Well it will do something but the resistors to limit the current in a high power LED is not a good idea.
Are these high power LEDs fast enough or is this switching speed far to high?
I would say that 2MHz is way too high. Why do you want it so fast? It only makes the decoding harder as you have to have an amplifier that is tuned to 2MHz which is hard to make for no advantage. This sort of thing normally works at 40KHz.
Does the parallelizing of the ANDs help in switching speed of the FET
A bit, but not as good as having a proper FET driver. It is the resistors that will slow things down.
The LEDs have an viewing angle of 135 degrees, will they do any harm to the eyes, if you look at them?
It could do because you have no blink reflex at IR. But that is only looking at it close up.
That's right, but efficiency is not so a problem and I don't know how to build a constant current driver, which is switchable so fast.
Using 2 MHz (also trying 1 or 5MHz) is an experiment. I'll try to use a shortwave receiver with a big lens and a photodiode as "antenna". I don't know, if this will work.
Are there simple to use FET drivers available for these frequencies?
Ok, I make same warning signs on the device, not to steer on it.
That's right, but efficiency is not so a problem and I don't know how to build a constant current driver, which is switchable so fast.
It is not a matter of efficiency it is a matter of not burning out your LED. The forward voltage will change with age and temperature. As these LEDs will get hot then the temperature of them can vary over several minutes.
I'll try to use a shortwave receiver with a big lens and a photodiode as "antenna". I don't know, if this will work.
No I don't know either but I would be astonished if it did. You need a photodiode that can respond at that sort of frequency. These are specialist devices used in fiber optics. Note that your big lens will not work well for IR as IR is absorbed by glass.
1- The IR LEDS are labeled "Serial"
The correct term is SERIES
2- output circuit values are correct:
4 * 1.6V = 6.4V (across four leds in series)
12V-6.4V = 5.6V (voltage to be dropped)
5.6V/.35A= 16 ohms (value of Rdropping to drop 5.6V @ 0.35A)
33 ohms || 33 ohms = 16.5 ohms (close enough)
Looks good to me.
I assume you have a reason for wanting to pulse the led at 2 Mhz.
Whether or not we need to know that reason is unknown at this time.
I am assuming an arduino digital output can drive four TTL AND gates.
I don't see any issues.
There's nothing wrong with the concept of a dropping resistor. Actually, if you do the math, it is underrated if anything because the maximum pulsed current ( 100uS pulse width) is typically 10 x the continuous current and the period of a 2 MHz pulse is 500 nS which is a factor of 200 less than the maximum pulse width of the pulsed current rating. I wouldn't be surprised if it worked with a dropping resistor 1/10 th the current value (two 3.3 ohm in parallel). I max = 12V/1.6 ohms with a pulse width 1/200th of the pulsed current pulse width allowed.
raschemmel:
There's nothing wrong with the concept of a dropping resistor. Actually, if you do the math, it is underrated if anything because the maximum pulsed current ( 100uS pulse width) is typically 10 x the continuous current and the period of a 2 MHz pulse is 500 nS which is a factor of 200 less than the maximum pulse width of the pulsed current rating. I wouldn't be surprised if it worked with a dropping resistor 1/10 th the current value (two 3.3 ohm in parallel). I max = 12V/1.6 ohms with a pulse width 1/200th of the pulsed current pulse width allowed.
Sorry, not really. The 100uS/10x specification applies only to a 10% duty cycle (for many of the pulse power specs that I've read). The duty cycle in this 2Mhz application is 50%, not 0.5%.
But if the LED can handle 350mA continuous anyway, it's a moot point.
That 350mA will cause the LED to heat up, when it does it will draw more current, and get even hotter. That's why the constant current driver is used, to limit the current flow as the LED heats up and lets more current flow if it's available.
Well, that one seems to operate with PWM at 600 KHz. So it would not be just a drop in for the OP's circuit. It's not designed for high frequency modulation, only at 100 Hz or so.
2 MHz, 50% duty cycle, is 2.5uS high/2.5uS low.
74HC08 looks to be plenty fast to keep up with that, with < 0.1uS times.
At 2 MHz switching speed, will the LED even turn off enough to be seen as off?
Is there a part number for it?
For use as a carrier, it would be best to have a crystal controlled frequency. Otherwise you are going to have a tough time staying on the signal at the receiver.
The switching speed of the LED is indeed a concern.
I explained that - the current draw will increase as the LED heats up. Fixed resistor will not prevent that.
Further discussion is pointless until we see a spec for the LED and see some info on how it's being cooled/it's heatsink/etc.
raschemmel:
I am not clear on why a constant current source is necessary
To make matters worse in this case you have four LEDs in series so any change in the forward volts drop per degrees C is multiplied by four, causing even more current to flow.
Anyway I don't know why we are arguing about this aspect, the whole project is dubious in my opinion.
Although you can not create a true current source with a resistor, you can reduce the dependence of current flow on the difference in voltage drop caused by junction temperature change. Just increase both the resistance and voltage. This will waste a lot of power, but it might get the job done reliably in a simple way.
