Low component count relaxation oscillator

Interesting. I suppose it's the thermal feedback that makes it oscillate--battery recovers, LED starts to conduct, LED forward voltage drops, current increases, battery droops, LED stops conducting, LED cools down, ....

Cheers

Phil Hobbs

It's not that uncommon for simple LEDs to flash, by themselves, at a few Hz. I suspect flakey wire bonds.

But probably not on three out of nine LEDs.

Just turned it on again, and only one flashes. After a few seconds it's two. Now it's back to three and one occasionally, but there's a clear mark/space ratio difference between the three.

Maybe it's been taken over by aliens, has mistaken me for Skybuck, and is trying to communicate.

Cheers

If so, do NOT let it get close to your gaming computer. It'll explode. Again.

:-)

Cheers, James Arthur

I've got an otherwise nice Lee Valley LED flashlight that has one LED that flashes at maybe 2Hz. (B) here:

I don't think it's a coincidence that it's the one in the center of the array.

There seems to be an actual LED flashing failure mode. I can't imagine what the physics could be.

It happened on my Mantis, which has a bunch of serial-parallel LEDs.

I see it on traffic lights, usually the green light. One section of LEDs dies, and occasionally blinks.

Ok, I took note the word "Cheap" and then "US: flashlight", does that mean we make junk?

The last time I checked, we don't make anything, anymore ;)

Jamie

Why the middle? Given Phil's thermal model, an led near the edge might have see more temperature change. (maybe you have a different idea?)

Led's in series would be nice, if you could make them fail as a short.

George H.

The LED chip is probably slowly arcing itself away after ESD damage. Some green, blue, and violet chips are crazy sensitive.

And no, you can't make a relaxation oscillator using a battery and a hot LED. The impedance of the battery is much too high for the 0.1V change per LED to matter. It can be done with an ultracapacitor but that's expensive.

Don't think so, when three of the nine LEDs flash they're in phase. I tested this by covering over all but these, then sweeping my eye across so as to see three sets of dashes persisting on my retina.

I don't really understand that.

Cheers

I wasn't sure if 'torch' was understood in the US, it's universal here.

A new Harley Davidson showroom has just opened on the outskirts of town here. I hired one in the 'States once, my ears are still ringing.

Cheers

Topical:

In response to other suggestions: Vf + Vbatt wouldn't be enough to cause oscillation, it would be a whole system thing for one -- and both effects being diffusion (thermal diffusion from die to package, chemical diffusion around the battery plates) will result in, at best, a real pole or pair, no complex pairs and certainly no RHP poles.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design

A thermal-analog-loop oscillation mode sounds unlikely to me. The ones that I've seen blink, they went sharp on and off, pretty much square wave.

I should have analyzed the ones on my Mantis that did that. I threw them away and replaced the entire silly array with a few Cree whites.

Blinking, very annoying:

original array:

What you heard was the tool bag bagging around that you need to carry with you, when owning a Harley. And let us not forget the needs of keeping the chiropractor in business, you'll need one at some point ;)

Jamie

They're connected together so one having minor shorts will cause the others to flash.

The LED's Vf may drop by as little as 0.1V when it's hot. That creates some positive feedback but it won't cause a simple battery + LED circuit to oscillate. Most batteries act as a constant current source when overloaded so nothing interesting will happen. Throw in a a $30 ultracap and it can oscillate from heating/discharge to cooling/charge cycles.

The LED blink failure thing seems to be opens, not shorts. I think.

LEDs are usually ohmic and PTC at higher currents, which is why people get away with an LED directly across a battery. The flashing that I've seen was really square wave, not soft like you would see from a thermal oscillation.

All nine are in parallel, only three flash.

I was thinking of some sort of electrochemical thing. On a partially discharged battery, the current taken by 9/9 LEDs causes the battery voltage to drop below that required for the three highest Vf LEDs, then the reduced current of 6/9 LEDs allows the battery voltage to recover.

There is some tenuous support for this in that when you first turn it on, only one flashes but another two are a bit dim, then after a few seconds, two then three start flashing.

Swapping the battery pack brings them all on again, although I think the three potential flashers are a bit dimmer than the rest.

Cheers

Any plated-through holes on that PCB?

-- Best regards, Spehro Pefhany Amazon link for AoE 3rd Edition:

Plated-through? Do you think I'm /made/ of money? This is cheaper than a cheap thing on International Cheapness Day. It's single-sided.

Cheers