IR Remote Control Issues

[Jumper]

Is it the correct wavelength? 950 nm is popular but so is also eight-hundred something and so on..... less correct -> less sensitive receiver

Is it the correct frequency? 40? 48? 56? kHz there are plenty to choose from.... pick the wrong one and you loose performance

Do you provide enough current thru the LED? Even if the LED is specified 50 mA MAX that usually means that it can handle 50 mA constantly. Since you have a 50 % dutycycle and then send bursts out it can take much more. There is a chart in the datasheet that shows max current for different dutycycles. Now you have 30-40? More is better...... many receiver datasheets mention distance when the LED currenct is 100 mA.

I am happily running my IR led's at 700 mA and that is more than enough to change channel for the whole apartment area.

Of all these questions the easiest to solve is frequency (if you have a scope). The others are more or less up to trial and error...

Break open the TV and see if the receiver chip has any markings, then you might get both FQ and wavelength.

/me

[Kamikaze47]

These are the specs on my IR LED:

Parameter                   min    typ    max   unit     test conditions
---------------------------------------------------------------------------------
radiant power               14     -      28    mW/cd2   If=20mA
peak emission wavelength    -      -      940   nm       If=20mA
spectral line half width    -      -      50    nm       If=20mA
forward voltage             -      1.2    2.0   V        If=20mA
reverse current             -      -      100   uA       Vr=5V
rise time                   -      -      200   nS       PW=10uS DC=10% Ifp=20mA
fall time                   -      -      500   nS       PW=10uS DC=10% Ifp=20mA
power dissipation           -      -      150   mW       Vr=5V
continous forward current   -      -      50    mA       -
peak forward current        -      -      1     A        PW=10uS DC=10% Ifp=20mA
reverse voltage             -      -      5     V        -
viewing angle               -      20     -     deg      -
operating temp range        -45    +25    +100  degC     -
storage remp range          -45    +25    +100  degC     -
lead soldering temp         -      -      +250  degC     5 Sec 1.6mm from base

PW=Pulse Width
DC=Duty Cycle

The frequency of 48khz i found by trial and error. It def works best at 48khz.

How could i get 700mA with this LED? Pref with 5v only?

[dhouston]

I've never heard of a 48kHz carrier. 32-40kHz is common and there are some that use ~56kHz What is the make/model of the TV?

[Kamikaze47]

Ok, I just tried it with a 10 ohm resistor, so 340mA through the LED. It only seems to have made a very small difference to the range, and I still have to point the LED *directly* at the TV for it to work - Even a few degrees make a difference.

I will try playing with the frequency again.

Also, do you think 50% duty cycle is the best bet? I wasn't sure what to use, and 50% seemed logical, but now i'm considering a lower duty cycle as that means the LED should theoretically be able to handler a higher peak current.

*edit* My TV is a Samsung LA40R71BDX

*edit2* Fixed mistake in schematic (thanks keithdoxey)

[Jumper]

http://www.sbprojects.com/knowledge/ir/nec.htm

Someone said Samsung uses NEC code 38 kHz with a dutycycle around 0.25 to 0.33

HOW IS YOUR TIMING?

Increasing IR current that much would have made a difference if that was the true problem.

Do you have a scope so you can meassure the output of the TV remote control LED? (and if you have a nice digital one you might even post the pic)


/me

[Kamikaze47]

I dont have scope unfortunately. So what I did was hook up an IR receiver to a PIC, and had it sample it every 10uS and output it to my PC which I then put into excel and made a graph. The pic attached has the signal from my TV's remote at the top, and the signal from my home made remote below. The scale on the bottom is in 10's of uS.

As you can see, the timing is pretty close. I would have thought if it was a timing issue, it simply wouldn't work at all instead of bad range?

I'm going to try 38khz at 25% and see how that goes.

[dhouston]

Samsung uses 38kHz and sends a 32-bit code similar to the NEC protocol but with a shorter lead-in pulse and with the bits interpreted differently. Samsung uses a 4.5mS start pulse followed by a 4.5mS space. Your example appears to be for the "1" key.

[Kamikaze47]

Samsung uses 38kHz and sends a 32-bit code similar to the NEC protocol but with a shorter lead-in pulse and with the bits interpreted differently. Samsung uses a 4.5mS start pulse followed by a 4.5mS space.

Sounds exactly like what i've observed. Do you know of a page on the net somewhere that has more detailed info on the protocol they used? I havn't been able to find any info on it.

[Jumper]

When you PIC-scoped the remote what receiver did you use? What is the range and sensitivity using the standard tv remote control and this receiver?

Then we can get an idea of the FQ used. If you have good distance with the remote then it must be your LED that has a major output power problem. It might be burnt due to over current along the evolution chain.

Have you tried to move one thansistor to the base of the other, so you do all the mixing on the base.

Maybe just loose both transistors and sink current with the PIC and make a software mixer by turning on and off HPWM (and connect the led there with a resistor)

what partnumber is the led?
what part number is the receiver you have used?

/me

[dhouston]

Sounds exactly like what i've observed. Do you know of a page on the net somewhere that has more detailed info on the protocol they used? I havn't been able to find any info on it.

I don't recall seeing any detailed description. There is a simple description at http://www.techdesign.be/projects/011/011_waves.htm and there are CCF codes on RemoteCentral.com and at http://ir.premisesystems.com/. There are a few detailed descriptions of the NEC protocol. It was the first and was copied with minor changes by many others. Here's a datasheet for an early NEC encoder chip - http://davehouston.net/Nec6121.pdf

EDIT: BTW, I find it easier to record the codes as .wav files using a soundcard. This way you can use a wave editor to view them. See http://davehouston.net/learn.htm

OTOH, if you sample with a PIC using a phototransistor, you can get the carrier frequency and eliminate any guesswork.

[ardhuru]

Ok, I just tried it with a 10 ohm resistor, so 340mA through the LED.

Just realised; if my math is right, you are *not* getting 340 mA thru' the LED with a 4.7K bias resistor; I would reduce that to 1 K; at the most 2.2 K.

Regards,

Anand