More IR circuit design questions

[cyborg5]

I realize I'm way off topic here but this forum has been so helpful with my previous questions about IR circuit design that I figured I would come back and ask a couple more questions. Much of my problem comes from the fact that I don't 100% understand how transistors work although I'm learning a lot.

In the previous thread you helped me design a circuit based on the output portion of the TV-B-Gone in which I have an Arduino output pin connected to the base of a NPN driver transistor via 1K resistor and that in turn drives four PNP transistors and IR LEDs. I know this is opposite of the TV-B-Gone but that's immaterial right now.

I understand that this circuit is pushing way more current through the devices and their maximum ratings and we get away with it because we're only sending brief pulses. However one of my applications may be to control and IR control toy helicopter. While I've got some data on the protocols used to control the helicopter, I'm not really sure (until I buy one) how frequently the transmitter sends a signal to the toy. My guess is that it keeps sending the code continuously as quickly as it can. Even though it's a modulated 38 kHz signal and not all of the bits are on, my concern is that if this circuit would be in continuous use over several minutes. The datasheet talks about peak amperage with less than 1% duty cycle and we are way over 1%.

To simplify things let's assume we have a single transistor driving a single IR LED and the base voltage is enough to keep the transistor fully saturated. We're also working on a 5 V system. My datasheet says that the maximum continuous collector current is 600 mA. If I'm applying my Ohm's law properly that tells me I need at least 8.333 ohms of load to keep the current at or under 600 mA. Is that correct?

Putting IR LED between the collector and +5V by itself I presume has pretty much zero resistance. Therefore to keep the transistor from overheating/blowing up etc. I would need some sort of a current limiting resistor in line with that to keep the transistor current under specification. Is that correct?

Given that the continuous current rating of the IR LED is 100 mA that means I would need an even bigger resistor than 8.33 ohms. Again assuming I understand Ohm's law that means at least a 50 ohm resistor given 5 V supply and a 100 mA maximum rating. Is that correct?

I'm pretty confident that I understand everything I described above. What I don't understand is in the TV-B-Gone circuit or in my NPN/PNP swapped version it looks to me as though when everything is switched on you basically have a dead short running through the LED and the driving transistor. Is there internal resistance in a fully saturated transistor that limits its current? Is there internal resistance in a switched on LED that limits the current? Or are you merely relying on the limits of your power supply to deliver current? If there is some sort of inherent resistance when these devices are switched on, where do I find that in a datasheet?

[gee]

A few things limit the LED current:

- LED forward voltage increasing with current. There's usually a plot of Vf versus current in the datasheet.
- Transistor gain - there's usually a Hfe versus output current plot in the datasheet. Output current won't exceed the base current (set by the base resistor and drive voltage) multiplied by the Hfe.
- Battery resistance. Alkaline AA's can't source much more than an amp, lithium AAs are capable of more.
- Wiring resistance, etc.

Unfortunately nothing's really reliable here - LEDs can have varying forward voltage, transistor hFE's vary, batteries can have different resistances based on their state of charge, etc... so there's no easy way to determine exactly what the LED current is going to be with the existing circuit.

I went with the constant-current LED driver circuit I poked in the other thread for this reason, I like circuits to work deterministically.

[john444]

Hi Cyborg,

We're also working on a 5 V system. My datasheet says that the maximum continuous collector current is 600 mA.

Check the Vf of the LEDs you are using.
the Forward Voltage drop is probably about 1.5V. Also, the LED's transistor probably has a saturated 'ON' voltage of 0.3~0.6V. Again, check the datasheet.
So, 5V - 1.5V - .6 = 2.9V, 2.9V/0.6A = 4.83-ohm

I would need some sort of a current limiting resistor in line with that to keep the transistor current under specification.

Yes, that is correct, the ~5-ohm resistor above. Although, it depends on the current capability of the 5V supply.
The TV-B-Gone relies on the internal resistance of the battery to prevent blowing the LEDs and transistors. Depending on how you are powering this thing, you may not need the 5-ohm resistor.

at least a 50 ohm resistor given 5 V supply and a 100 mA maximum rating.

Just like above, you have to subtract the Vf of the LED and the Vsat of the transistor. (5V - 1.5V - 0.6V = 2.9V, 2.9V/0.1A = 29-ohm

Is there internal resistance in a fully saturated transistor that limits its current?

Not really. It has more to do with the base current and the transistor's gain. If the transistor can stay "fully saturated" and below it's power dissipation, then the "on resistance" will be nominal because it is related to the somewhat fixed Vsat voltage.

Hope this helps, John

[cyborg5]

I'm understanding this much better now. I guess I was little confused about the term "Forward voltage". I thought it was just the minimum voltage necessary to turn on the LED. I wasn't thinking of it in terms of a voltage drop across the device. Makes sense now.

[john444]

Cyborg,

I thought it was just the minimum voltage necessary to turn on the LED.

Almost. It is really the approximate LED voltage drop.
It is generally not practical to supply a fixed voltage to operate LEDs because the current changes very rapidly within mV of the forward voltage. The usual way to operate LEDs is to provide them a current (set by a resistor) and let the forward voltage be whatever the LED wants it to be. It should be close the the datasheet Vf but can change slightly with temp and current. For example, the Vf will only change about 10mV when the current through the LED changes from 1mA to 10mA. Try it yourself when you have the chance.

Good Luck, John

[cyborg5]

Another thank you to all the people who helped me with this project from this forum. I thought you would like to see the results of your efforts.
http://tech.cyborg5.com/?p=192

[john444]

Hi Cyborg,

Nice write-up.
And, your boards look great.