Filament LED

[night.hoodie]

What's new on those filament LED bulbs ? Any feedback on their life span ?

http://greenbuildingadvisor.com/blog...filament-bulbs
via /.

[Anders Hoveland]

I wonder if these LED filament designs can really achieve the high enough light outputs necessary to replace a 100 Watt incandescent bulb, for example.
It is more difficult to dissipate heat directly to air without some sort of heat sink.

My guess is these filament bulbs are mostly decorative.

[angerdan]

One regular filament-stick deliver around 105lm. There are currently lamps with 12 sticks and 800lm.
E27 B22 4.6w 8.2w LED Filament Lamps Bright lights High Power SMD bulbs | eBay

[night.hoodie]

I didn't RTFA, but the /. summary claims something about these things not needing heatsinks. So there should be some interest in these forums for not only how or why it is achieving the quality of light that is claimed, but also why its not throwing off more heat. /. has gone down hill, and its rarer to see some solid dialogue in the comments, someone pointing out the tell tale sighs of BS or perpetual motion machines, or perfect efficiency. So as soon as I saw it, searched here, found this thread, and posted link. I want to know what sorcery this is, without having to do any real work. I am qualified to request and expect something for nothing because I am American. TIA.

[PhotonWrangler]

40 watt 4-strip LED "filament" bulb, thermal images. Ambient temp was 72 degrees, bulb operating base-up in an open fixture -

After 30 seconds the temperature of the hottest spot was 84 degrees.

LED_filament_30seconds_84_degrees

After about three minutes, the temperature stabilized at 93 degrees -
LED_filament_5mins_93degrees

[CoveAxe]

My guess is that the inside of the bulb is pressurized with air or nitrogen. This will conduct heat much better to the outside surface, which has a much higher heatsink area than normal LED bulbs.

The lack of sophisticated electronics on board also likely helps. If I remember right, the power factor on these bulbs are just horrible (like around 0.6 or lower). You can get rid of a lot of heat-generating components when you don't care about that kind of stuff.

[tino_ale]

Very interesting. I am certainly not an expert in thermal imaging but it looks like the heat is rather well spread in the bulb !

[SemiMan]

My guess is that the inside of the bulb is pressurized with air or nitrogen. This will conduct heat much better to the outside surface, which has a much higher heatsink area than normal LED bulbs.

The lack of sophisticated electronics on board also likely helps. If I remember right, the power factor on these bulbs are just horrible (like around 0.6 or lower). You can get rid of a lot of heat-generating components when you don't care about that kind of stuff.

Air only transmits heat if it moves. I do imagine that the design of the LEDs could create a bit of a chimney effect. I wonder if we will see that the LEDs at the top will degrade faster.

[SemiMan]

40 watt 4-strip LED "filament" bulb, thermal images. Ambient temp was 72 degrees, bulb operating base-up in an open fixture -

After 30 seconds the temperature of the hottest spot was 84 degrees.

LED_filament_30seconds_84_degrees

After about three minutes, the temperature stabilized at 93 degrees -
LED_filament_5mins_93degrees

If I am not mistaken, standard glass, as what would likely be in these bulbs, would not transmit any of the long wavelengths that would be representative of the filament itself, i.e. you cannot image the filament from outside the bulb in the IR. Any reading would likely be highly erroneous, so what you are doing is imaging the glass which due to the surface area would stabilize at a low temperature, but that is not at all indicative of the temperature of the filament inside which is insulated by a somewhat non moving insulating medium.

[PhotonWrangler]

If I am not mistaken, standard glass, as what would likely be in these bulbs, would not transmit any of the long wavelengths that would be representative of the filament itself, i.e. you cannot image the filament from outside the bulb in the IR. Any reading would likely be highly erroneous, so what you are doing is imaging the glass which due to the surface area would stabilize at a low temperature, but that is not at all indicative of the temperature of the filament inside which is insulated by a somewhat non moving insulating medium.

This did occur to me. The glass will block some of the longwave IR energy, however it will also gradually warm up to a temperature that gives us at least a comparative reference. I've tested conventional LED bulbs where the temperature of the hottest portion was usually the electronics in the base, which would reach into the 130s. Granted the base is designed to conduct heat to the outside, so the thermal path from the semiconductors is much better than the path from the filament LED to the glass wall.

Unfortunately I'm not willing to break the bulb open to image the LED filaments directly. Maybe ElectronUpdate has already done this? He has the same thermal imager that I do.

[brickbat]

...so what you are doing is imaging the glass ...

Agreed. Further, unless you've compensated for the emissivity of the glass (which is less than 1), the readings are low.

[night.hoodie]

Well, I, for one, don't need to hide my embarassingly childish amount of envy behind valid scientific criticism, and I'm not afraid or ashamed nor too proud to say... damn, nice thermal imager you have there. Thanks for posting the awesome thermal images. Does it work on cats?

[Steve K]

Air only transmits heat if it moves. I do imagine that the design of the LEDs could create a bit of a chimney effect. I wonder if we will see that the LEDs at the top will degrade faster.

apologies for being pedantic, but my guess is that it is more accurate to say "air only transmits heat well if it moves". Conduction still works even if convection isn't being used, right?

[PhotonWrangler]

Well, I, for one, don't need to hide my embarassingly childish amount of envy behind valid scientific criticism, and I'm not afraid or ashamed nor too proud to say... damn, nice thermal imager you have there. Thanks for posting the awesome thermal images. Does it work on cats?

Lol, yes it works on cats and other mammals also, including the bipedal ones.

[PhotonWrangler]

apologies for being pedantic, but my guess is that it is more accurate to say "air only transmits heat well if it moves". Conduction still works even if convection isn't being used, right?

It also occurred to me that if I were to remove the glass globe and re-image the filaments directly, it would still be inaccurate as the filaments would be exposed to cooling air currents in the room. The most accurate test would be to drill a small hole in the base of the bulb and snake in a thermistor, as electronupdate has done on several of his reviews.

[SemiMan]

apologies for being pedantic, but my guess is that it is more accurate to say "air only transmits heat well if it moves". Conduction still works even if convection isn't being used, right?

Actually technically it is always moving .... even if a small bit. Non moving air is pretty terrible at heat transfer. Things get so hot that eventually you get movement and/or radiative becomes dominant.

Semiman

[brickbat]

...Air only transmits heat if it moves....

The interesting ( to me, anyway) part with these 'filament LED lamps' is I've read (earlier in this thread) these lamps are helium-filled. Helium has a different heat transfer ability than air, and I recall it being a lot better.

If true, it's kinda everything good about Switch's design, without all the bad associated with a liquid-filled bulb...

on edit: Looks like OSRAM/Sylvania holds a relevant patent:

http://www.google.com/patents/US2004...0#legal-events

Which might be part of the reason these 'filament lamps' seem to be curiously absent from the US market...

[PhotonWrangler]

The interesting ( to me, anyway) part with these 'filament LED lamps' is I've read (earlier in this thread) these lamps are helium-filled. Helium has a different heat transfer ability than air, and I recall it being a lot better.

Looks like OSRAM/Sylvania holds a relevant patent:

http://www.google.com/patents/US2004...0#legal-events

Interesting. If they're helium filled, it seems they would light up with a peach colored glow in the presence of a strong RF field. This could be a means of verifying the gas used.

[brickbat]

Sure, or suck out the fill gas and see if you talk like a chipmunk

[Steve K]

Interesting. If they're helium filled, it seems they would light up with a peach colored glow in the presence of a strong RF field. This could be a means of verifying the gas used.

very interesting and a bit tempting... did I mention that I work in the EMC field and spend a lot of time with high power RF amplifiers?
Any idea what field strength is required? Our lab generally tests around 140V/m.

oh.. we also have RF shielded cameras in the test chamber. It's better than standing around in a 140V/m field while waiting for a bulb to glow (or monitoring more business related things)

[degarb]

It occurs to me that a helium filled design, will likely make it more fragile and prone to factory defects. The choice of reliable brand may be important. Greenbuildingadvisor article alludes to a percentage of initial failures.

The pattent thing is interesting. I wonder if Big Brother will step in to identify and help negotiate a settlement, as to part of the national effort toward energy independence, since 75 lpw to 105 is nearly a 50% improvement on a fledgling technology.

[SemiMan]

It occurs to me that a helium filled design, will likely make it more fragile and prone to factory defects. The choice of reliable brand may be important. Greenbuildingadvisor article alludes to a percentage of initial failures.

The pattent thing is interesting. I wonder if Big Brother will step in to identify and help negotiate a settlement, as to part of the national effort toward energy independence, since 75 lpw to 105 is nearly a 50% improvement on a fledgling technology.

A good portion of that "efficiency" is from a design that has no power factor correction, is not dimmable, and has brutal headache inducing flicker. Making a 105 lumen/watt bulb today is pretty easy when you throw away all the typical parameters that define a good bulb.

Semiman

[brickbat]

What's your basis for the flicker condemnation? I thought electronupdate couldn't measure any flicker...

[SemiMan]

With the exception of one that had exceptionally bad power factor, everyone I have had my hands on had wicked flicker.

[brickbat]

For other readers not familiar with power factor, could you explain why they should care? (I'm talking about average residential users here...)

[CoveAxe]

For other readers not familiar with power factor, could you explain why they should care? (I'm talking about average residential users here...)

Well, just in case here's a simple explanation of what power factor is.

To put it simply: Low power factor means that you will have current surging back and forth over the transmission lines that aren't doing anything useful. Since transmission lines are resistive by nature, that extra current will be mostly lost as heat. So basically the power company will be generating power that will get wasted. This will increase everyone's bills.

Now if it's just one or two bulbs in the neighborhood with a poor power factor, that's not a big deal. Those won't even register. If every house in a city uses them though, that could create current surges large enough to start wasting a lot of energy. So while it's not really important from a residential perspective, it's very important from a power company or grid perspective.

I'm not familiar with the electric code or UL criteria, but I'm fairly sure that almost all consumer electronics and appliances must have a reasonable power factor to be approved. So things like your air conditioner, refrigerator, etc all have power factor correction so that the power factor is very close to 1. All of the normal retail LED/CFL bulbs I've seen for sale have a power factor of 0.9 or higher. I've only seen the filament bulbs with low power factor being sold on ebay or through small importers (i.e. not approved by any US entity).

[brickbat]

Yep - that's the basic principle alright. I'd just like to see it put in perspective with real numbers.

...A good portion of that "efficiency" is from a design that has no power factor correction...

And, if this is really true, it seems like we trade good PF for efficiency - so what's the point?

[idleprocess]

And, if this is really true, it seems like we trade good PF for efficiency - so what's the point?

Industrial customers are usually billed for low PF in addition to MWH consumed; as such it makes sense for them to balance the costs of better power factor equipment against their own PF correction gear at the utility connection point.

For residential customers, the utility has typically dealt with poor power factor by installing huge PF correction gear at the substation, however with the roll out of residential smart meters they now have the potential to bill residential customers for low PF. So soon enough it may become relevant to you as well.

[CoveAxe]

And, if this is really true, it seems like we trade good PF for efficiency - so what's the point?

Here's a really good article that gets into the meat of it.

You are also not really changing the efficiency that much for the power factor correction. You are trading off other aspects like flicker though.

[SemiMan]

When you only conduct at the top of the waveform you get low conduction losses and high efficiency ... And potentially wicked flicker.

Posted by really crappy Tapatalk app that is questionable wrt respect of personal data.