Resistor rating for IR LED for WiFi IoT remote

Here's a bit more information:

The github repository for the library I'm using to handle the remote control details for the heat pump are here: https://github.com/crankyoldgit/IRremoteESP8266

I found a few places that discuss the protocol used for my heat pump, including this one: http://rdlab.cdmt.vn/project-2013/daikin-ir-protocol

In summary, it appears to be 35 bytes sent, with (roughly) at most a 50% duty cycle, but it depends on the bits - if a bit is 0, the duty cycle is about 33%. So Wolfgang's assumption of a 50% duty cycle is a safe choice.
The total signal time appears to be about 41 milliseconds if all bits are 1 (= 50% duty cycle). if the bits are 0 the time will be longer but the duty cycle will be lower.

I should sign off for the night - tomorrow is chicken wrangling day. sigh. my life has gone weird LOL. they're not my chickens it's a small flock of 7 so hopefully we'll get their feet treated without too much drama

Thanks,
-Nico

Hi all,

I just found some additional information on Adafruit that I think will clear up my resistor current worries.

They are using a Raspberry Pi Zero with a QT-Py-Hat as a WiFi controlled IR remote control, which is exactly what I'm aiming for too, so I have hope it will also work in my setup.
This is the Adafruit page with all the details: https://learn.adafruit.com/iot-ir-remote-using-raspberry-pi-zero-w-and-qtpy-hat/parts-required

They recommend using 2 LEDs, where one is a wide angle LED and the other one a narrow angle LED. Each LED gets a transistor, which are both switched with a 3rd transistor that is controlled by the micro controller.

The recommended IR LEDs are rated to 100mA with 1amp max, and 1.6v forward voltage. My LEDs have the same spec, with higher peak amps, so I assume they will behave in a similar way.

For the resistors, they use ("optional") 33 ohm resistors, which if the online calculators are correct, comes out to the basic 100mA rating of the LED. Their explanation says these are optional and indicate they are only there to limit current if the LEDs draw so much power that they crash the main board. They are just using basic 1/4w resistors in that case too.

They also have this IR-only project, using the same kind of circuit.
https://learn.adafruit.com/building-an-infrared-transmitter-and-receiver-board
The schematics for that are open source too: https://github.com/cyborg5/IRIO3 

"The first option is to add a 33 ohm current limiting resistor in line with each of the LEDs. Without them, this circuit will drive the IR LEDs beyond their rated continuous current capacity. However because IR signals are intermittent and modulated it is safe to do this."

This is a close-up of the schematic of the QT-Py-Hat on the Adafruit website, which has an IR transmitter on it. It's open source. (I've erased the bits that don't apply to my case).

So, I'm guessing if they are leaving out the resistors altogether and aren't worried about burning the LEDs, then I probably don't need to worry about the resistor rating too much either, especially if I keep it within a reasonable range.

-Nico

One more question...

Is this safe math I have here?

If I run the LEDs at 1amp, using a 3.4 ohm resistor (1/4 watt rating), the power is 3.4 watts.

But the duty cycle while sending is about 50%, so that brings us to about 1.7 watts.

That only runs for about 50 milliseconds.

I can very easily limit the number of times the signal gets with a software delay.
It makes sense to only allow one message per second, which cuts it down to 1/20th of the time. In actual use, I only expect it to send 2 packets a day (night mode and day mode).

So if I limit sending to at most once per second, then the average power over that second is about 0.09 watts, or only about 1/3 of the rated 1/4watts the resistor can handle.

Is it safe to assume power dissipation levelling over a 1 second period? I guess it only needs the first 1/3 second.

ps, by "safe" I mean I'm more concerned over fire safety risk than about resistor failure.

Thanks!
-Nico

If you lose a parallel resistor, the resistance goes up and the current down.

How will that scenario lead to led burn-out?

Hi All,

It works!

It seems I was overthinking things, as usual :D

I just tested a 2-LED circuit using 22 ohm resistors, which runs the LEDs at about 150mA, and it seems to be plenty to reach the heat pump.
So my aiming for 1 amp was really just far more than is necessary. And that would also be over the rating of the transistors I'm using  
I think the reason for it not working before might've been due to the way I was using the transistors - the proper circuit uses an extra transistor just to trigger the LED-powering transistors.

The math:
If we assume it's still a 1.6v voltage drop on the LED, then the voltage drop across the resistor is 3.4. 
At 150mA x 3.4 volts x 50% duty cycle, we get about 0.25 watts, which is exactly the 1/4 watt rating of the resistors.
Voltage drop at 150mA will be a little higher, and 50% duty cycle is a theoretical worst case which never happens (bits set to 1 have a longer pause), and there will always be pauses between transmissions much longer than the 50ms packets too.

So I feel much better about the current running through the resistors now - it peaks close to the 1/4 watt rating but doesn't go over, so even accidental continuous transmission should not be an issue.

Here is my cleaned up circuit, but instead of 33 ohms, I'm using 22.

And here is a picture of it on a little breadboard, powered by a Wemos D1 Mini.

Thanks again for all of the excellent advice!

-Nico

I like the minimalist approaches.