came across this old thread, but I had been thinking about same thing.

I don't think you can do it with just RF. In an open room, you can use a combination RF and ultrasound beacon. Send RF and ultrasound signals at the same time. The detector could figure the time lag and hence its distance from the beacon.

There is an MIT system called Cricket that takes this approach: http://nms.csail.mit.edu/papers/bodhi-thesis.pdf It's a good read. Cricket does more and the paper has background.

I played around with this and it seems better at night w/o sunlight coming in through the window and most lights off.

Overall there was about a 90 degree range that worked from 6 feet. This suggests the issue is with the receivers noise suppression.

I was thinking of trying a few receivers with different noise reduction. The retrievers have the same 3 connections, so they are easy to try. (The biggest hurdle is the cost of the shipping.) These are the receivers, I was thinking of trying: TSOP4838, TSOP34838, TSOP8238 and TSOP1138.

Also would like to try tsal7200. It is supposed to have greater range.

Any suggestions on receivers and emitters is appreciated.

"To make it omni directional take a piece of perspex and drill a hole half way through it. Mount the received pointing into this hole and mount the whole thing at horizontally. "

I assume perspex is English for acrylic or plexiglass. Does it matter the shape of the perspex, like would a 1" circle with the led mounted in the center work? I guess this is like a poor man's lense.

I made an Infrared transmitter and receiver, but it seems to only work when shooting directly on. It uses TSAL7600 Infrared Led and a TSOP2438 38 MHz Infrared Receiver.

Any ideas to allow it to work better from multiple angles? Anyone add a lens?

LadyAda has kit you might like: http://www.adafruit.com/index.php?main_page=product_info&cPath=6&products_id=5

Now I haven't bought the kit, but it sounds like lots of fun. And even if you don't want to buy it, you can read the instructions.

I believe it uses a shift register or latch for the same purpose you want: to conrol leds. She also has some kind of program for creating the image.

If you are looking for a simple example, try http://www.arcfn.com/2009/08/multi-protocol-infrared-remote-library.html

You can run the send examples without any hardware. It just sets a pin high. It doesn't matter that nothing is connected to the pin.

The first couple of paragraphs of his blog contain the link to the library and explain how to install it. It looks real simple.

s/MHz/KHz/p
It uses a 38 KHz carrier.  


The 20 Ohm resistor does seem like a lack of power, but it's not clear that the 47 Ohm resistor causes a power problem.

I'll try the webcam.

The batteries are regular, non-rechargeable batteries. They are fairly fresh. I don't have a battery checker, but if the power dropped, the arduino would probably reset.

Is going over 5 volts ok?

If 5.5 volts is ok, this doesn't like a battery problem.

The TSAL 7600 can take 200 mA peak. I'm using NEC format. That uses a 38 MHz carrier. When the signal is high, the wave would be high 1/2 the time. So overall when transmitting it is on around 1/4 the time.

The current through the LED would be: 5.5 volts/50 ohms = 110 mA

That sounds good too. I tried a 25 ohm resistor, but that just made it much worse. (20% success at 10' instead of 80% success at 10', head-on.)

So I don't understand why dropping the resistor caused performance to deteriorate. And it doesn't sound like there is a problem with the amount of power. But maybe the constant current supply is worth doing? Do people get good results without doing this?

Infrared is kind of frustrating. It is hard to tell what is going on.

Instructables has a circuit http://www.instructables.com/id/Power-LED-s---simplest-light-with-constant-current/, but this seems to be for a much more powerful LED, like something you would wear.

This only needs 200 mA peak.

How much power can 4 AA's make? Has any one played around with different power circuits?

What is a good power supply for an Infrared LED, TSAL 7600?
http://www.vishay.com/docs/81015/tsal7600.pdf

I'm using 4 AA batteries that are about 5.5 volts. The LED is in series through a 47 Ohm resistor. It is powered through a transistor to overcome the Arduino's current limit.

How do you tell from the spec sheet what the maximum voltage and limit resistor to use?

Ok, i'll try it without the resistor. The arduino digital output switches a mosfet transisitor, so there is no danger of burning that out.

If the Arduino is powered by USB, how much current can I draw from Vin? Can I use that or any other pin to power motor?

Can SN754410 H-Bridge be powered in same way? Can this harm Arduino?

I was thinking of building shield with this chip as it sounds pretty simple (as opposed to buying a shield. Does that seem reasonable?