Philips 10.5w 800 lumen premature failure

I've purchased about a half dozen of these bulbs and three of them have been running in my bathroom mirror fixture for a few months. Recently one of them started flickering intermittently. Sometimes it would go for an hour with no flicker, other times it would flicker every minute or two. The bulb doesn't respond to tapping or shaking so I think this is not a cold solder joint but a component failure. Intermittent wiring or socket contacts were also ruled out

I've started to disassemble the bulb and I thought you'd like to see a couple of photos.

Removing the cover reveals 38 SMT LEDs on a metal heatsink plate. All of the LEDs are numbered and it has the phrase "entry range 38 up" printed on the board. Underneath the LED board is a press-fit heatsink with yellow heatsink compound on it. I'm trying to get past this so I can inspect the driver electronics.

Very disappointing to hear,especially with what must be really limited hours of actual usage,on a respected brand name bulb.

Electronics fail randomly after 1hr and 100000hrs ... It happens .. Not really anything to be dissapointed with.

It probably had about 40-50 hours of runtime, running base up in an unenclosed fixture. I'm not whining about it - I realize that this is still a relatively new field for LEDs and manufacturers are still finding their way.

LEDs in bulbs are somewhat new, but electronics are not new at all. Power supply technology is very well understood.

Electronics fail randomly for a very wide range of reasons. You cannot predict when it will happen, but you can predict the failure rate. That failure may happen after 1 hour, 1000 hours, or 1000000 hours, though in a case of an LED bulb, it will likely have failed already for non-random reasons.

Failure rate is also directly tied to temperature and increase exponentially as the temperature increases.

Unfortunately you got the bulb that failed and it happen to fail when it did ... and could have just as easily ran 50,000 hours.

When you have a large installation, even with the most reliable designs (redundancy aside) you will have regular random failures ... likely several or more per year.

Semiman

I'm working on tearing the thing apart and so far I've found that the entire bottom half of the lamp is metal with a plastic coating. I also looked at the traces on the LED board and found that they're all wired in series with no resistors on the board. So assuming 3.2v x38 LEDs = 121.6 volts feeding the LED string.

Really makes one wonder whether some of these LED bulbs will really last as long as they claim they do. Especially the lower cost ones. Better hold on to the receipt and warranty if you have one.

I think if you stick with name brand products at least at the consumer level that you will be safe. Yes there will be failures like anything else and yes I would keep the receipt. That said, I can't remember even the last time I had a CFL fail. I stick to the better brands. Costs more for the trip to the store than to stick to good quality brands.

Semiman

I've torn the lamp all the way down and added a few more photos to the set. This lamp is built like a tank and the pcb is entirely enclosed in a hefty aluminum housing. A few findings:

1) The power supply does not have an isolation transformer between the mains and the load. Not that it matters - this is a light bulb, not a radio.
2) As I was picking out the rubbery potting compound, the small yellow choke at the wide end of the pcb simply fell out. Apparently cold solder joints.
3) The neutral wire on the supply side of the pcb had some solder on it but it looked like it wasn't soldered to anything inside the base. Maybe it's press fit?
4) I re-soldered the choke back onto the pcb and connected it to ac power but got nothing at the output. It's possible that I've damaged something during the tear-down but I don't see any damage.

Check the PCB traces around the perimeter of the board to make sure you didn't cut through them by accident.

PhotonWrangler said:

.....
4) I re-soldered the choke back onto the pcb and connected it to ac power but got nothing at the output. It's possible that I've damaged something during the tear-down but I don't see any damage.

Click to expand...

You did well to get it de-potted to this degree. It's very easy to break components (especially glass diodes) or crack leads when de-potting.

Thanks for the tear-down report and photos! It's always fun to see what the manufacturers are up to.

These things are pretty cheap, and show some ficker even when brand new. It doesn't really surprise me that they get worse as the capacitors age. Anyone got specs on that Samxon capacitor?

The electrolytic capacitor is probably the component most vulnerable to degradation/failure over the long life of the LED lamp. However, if the lamp failed sooner, it was probably something else. It could have been an overheating transistor (resulting from a mistake in assembly, or overall design that did not adequately consider overheating of the circuitry). If the heat is adequately dissipated, and it is fed with a stabilized current supply, the LED chip itself should last nearly forever. It is the circuitry in the power supply that is likely to fail before then.

Arilou said:

These things are pretty cheap, and show some ficker even when brand new. It doesn't really surprise me that they get worse as the capacitors age. Anyone got specs on that Samxon capacitor?

Click to expand...

Flicker is likely far more a consequence of architecture versus the output capacitor though the output capacitor will make it worse and may result in other power supply issues.

Semiman

The Samxon capacitor is 47uf, 160vdc and rated at 105 degrees centigrade. It does not show any physical signs of degradation. The choke does show a little bit of insulation oozing, indicating that it gets rather warm.

I wonder if there is any circuitry design that could prevent all flicker (giving a smooth current supply to the LEDs) without an electrolytic capacitor.
(and yes, I realize that most LED bulbs do have electrolytic capacitors and still flicker)

Anders Hoveland said:

I wonder if there is any circuitry design that could prevent all flicker (giving a smooth current supply to the LEDs) without an electrolytic capacitor.
(and yes, I realize that most LED bulbs do have electrolytic capacitors and still flicker)

Click to expand...

If there is enough room - the advantage of electrolytic capacitors is their capacitance per volume.

Anders Hoveland said:

I wonder if there is any circuitry design that could prevent all flicker (giving a smooth current supply to the LEDs) without an electrolytic capacitor.
(and yes, I realize that most LED bulbs do have electrolytic capacitors and still flicker)

Click to expand...

I replaced most of these Philips 3000K bulbs with EagleLight and Aluratek ones. These appear to contain switching power supplies which can compensate for the ripple and give a smooth current supply to the LEDs. I haven't torn them apart to see what kind of capacitors they use.

Anders Hoveland said:

I wonder if there is any circuitry design that could prevent all flicker (giving a smooth current supply to the LEDs) without an electrolytic capacitor.
(and yes, I realize that most LED bulbs do have electrolytic capacitors and still flicker)

Click to expand...

Yes, but it all comes down to practicality and that is what electrolytic capacitors have going for them. Lots of capacitance in a small space at a low cost. That is not easily replicated in other technologies today. If space or cost is not a concern, you could easily eliminate the electrolytic capacitors.

Semiman

SemiMan said:

Yes, but it all comes down to practicality and that is what electrolytic capacitors have going for them. Lots of capacitance in a small space at a low cost. That is not easily replicated in other technologies today. If space or cost is not a concern, you could easily eliminate the electrolytic capacitors.

Semiman

Click to expand...

with the caveat that I like to tinker with hardware... I'm inclined to think that you could get rid of electrolytics, but you'd have to install some other form of energy storage in order to keep the LEDs powered when the AC waveform is near zero volts. How does a 5 kg inductor sound? (I just pulled that number out of my backside, but it might not be far from the truth). Supercaps might be smaller than electrolytics, but they are expensive, and I'm not sure that they are much more reliable.
Maybe just keep some deep-cycle lead-acid batteries in the house and use them to power your lighting? Perhaps just some low-dropout current regulators to control current?
Or convince the region or country to switch from AC power to DC power?

None of these are small or cheap solutions. I suspect that the better solution is to use better electrolytics and apply less stress to them. It will make the power supply a bit larger and a bit more expensive, but will likely be more cost effective than a short-lived bulb.