I've noticed that various LEDs which are designed to be dimmed have a minimum brightness - if you dim them from bright, they go gradually dimmer and then turn off while still partly lit; if you increase the brightness from zero, they come on suddenly at partly lit, and then increase from there.
This applies to Philips Hue bulbs and to various makes of Christmas tree lights. Is there some inherent design problem with a minimum brightness below which the LED will not produce any light? Given that the LEDs produce high-frequency flashing light (you can see separate images of your finger if you move it quickly across the light) I'd have thought that the dimming is achieved by varying the mark:space ratio. Will LEDs not turn on if the M:S ratio is below a certain level? I could understand it if the LEDs were fed with DC from a constant-voltage/variable-current supply, but that would not produce the multiple images when you move your eyes, so evidently that's not how these lights dim the LEDs.
That seems to be a function of LEDs. I converted the interior lighting in the old car to LED. Nice to have lighting bright enough to read by without burning the lamp housings of the old 6 watt units. And I added a fade in, fade out device. Which just ramps up and ramps down the current flowing through basic LEDs. They fade out nicely - but come on with a jump.
The problem is with the control circuitry not the LED itself.
Making dimmable LEDs on mains is somewhat tricky since the most obvious design will simply draw ever more current to provide the required constant output if you modulate the incoming waveform by chopping part of it off. My friend with dimmable LEDs and officially compatible controllers seems to blow them up with monotonous regularity.
The work around to make is somewhat imperfect. The small load LEDs present can also cause problems for the controller as there is a strict minimum power that it is able to control reliably. I suspect that is why it snaps on at a particular brightness setting.
The requirement is for the control electronics to figure out what brightness you actually want and then they add some hysteresis so that it doesn't hunt. The result is that you have a dead band.
A bare modern high efficiency LED with 10uA flowing through it is bright enough to see by once you are dark adapted. Some designs of domestic mains LED bulbs can pick up enough capacitively coupled current to glow very faintly even when nominally off (particularly when on 2 way switches with a nice long run of cable between them).
Hysteresis perhaps? I don't know much about modern leds, but I was under the impression that the white ones actually contain more than one and the linearity of light is very poor at below certain currents, but if as you say its done by duty cycle, then maybe there is more to it than just non linearity, perhaps actual switch on voltage varies and although it could be a square wave its more likely to be a ramp. Brian
It's true that once LEDs used a few of different colours to get white. And this might still be the case with cheap ones.
Decent ones these days can give an approximation of daylight or warm white in a single LED. For ordinary domestic use.
Think if you want a decent continuous spectrum luminair - like for film etc work, you may well use a mixture of white and spot colour LEDs. Like found in operating theatres.
White LEDs are a blue LED with a phosphor coating. Blue light is energetic, and makes good "excitation" for the phosphor.
It's a single LED diode, as an electrical device. The Vf and If are coupled, so if the If went from
20mA to 1000mA, you would expect the Vf to go above the 2.5 volt value, to maybe 3V or 3.5V .
The linearity of light, is much better using the PWM method, than using the "resistive limiter" method. To make the "last step" in the PWM method seem less abrupt, you simply need to run the circuit with a faster clock. (Like on for 1 cycle, off for 9999 cycles, repeat at least 200 times a second, so 2MHz.)
Large LEDs can be switched at 10MHz or so. This was a relatively popular project to set up free space optical networks. They were switching LEDs at that speed.
formatting link
But if much smaller LEDs are used, they can be switched at 1GHz to 3GHz, for in-room communications. The 10MHz project is switching higher power devices.
The most current I've ever heard of, going through a LED, was around 17 amps. It was done, by sweat soldering a power LED to a copper anvil, so that the copper anvil could act as a heatsink while the LED was tested for a few seconds. The LED doesn't actually put out that much extra light when abused like that. If doesn't put out 17x as much light as it would at 1 ampere. Maybe the light is only 3x as great.
Many of the light bulbs you see, use array LEDs, which raises the forward voltage up to levels better suited to mains powering solutions. You can still PWM such things.
I would think, using a DC supply and your own extended-count PWM design, you could make the last vestige of light very close to zero. So the last step didn't seem as abrupt.
Modern power supplies, some of them "cheat" to remain stable. A coveted rating, would be for an ATX computer power supply, to run with no load on it at all. But inside the supply, sometimes there is a 5W to 10W load, which ensures stability when the open circuit test is done. So they don't really run at zero load. The SMPS PSU inside a LED lightbulb, might have similar issues, where its behavior when PWMing to extremes, uncovers a non-linear part of the control loop. Simply squelching the operation of the circuit, is "safer" than using a resistive load to bodge the thing to look like it runs down to zero.
You'd need Big Clive to tear one of these apart, to see what circuit choices they've made. I actually avoid dimmable bulbs, because in non-dimming applications, they fret a bit and "wobble". I want a nice light that puts out a constant level, all the time it is operating, with no signs of jitter in the output light level.
I've even had bulbs here exhibiting "coil noise". And those eventually, I got sick of that, and replaced them with a similar bulb not making that noise.
It can happen with some dimmers and lamp combination, it can even vary with one lamp or two on one dimmer. My present living room wall lights, which I have just swapped to higher wattage LED Lumilife lamps, gave [roblems with the existing dimmer, so I swapped the dimmer for a LAP auto selecting positive or negative switch. That works fine for the one with two lamps, but there is a single walllight too - That goes out every few seconds briefly, on the full setting.
I have a V-Pro dimmer running a single LED - 100 watt equivalent. With the V-Pro, you can programme the maximum and minimum setting. I've been able to set the minimum one OK - but not the maximum. The LED comes up to full brightness about 2/3rds of the knob travel. But if you put it to full, the light will go out and come on again at random. There are two other modules on the same plate that work as intended - but with a higher load. Not got round to swapping them, to see if it's a fault or 'feature'.
Mine start to show some light at about 10 degrees clockwise from off, just the barest glimmer if the room is completely dark. There is a one second delay from switching on, before the LED's light up, I guess that's the 'automatic polarity' working out what to offer the LED.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.