# IR switching of battery-powered LED

#### ryanlloydmiller

Joined Oct 31, 2011
15
I've spent the last couple days learning about electronics trying to come up with a solution that has thus-far evaded me.

I need a small, simple, and fast circuit to take pulsed IR light and a 3.6 v battery to switch on a rather powerful LED (2.5-3.5v 0.3-1 amp) only when the IR is flashing at 40 khz or whatever. Small meaning fitting on ~ 2 x 1 cm board (not including battery). Fast meaning <0.5 ms on and off. I'm swimming in photodiodes, phototransistors, transistors, relays, opamps, analog switches, etc and I need something concrete. Any plans I can find either don't tell you which parts or list part numbers that are now obsolete. Please help!

#### Bernard

Joined Aug 7, 2008
5,784
Over what distance must it operate? A little tighter specs on LED would help.

#### thatoneguy

Joined Feb 19, 2009
6,359
Are you capable of working with SMD (Surface mount components)?

A wideband 38kHz receiver can be had at Radio Shack for cheap, a faster one from an electronics warehouse. Amplify that signal, then use the signal for a constant current LED Driver, which the choices are mostly surface mount. The transistor would also need to be surface mount, in addition to bias resistors, etc.

When you say "Fast", do you mean on within 100uS or 1mS? Does the off time speed matter, or just the time to turn on the LED once an IR signal is received?

How wide of an angle will the IR Receiver need to cover for the possible input? Is it guaranteed that only one IR Transmitter will be operating at any given moment? I read it as only a carrier wave needs to be present, no modulation/code sent?

If this matches what you are trying to do, I'll scrounge up an LED driver (the hardest item to fit in the area required. Also, the LED posted may need more copper for a heat sink than $$2cm^2$$, page 19 of the datasheet shows a 2cmx3cm PCB heatsink in addition to heatsink for the driver for 1A operation, so the size will be closer to 6cmx4cm or so for the full circuit (all SMD), with plated through etched holes - requiring a professionally made PCB. So simulate solution much and build on large scale before ordering PC Boards.

#### ryanlloydmiller

Joined Oct 31, 2011
15

I think I can work with SMT, assuming it takes a fine-tipped soldering iron and a steady hand, but I can't say for sure.

By "fast" I mean on and off within ~0.5 ms.

I'm not too worried about the angular coverage...I can place as many IR LEDs as I need to in order to cover the area. So multiple LEDs will be delivering the 38khz carrier signal with no modulation/code.

One important point: the LED will never be on for more than about a second, and usually more like 100 ms, with several seconds off in between times. So I'm hoping a heatsink is not required. Also, it will typically be running off around 0.5 A, I expect.

Thanks again.

#### ryanlloydmiller

Joined Oct 31, 2011
15
I'd rather not use a LED driver. I think a serial resistor is good enough for my purposes at this point. The part I need help with is the amplification.

#### Bernard

Joined Aug 7, 2008
5,784
I think it fits: Heat sink, HS, 1 X 2 cm copper,or Cu clad board, IR detector-TFM5360, 2N3906, PNP or SM equiv, FZT849 SM NPN,med power, LXML-PBO1-0040 LED, & 3 resistors. The IR works fine on 3V, specs @5V & gives 36kHz bursts as 3V neg. going detected sq pulses, worked with 3 different remote controls. 9.9mm wide, 5.6 mm thick, 12.5mm tall, [ height limitation not yet specified], mounted dead bug style with 2N 3906 attached. FTZ mtd. vertical with collector soldered to HS.LED flat, glued or soldered to HS. Light in 90 degrees to light out.

#### Bernard

Joined Aug 7, 2008
5,784

#### Attachments

• 130 KB Views: 41

#### tylernt

Joined Nov 2, 2011
2
I'm not too worried about the angular coverage...I can place as many IR LEDs as I need to in order to cover the area. So multiple LEDs will be delivering the 38khz carrier signal with no modulation/code.
If the multiple transmitters are not synchronized, won't the detector see what looks like a constant-on IR signal instead of a wave?

#### Bernard

Joined Aug 7, 2008
5,784
Lost last post: Change FZT base resistor to 56 ohms.

#### thatoneguy

Joined Feb 19, 2009
6,359
If the multiple transmitters are not synchronized, won't the detector see what looks like a constant-on IR signal instead of a wave?
That is what I was thinking, needs to be a way to synchronize the IR emitters, or ensure only one is on at a time to make it through the 38kHz bandpass filter.

#### ryanlloydmiller

Joined Oct 31, 2011
15
Thanks everyone. All IR emitters would be receiving the same signal. I'm thinking 4 LEDs in 2 parallel sets of 2 in series. Wouldn't this work?

#### ryanlloydmiller

Joined Oct 31, 2011
15
Bernard: Thanks for your help, its exactly what I was hoping for.

If I understand your terminology correctly, you're suggesting mounting everything on a copper board (which doubles as a heatsink [for the LED and FZT849 transistor?]) with the pins sticking up in the air and soldered together? I didn't mention this, but ideally the IR will come in 180 deg to light out. This should be doable I reckon by gluing the IR and LED edges to the edges of the board.

Also, is there a reason the TFM5360 was chosen for IR detector? Its a little difficult to source (not at Digikey or Mouser) and its fairly large. I have a 40 khz receiver from Radio Shack thats about half the size and works pretty well...but it seems that you're saying the TFM5360 outputs something different?

#### Bernard

Joined Aug 7, 2008
5,784
The TFM5360 was in hand, 180 deg receiving angle, poor output- it dies if loaded more than 680k to gnd & 1.5k to VCC, so 2N3906 would not work- needs a buffer or 3V gate FET-p ch. Could never find a data sheet on it.
Do you have specs on 40kHz detector? Looks like you understand my suggestion on mounting.

Last edited:

#### ryanlloydmiller

Joined Oct 31, 2011
15
This is the one I have but I can get whatever. In order for your circuit to work, it just needs to pass current when no IR is present and switch off when IR is detected, correct?

#### Bernard

Joined Aug 7, 2008
5,784
Did the package come with any information? A CMOS IC, 4049 or 4050, could be used as a buffer between detector & 2N3906 if it has poor drive capability. There must be a better choice of circuits.

#### Bernard

Joined Aug 7, 2008
5,784
Probably both IR detectors are about same operation, so adding a 2N 3906 as emitter follower to drive other 2N3906 might work. Will be back monday?

#### Bernard

Joined Aug 7, 2008
5,784
RS IR detector has better sink capability. Tested with 3Ω paralled with 100mA white LED as load. FZT E-C .i2 V. Nice bright flashes.

#### Attachments

• 202.2 KB Views: 28

#### ryanlloydmiller

Joined Oct 31, 2011
15
Excellent. I'll try it.

What's the advantage of using two transistors? Couldn't I just use the 2N3906 with a resistor/LED on the emitter?

Last edited:

#### Bernard

Joined Aug 7, 2008
5,784
The IR could not deliver the necessary 50 mA base current, & 500 mA collector current would max out the 2N3906. If you had just a little more V, then possibly could use a PNP darlington [ possibly NTE2342, 1A, gain 2000, TO92 ] as direct drive.