White LED strobe

[jollyrgr]

At a ham fest I saw a vendor with an attention getting strobe. But the strobe was not the normal xenon type that have been around for years. This was a white LED strobe. But the LEDs were being pulsed so quickly that it resembled a xenon type strobe. And it was as bright if not brighter than a xenon strobe.<p>I have several questions.<p>Does anyone know of a low cost source of white LEDs? Sellers on eBay sell these at about $0.25 each in quantity. The only problem is most of these sellers appear to be in Hong Kong. The cheapest US source I have found wants $1.40 each. (I know there has to be a US vendor selling these cheap somewhere.)<p>Has anyone made an LED strobe? I've made LEDs flash but I have never had one appear like a strobe (except one brief flash and the LED didn't work anymore). If so, what kind of arrangement is used (i.e. duty cycle, current pulse etc.) to make an LED appear more as a strobe instead of a flashing LED?<p>Ideas?

[bwts]

Not sure about the duty cycle (id go 4 50%) but as long as ur frequency is below 50Hz then the flashing (strobing) can B detected by the human eye. 555 timer could do this! (unless there's somefing obvious ive missed)<p>B)

[frhrwa]

not sure of pricing or which led your looking at, but this place has been really good with the pricing on the LED replacements that I needed..
http://www.superbrightleds.com/leds.htm someone else in an earlier message pointed me here..

[Edd]

Hey Jolly Roger:
From the aspect of the minimum time to effect response of the human eye and its retentivity , any time greater than 100 millisecs on a single light burst would be overkill from the taxing on the LED.
Taking a look at the power specs of the particular LED that you are working with then:
On the total power drive to the LED then you might consider the capability of an increase of the drive level proportional to the ratio of off time in respect to on time, thus being able to hit that device with quite an intermittent high drive level, proportionately speaking.
Now with the pricing of the white units the smart thing to do would be to use a red LED (Cheeeep) unit within its spec paramaters as a “sacrificial lamb(s) to see how hard it could be driven “ before moving on to attempt a white unit…naturally with its higher semi junction voltage specs.
On failure mode ; probably unknown to the unitiated:
One last consideration, the fine gold whisker coming down to ball bond to the LED semi die is of some import. On hi current overdrive it tends to start a warm up…like a filament…..at a certain heat level it starts to discolor the acrylic lens potting medium up to a point where…under observation of a hi power strereo microscope…it takes on the appearance of a microscopic black “fly speck ?” which then progressively further enlarges in size on overdrive. There is a final, non regressive, diminishment of optical center light output around the spot on a unit overdriven to that point. As a comparative analogy, much like the onset of glaucoma on the human eye….except to a state of blackness instead of cloudiness.
Certainly this happens on all colors of LED’s ….its just easier to detect on a clear lens.<p>73's de Edd
[email protected] .........(Interstellar~~~~Warp~~~Speed)
[email protected]........(Firewalled-Spam*Cookies*Crumbs)

[sundancer87]

Hey EDD, how about doing us, the uneducated, a favor. Perhaps you could spend less time and verbiage trying to dazzle us with details and baffling us with bullshit and just answer the question in terms that most of us understand.<p>I had to read through (several times), your rather long-winded explanation of duty cycle and timing just to figure out what you were attempting to convey.<p>Now I have a headache and I'm sure Jolly Roger hasn't a clue as to what you thought you said.<p>I thought we were here to help those with questions, not to impress them with worthless ramblings.

[russlk]

Sundancer87, please don't look a gift horse in the mouth. So you had to work a little, it won't do you any harm. Edd's info was all good and I understood him fine.

[cato]

Ahmen Russ, I liked, and was grateful for, Ed's post too. It's the questions which come in the form of half sentences and incomplete thoughts that drive me to distraction...

[bodgy]

Ah a man who knows exactly how not to gain clarification on something that is not clear to him/her/it . <p>Something along the lines of Edd, I didn't understand ------ is possibly less likely to get his back up, I know it would mine if I were the poster being harangued.<p>Colin

[jollyrgr]

Thanks to all that replied. And Edd, I understood you.<p>Some clarifications and answers.<p>I can make an LED flash with a 555 just fine and know how to adjust duty cycle, use driver transistors, and bias a diode for steady state current. And I know that you can overdrive or exceed the current limit for short bursts without killing the LED. My junkbox contains a number of high brightness red LEDs that will probably let their smoke out when I actually start testing this circuit. <p>The flashing device I saw, though, was very unique in the fact that it did not look like a flashing LED like the normal ones I've played around with in the past 20+ years. Maybe it's because they are white and super bright but they did not even look like LEDs from across the room. It was only when I got right near this vendor did I notice that the device was made using white LEDs. The flasher was home made but the vendor claimed he did not make it himself. Searching the Internet has led me to guess that it might be or at least based on the Ramsey LEDS1 kit.<p>Edd did bring up an interesting point about increasing the duty cycle higher and higher. A typical strobe light for motion stopping has a burst on rate of about 1mS. But all my thinking in this direction just seems too illogical, hence why I posted the question. It would seem that if an LED had a steady state current typical of 20mA at a 1mS burst it could handle 2 Amps of current. This could be achieved by having a large capacitor in parallel with the battery.

But logic keeps pulling me back saying that there is no way that an LED could handle two amps of current repeatedly even for short bursts.<p>So I'll ask this....Which line of thinking is correct? An LED can handle two amps for one mS. Or two amps through an LED for one millisecond will vaporize the little gold wire the first time this happens.<p>[ October 31, 2003: Message edited by: Jolly Roger ]</p>

[Chris Smith]

In simple English, the shorter your pulse time, the more amperage you can pump through the led with out burning it out. This was the first method used to make pulsed lasers out of glorified LEDs, some twenty plus years back. <p>They would dump up to 80 amps or more into a mere LED, [slightly modified] at rates of nano seconds and let the laser recover and cool down before repeating the trick.<p> This actually comes to the equivalent of one second pulse on time, followed by days of OFF time.<p> BUT, in actual fact, the time on was 10 to 200 nano seconds, followed by 980,000 nano seconds off time, repeated thousands of time per single second. <p>You can buy a strobe unit from All Electronics for Six bucks, [called a flasher with solar panels] and replace the green unit with a white one, or build your own using very fast switching techniques and high voltages, as long as the on time including rise time is very short. <p>My first LED driver drove a led [20 milli amp led ] at one amp, 12 volts, at 50,000 cycles per second, and out shone a four battery mag light constantly with light that would blind the naked Eye. <p>A strobe unit, one pulse per second, can be driven at this ratio in theory and much higher. <p>If 100 milli amps is the normal rate, constant, then a 1 billionth of a second pulse can equal 1 billion times 100 milli amps, or 1/10 of 1 billion amps. That’s 100 million amps! <p>Pure theory is here, and switching times are short of impossible, but reduce it to 5% duty cycle, and 500 million nano seconds and you come up with a realistic number of 500 nano seconds of on time, at 10,000 amps still???? <p>I think the math is right, as it must add up to the original, divided by time which is a multiplier.<p> But any way you get the point, the led can and will run at several hundred volts, delivering mega amps, for very short times, and still not burn out, if you do it right.<p>[ November 01, 2003: Message edited by: Chris Smith ]</p>

[Tommy volts]

Jolly Roger<p>Did you build your strobe yet? If not, you may want to check out this kit.White LED Strobe Light Kit

[toejam]

This brings up a thought. I wonder what would
happen if one were to hook up a
led in series with a strobe tube and fire it?

[Chris Smith]

Toast, one flash and its over.<p>The cap will dump way too much amperage over way too long of time until the LED is burnt. <p>Right Idea, but you have to shut off the 400 plus volts about 200 nano seconds after you start it up.<p> Then you can spot the led flash at about 20 miles!

[toejam]

Chris, is there a way to switch stuff that fast that be done on a limited budget?

[N9AOK]

I tried this with UV LED's this year for halloween and had some success. Although, its hard to judge brightness due to the light being UV instead of WHITE I was able to use a duty cycle of about 10 to 20% with an on current of about 100ma. These have run for weeks without fail (yet). <p>I have run more current through them (accidentally) but its hard to judge the light output beyond about 100ma. I dont think they got much brighter in my opinion.<p>LED's are very directional so with 5 UV-LED's I got about a 5' dia spot on a wall about 12 ft away. In a completly dark room everything in front of me that was UV reactive lit up pretty well. <p>I did buy some HONG-KONG white High output led's. Why not buy from hong kong? anyway, I haven't done much work with them yet but putting them circuit where my UV-led's were they were mega bright!! (technical term)..sorry, I wish I could be more specific but I need something to pick up light so I can evaluate light output on my scope. Anyone have any suggestions..I tried a variable resistor from a motion light (night day sensor) but had poor results. <p>Todd Snyder