LED lamps turning off and on independently

We have several mains voltage LED downlighters and a couple have taken to turning off after they've been on for a while and then turning back on after a couple of minutes.

It would seem that something is overheating perhaps? Is it likely to be the transformer or the lamp?

Reply to
Frank
Loading thread data ...

LED lamps and the SMPS in the base, really could use some airflow.

This is why it is not recommended to place them in sealed globes, such as the globe in my kitchen.

formatting link
While you could punch a hole in the top of a pot to allow the hot air to escape, that might be a violation of a fire code. I don't know the rules regarding modifications like that.

In any case, pots were designed for incandescents, and the ceramic and metal in incandescent bulbs can take lots of heat. Relatively speaking.

They make LEDs on silicon carbide substrates, to extend the temperature range, but the problem is, some of the environments are just too miserable for such incremental improvements to help a lot. The article above mentions the electrolytic capacitor failing, and sometimes when the lamp dies in only three months, it's the cap that fried, not the LEDs. The capacitor might only be rated for 2000 hours at that temperature.

In the SMPS controller chip, it is easy to add an overtemperature detector, as substrate diodes are easy to do in silicon. It could be purposely switching off, or, the SMPS oscillator could have stopped switching for some reason.

It you take the lamp out of the pot and operate it in open air on your table, I bet it will run all day long without winking out.

I have some lamps here, that are failing prematurely, and they *do* have airflow. And that's how you know, how "cheerful" LEDs have become.

Paul

Reply to
Paul

Paul snipped-for-privacy@needed.invalid wrote

Much more likely to be the 'transformer' which isnt actually a transformer.

None of my Philips Hues need that. Not even warm even when the room temp is 48C

Works fine with the best bulbs.

But a failure of those wouldnt see the light cycle.

Not aware of any that do that and its unlikely to have just started doing that now, in winter.

Reply to
Rod Speed

Thanks, but there's no pots involved. They're open to the void above the ceiling and below the roof which is why I'm not sure if it is overheating and, if it is, which component, lamp or transformer is the problem.

Reply to
Frank

How much power do the LEDs use, compared to the original lights ?

With a multimeter, check the voltage across the LEDs. In other words, what is the transformer producing right now. Is the output normal or abnormal ?

You can try substituting one legacy incandescent for one of the LEDs, and see if stability improves. (This would be an attempt to alter the power factor of the load presented to the transformer.)

Paul

Reply to
Paul

cooked capacitors as a first guess ...

Reply to
Andy Burns

They dont normally work intermittently like that.

Reply to
Rod Speed

You could try swapping them with ones that don't do it to see if the fault moves with the bulb. That should show what the issue is. How hot does the lamp or other components get?Normally LED lamps just die and stay dead if its their internal psu. Brian

Reply to
Brian Gaff

If an Led power supply gets hot its hardly being very efficient is it? Are these all the same maker etc? I remember also aperson telling me that the colour temp of some supposedly the same leds can be different, indeed variations exist inside one light. I've never chopped one to bits to see what is there, but you can bet they have cut out all unrequited parts and used the cheapest chips and Leeds they can get. Brian

Reply to
Brian Gaff

Could just be a dry joint of course or a faulty device where the wire loses connection inside the plastic. Many years ago a N1501 VCR randomly stopped. Several people failed to find it, but I used my brain and fond the problem almost immediately. Four silicon rectifiers in a bridge. Two oof which had the external wires only making contact by mere luck. Take them out and the wire fell out. I changed the lot and no more problems. They were Philips light blue bodied bullet shaped and the wires, in my view had been bent too close to the flat end of the body. In such a confined space as a lamp I could well imagine something similar. Brian

Reply to
Brian Gaff

Maybe the driver circuit in bulb is shutting down. What some manufactures have been including in their bulbs is temperature protection IC. The LED current is reduced with a rising temperature and possible shutting off if a threshold is exceeded.

formatting link
In that video BigClive is comparing two different lamps - one of which is over driving the LEDs

LEDs and their electronics do not like getting hot irrespective of the brand.

How hot a KED bulb gets often depends on the light fitting and many downlight fittings that originally took halogens are unsuited for some LED replacements.

Reply to
alan_m

If they're mains LEDS there won't be any transformers ...

Reply to
Andy Burns

Might be, in an SMPS. Not all are so crappy that they use capacitors for current limiting

Reply to
The Natural Philosopher

But not user-accessible like 12V lighting trafo within the ceiling void, likely to have to break the lamp open to see it ...

Reply to
Andy Burns

I had a bad batch of 12V led downlighters a few years ago which all started cycling on and off just as described. It turned out that the bond wires between the leds and the alumina substrate were not well bonded. As they heated up the bond opened up and the led turned off. As they cooled the connection was made again. These were "chip on board" leds with a series/parallel matrix of leds on a rectangular substrate, covered in phosphorescent gel. They were from TLC who replaced them all several times over under warranty until eventually they started working reliably. John

Reply to
John Walliker

The Philips Hues don't get hot by design.

Reply to
Rod Speed

Philips Hues

"A light which is off, consumes a standby wattage, the amount depends on the model, typically between 0.2 and 0.4 W A light which is on, consumes the standby wattage + % of brightness * (wattage – standby wattage) When a light is on, a brightness setting of 5% is the minimum used for the power consumption calculation The above also applies to the Hue bridge, with an estimated 50% average continuous power consumption

This is a basic estimate, without knowing the exact power consumption (which depends on a combination of factors like brightness, color, temperature, etc.) this is the closest estimate possible. Unfortunately, the Hue bulbs do not measure their own power consumption. "

"Smart LED bulbs consume very little power, especially when compared to incandescent bulbs. An incandescent bulb uses 60 watts to emit 800 lumens. However, a Philips Hue smart bulb delivers the same amount of light using just 9.5 watts." <===

I just went to the bulb cupboard, and pulled the cheesy Philips non-Hue. This is the spec on the side of the box.

60W 800 lumens 8.8W 15,000 hours

So in fact, there is nothing magical about a Hue.

If I need a 40W lamp, I screw in a 40W lamp. If i need a 60W lamp, I screw in a 60W lamp. If I need a 100W lamp, I screw in a 100W lamp.

All of the bulbs in the cupboard use SMPS. There are no low power capacitive-dropper bulbs, no Dubai bulbs. And the laws of physics have not been repealed. The bulbs all use electricity in proportion to the light output. There are no Sparkle Ponies in the room at the moment.

If you set your Hue to 100%, because you need the light to see, then your bulb uses 9.5W (according to the promo prompt on the Philips page) and mine uses 8.8W. Waste heat comes from the package, the SMPS gets warm inside due to the modern bulbs having less heat dissipating surface area.

And it's not just the "15,000 hour" bullshit rating of the bulbs. The bulbs so far, are blowing out a lot faster than the 15,000 hour rating implies. Which is... not good.

The bulbs now have a "switching cycle" rating, but even taking that into account, they're still blowing out too soon. These are not bulbs that are failing because "the LED output has dropped to 70% of the initial brilliance". These are things like SMPS failures. The bulbs are never in globes, the bulbs have convection air to use.

The best SMPS I know of, were made at work, with a computed reliability of 100 FITS. That's ten million hours (5x a hard drive). These are smaller devices with not very impressive power output. But they have the best FITS numbers I know of. Because someone asked the engineers in the power supply division to "make a good FITS product".

Larger power supplies weigh in at 3000 FITS, by comparison. Thirty times less reliable.

If we were to measure the field reliability of light bulb FITS, what would we find ? We would find "utter s**te". That is what we would find. So don't be telling me how wonderful light bulbs are. They're little better than incandescent now, once they removed the heatsinks.

Back when I was testing CFLs, and the SMPS in the base of the CFL would fail and make a stink, it was the same. Unreliable. Failing before their time. When the last CFL made a stain on the ceiling, that was the last straw. I still have CFLs I won't use, sitting in cupboard, because I know how they will leave this world, with a stink and a stain.

At least the LEDs blow, without a chemical halo. I guess that is progress.

Paul

Reply to
Paul

The LEDs are blue emitters, with a yellowish-tinted phosphor coating on top of the blue LED. This converts the light into some shade of white.

Some bulbs, have tossed in a couple red LEDs (without phosphors), to adjust the spectrum a bit.

You can spend more money on the phosphor composition (precious materials) to make more corrections to the spectrum.

The higher the colour temperature (like 9000K for light humans don't want), there is a thinner phosphor coating on those, and the "base lumens" is greater. The phosphor wastes some of the energy, so you get fewer lumens with a thick phosphor coat, but a more pleasing colour. The 9000K lamps might work on a bicycle, as an example of an application. You're not reading a book when riding a bicycle :-)

Philips made some remote-phosphor lamps. The room I'm sitting in, is lit by one of those. The LEDs in the lamp inside are blue (if you smashed the lamp envelope, you would be seeing blue). The plastic shell on the lamp, has the phosphor coating. This is to avoid the "phosphor smothering" problem. But remote-phosphor lamps have come and gone. They don't make those any more.

Paul

Reply to
Paul

Paul snipped-for-privacy@needed.invalid wrote

Irrelevant to whether it gets hot.

Duh.

Irrelevant to whether it gets hot.

Irrelevant to whether the BULB gets hot.

So what was the point in quoting this ?

Which explains why it doesnt get hot given the size if the bulbs.

No one said there was. What was being discussed was whether they GET HOT.

Clearly they don't if they only consume 9W and are big enough so that that doesnt run anything hot.

You quite sure you aint one of those rocket scientist fellas ?

You quite sure you aint one of those rocket scientist fellas ?

You quite sure you aint one of those rocket scientist fellas ?

That's not true of the E27 Philips Hue bulbs.

I haven't lost even a single one and most of them are on for hours daily given that I do multiphase sleeping which means that I spend quite a bit of time up and about at night.

Yours arent HUEs.

Same with mine.

<reams of irrelevancy to whether Philips Hues GET HOT flushed where it belongs>

My Philips Hues are MUCH more reliable than incandescents ever where.

Reply to
Rod Speed

probably

Yes, it's a failure pattern of both LEDs & their ballasts so who knows

Reply to
Animal

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.