LED Strobe questions

Hello!

Can luxeon-type LED's be used effectively for a simple strobe?

Example LED:


My goal is a bright strobe that is reasonably noticeable in daylight to people who are looking at it and waiting for it to flash. It's a signal. It needs to be effective out to 400 yards. It doesn't need to "catch attention". It only needs to effectively signal an event to observers who are already keenly staring at it and waiting for the signal flash to occur.

Secondary goal: Cheap. Simple. Reliable.
Power source is a battery.
Remote trigger is a vibration sensor on a fixture.
Signal strobe is an array of three of the above LED's mounted in 30° flashlight optics, oriented at the observers.

So my first questions are about the suitability of pod of three of these emmitters for use as a strobe. If the duty cycle is 50% and the frequency is something eye-catching like 3 cycles per second, does this allow enough time for the emmiters to come up to full brightness for each flash?

That's my current bottleneck. I don't know how much time is required for an LED of this type to ramp all the way up and then all the way back down, assuming full recommended current is switched at it in a strobe-type time frame.

I would not worry about the rise/fall times in your case.

I would worry that the lumens would be high enough and the pulse length is long enough to be detectable.

You can exceed the rated maximum current for short pulses;

Item Symbol Absolute Maximum Rating Unit
Forward Current IF 25 mA
Peak Forward Current IFP 100mA

Kam, the LED shown is a high power one, continuously 700mA or more. You are talking about an ordinary little 20mA LED.
Our vision sees a flash duration of 30ms or longer at full brightness. 3 flashes per second at 50% duty cycle is a duration of 167ms so the light will be as bright as if it was continuous.

Cool! Thanks to both!

I was wondering whether it was safe to spike a little over max rating in a strobe application.

Yes - these are 5W 700ma emitters. They don't have a lm rating, but I suspect it's ~ 200.

That's good info about the human eye capability. As you say, each flash will be "on" for about a 6th of a second. My initial concern was about the rise/fall requirement, and whether common high-power emitters are able to achieve full brightness in 137ms (using guru's example, 167-30=137ms assuming that you want full brightness sustained for 30ms).

If so, the next step is a driver circuit that can keep up with a timer circuit that sends signals at the described frequency.

My initial thoughts are for overall design: [piezo - op amp - 555] +relay+ [driver - led]

(the op-amp is recommended to shield the timer from potential v spikes from the piezo)

This is basically a door knock sensor with led signal. The actual project is a bullet impact sensor for a 24"x24" hardened steel target positioned at up to 400 yds from an active shooting position. The local noise level prevents audible impact from being a reliable hit indicator. The solution is a vibe sensor on the back of the target that triggers an LED signal strobe positioned 10' away from the target.

There are commercial variations available, very expensive. Most seem to use an arduino kit or a microprocessor design. My goal is to build four kits with an old-school thru-hole design. I'd like to keep the kits under $20 each. I already have boards, resistors, other common stuff.

Measurement Specialties offers floppy piezo sensors for less than $3. The emitters are < $2. I already have a fistful of 555's. Op amps are $3. Relays are cheap. I think it's doable.

My next bottle necks:
1 - Driver circuits. I built some daedal constant current drivers several years ago. they're still going strong. Maybe i can use something like that, but I don't have any experience with the whole rise/fall parameters. I just don't know what is fast vs what is not.

2 - Layout. Given 12' of cable built into the design, I presume the best approach is to put everything but the LED's in a magnetic-mounted box that slaps on the back of the target. Maybe its a two layer box where the piezo is exposed to full shock in layer 1 but the rest of the circuit is cushion-protected in layer 2. The driver then pushes the necessary I across 12' of cable?

Thoughts?

Thx!

I guess this is going to be powered by a 12V car battery? The LEDs, what are their ON voltage?, it looks as though it will be 4 V, which is unfortunate. The easiest way to drive these sort of beasts is to string say 8/9V worth in series the a power transistor then a very low value resistor across the battery. You then sense the voltage drop across the low value resistor and use this as one input to an opamp, with the other going to a reference voltage, the opamp driving the power transistor. So now the current through the LEDs will be stabilised. Use your 555 to switch the reference voltage ON/OFF. You could use a reflector from a cheap torch to "focus" your LEDs light.
Frank

The rise/fall time of the LED light is insignificant, I know of someone using them to transmit a data signal where the pulsing is several million per second. I would be more concerned about how to stop them flashing. I assume you want several flashes as the target is hit then it should reset in readiness for the next hit. You can use a 555 to generate the flashes and maybe a second one to set the flashing duration but two timers and an op-amp makes me think in the realms of a simple microcontroller. It would be cheaper, smaller and more 'customizable' .

Brian.

Thanks guys! I'm pretty much a noob, and your feedback really helps!

The emitter is rated 6~8v.

Batteries I already have available: 12v marine and 6v golf-cart (both are large deep-cycle). Fully charged, they run at ~14v and ~7v, resp.
I also have a dozen 3.6v 14500 batteries.

Job parameter: starting with a fresh battery, the signal apparatus must function for 6 hours in standby mode and be able to signal up to 100 events during that time. So the big batteries are more than adequate. If possible, it would be nice to get it done with 2 or 3 14500 in series. I guess this will depend on circuit voltage drop and also standby drain.

I need to learn more about 555's. My perception is that they generate an output signal during an interval bounded by incoming trigger and threshold signal events, and the freq and duty cycle of the output signal is determined by cap and resistor selection. This seems like a good match for the behavior of the piezo (voltage spike caused by membrane flex). If the piezo is the trigger, the 555 will only output a signal when the piezo is wiggling, and the 555 will stop outputting a signal when the peizo stops wiggling? If it's that simple, then I'm inclined to take that approach.

If I'm missing something, then I guess I have more research to do. As stated, "old-school & cheap & reliable" are my initial priorities, but I'm certainly not opposed to learning some basics about simple microcontroller circuits.

Thanks!

Two series power white LEDs will work fine directly across two 3.6v 14500 only when Charger is OFF.
They will be much dimmer at 3V per LED which is when the SoC of LiPo drops to 0.

MOSFET switch must be much less than ESR of LED's which is around 1 Ohm @ 1Watt.
So choose <<100 mOhm. with logic level drive.

Your assumption about the 555 is not quite correct. They can be used in two modes: Astable and Monostable. In Astable mode it produces a signal that oscillates (flashes if you prefer that analogy) continuously, in monostable mode it produces a fixed time interval which starts with the trigger and runs for the duration of a period decided by the resistor and capacitor values.

Question: when triggered by an impact, what exactly do you want it to do?

A monostable will produce a single flash of light then wait until next triggered. If you want it to flash several times after an impact, you can do it with a combination of a triggered monostable to set how long it flashes for and an astable to create a rapid flashing. The astable can be held in a reset (dormant) state until the monostable wakes it up, after that it flashes until the monostable 'times out'.

Brian.

betwixt said:

.... you can do it with a combination of a triggered monostable to set how long it flashes for and an astable to create a rapid flashing. The astable can be held in a reset (dormant) state until the monostable wakes it up, after that it flashes until the monostable 'times out'.

Brian.

Click to expand...

Thanks! That makes sense. Use two 555's.

Ideally, impact will cause the strobe to flash like hell for 3 seconds and then reset and be ready. The shooter's objective is to hit the target once during a timed stage of targets at various distances with various firearms. 3-Gun!

Thanks also to SunnyMan. My 14500 are L-ion and I only have them in the charger when I'm charging them. The rest i didn't understand, but cheers anyway.

I'll try to come up with a schematic in a few days and maybe you guys can tell me what you think.

Audioguru said:

Kam, the LED shown is a high power one, continuously 700mA or more. You are talking about an ordinary little 20mA LED.

Click to expand...

The principle is the same, although of course the parameters are different.

However, after pondering, I think it is making it more complicated than necessary. Just flash the LEDs for a second, and repeat N times. Whatever value of N makes you happy.

As for a current limiter for the LEDs, there is a nice 2 transistor design [in 2 versions, the other uses a mosfet and transistor] around - a search on the internet should uncover it. It is pretty voltage supply tolerant [current limit is not so dependent on the voltage supply]. If the OP can not find it, I might have it in the dark deep recesses of my computer.

The OP mentioned using an op-amp to limit the voltage to the 555. I would just use a resistor and some diodes [say in4148's] to clip the voltage - never mind the op-amp.

I found a different 10W emitter. Specs:
V-f = 3.0 to 3.5v.
I-f = up to 3amp, but they are plenty bright at 2amp.

14500 battery = 3.6v nominal, 4v peak.

If v-battery matches v-LED this closely, why not direct drive the LED with nothing more than a small resistor in series?

Calculating a value for the resistor: assuming 4v from battery and 3.3Vf = 0.7/2 = 0.35 ohm resistor. Basically zero. But I-f squared over R says it would want a big wattage resistor.

If the calculated R for series resistance is that low, why not just direct drive this 10W led from a 14500?

Thoughts?

Resistance is a very relative thing, if you consider 0.35 Ohm is low and can be left out, do some math. Suppose you didn't take it out but changed it to say 0.1 Ohms, recalculate the current and see the result, then try 0.05 Ohms and do it again. When dealing with high currents even a small resistance is very significant!

Leaving it out would, if the battery was charged, kill your LED in a fraction of a second. LEDs are constant voltage devices, up to their design forward voltage (Vf) they gradually start to conduct but they will try to draw as much current as they can, and overheat in the process, if you force higher than Vf across them. The resistor is necessary to keep the current in check by dropping the voltage between the battery voltage and Vf.

I'm not sure you really want a single high power LED for your application. I'm guessing from your description that the purpose of this device is to let someone know they hit a target over the length of a shooting range. For that you might find it better to use an array of lower power LEDS. Low power LEDs (5mm or 10mm types) tend to have a narrower light beam which can be better aimed at the shooter and by spreading them out (5 x 5 matrix?) you make them easier to see over a distance. A single high power LED will be a tiny spot of intense light that can be seen over a wide angle. Its your choice of course but before working out the complexities of pulsing high current LEDs, consider whether they are really what you need.

Brian.

A lithium battery is usually charged to 4.20V, not 4.0V. Then the current in a simple resistor will be too high when the battery is fresh out of the charger. Then the LED gets dimmer and dimmer and .... as the battery voltage runs down.

Let's cut to the chase...

Ok you can power up one or more 3W LED's directly off a fully charged LiPo and pulse it with a 50 milliohm MOSFET which is reasonable as each 3W LED is around 300 mOhm ESR.

But trying to detect it at 300 Yards is a different matter.

If you calculate the arctangent angle of a emitter radius at 300 Yds and if it is less than 0.05 degree, the brightness will drop out more rapidly than 1/R squared.

YOu might be able to see a stop light due to it's diameter, not just brightness.

You can see faint stars which are incredibly bright due to some sparkly diffusion in the atmosphere.

Thanks again for all the feedback.

I played with the 3.5 to 4.2 battery voltage range, across various forward-current / resistor scenarios, and i see now that current apparently has a near exponential relationship with V in a DD circuit. Got it, thanks.

I'll be using some 15° optics on the emitters, unless i can find something narrower.

My current plan is at least 3 parallel emitters on the strobe. Redundancy -> reliability.

Cheers!

DS