NanoLight LED Bulb - 4000K at 133 lumens per watt

[SemiMan]

I know it is a poor image but here is a frame containing electronics from the back of the circuit board.

[/url][/IMG]

Some quick thoughts:

- Hope it works as "advertised", but I am skeptical, at least from a practical standpoint

- Quoting 133 lumens/watt AND saying it has been tested at independent labs, but not putting a copy of that test report in the Kickstarter makes me a bit suspect. I will believe it a lot more if signed by a recognized lab. Having a display show 1600 lumens is just about meaningless unless the test was run by someone who knows how to calibrate and operate the equipment. That is a generic spectroradiometer. It would need to be calibrated with a standard source in the integrating sphere. Was it done?

- 4000K may sound great, and I do like 4000K, but that is likely the lowest color temp they were able to get high lumens/watt LEDs at. Of course, those high lumens/watt bulbs have poor CRI. Most of the ones at 4000K with high lumens/watt have poor CRI, 70 or less. That is a fine light for outdoor lighting (what it was designed for), but not so great inside. Think those nasty CFLs that many people do not like.

- Though they claim it runs cool, we are still talking say 8 watts dissipated on a fairly small surface area. I would expect that to be warm, not hot, but not cool either. Interested to know what the coating is. You can increase emissivity with coatings to a point but that mainly helps when the surface is warmer. You still need area.

- Given they say the LEDs are Korean, and they look like Seoul Semi .... what do you think? I am not sure Seoul even has anything that is this efficiency? Remember that is 133 lumens/watt at temp, with AC losses. They could have a real hard time with color consistency bulb to bulb. This is one of the limitations on ultra efficient LED products today. Sure you can buy the brightest bin, but good luck getting it reliably at the same color co-ordinates ... even when willing to pay more.

- The 12W version looks like the LEDs are behind the board. How are they transferring heat to the board?

- The 10W version looks like the LEDs are sitting on top of the board? I wonder if that will pass safety certification? There is no discussion of that and no isolation in the circuit. Even the 12W version could be iffy in that measure.

- The low PF means no EnergyStar. There are so many off the shelf driver designs that dim, this should be an easy upgrade and some are pushing high 80s - 90% in this power range. Would be nice to see an upgrade.

- The assembly with the folding boards is pretty cool. It will need to be sturdy to ensure it can't easily fall apart if someone screws it in/out while turned on. That goes back to the safety/approval thing. Even if it was spray encapsulated, I still think it is interestingly viable but that would kill the heat dissipation.

- Multiple point sources, but not a lot of them, tend to create unfavourable shadows at close distances. Likely not an issue for the 12W, but could be for the 10.

- The design occured in North America, but the testing that was shown was at 215V ... so likely in a lab in Asia? ...


I am tempted to get one just to try it out, but have asked some questions on the Kickstarter first.

Semiman

 

[SemiMan]

- Also just noticed that the lumens on the meter were steadily decreasing even during the short duration of the test, hence the bulb had not thermally stabalized

- Power seemed to bounce around when would expect to be dead steady ... or at least monotonically going in one direction

 

[brickbat]

...
- The assembly with the folding boards is pretty cool...

It is. But when you mull it over, it seems that's the only thing that's the tiniest bit unique or novel about this lamp. I mean, other than that, what have we got here? A handful of off-the-shelf LEDs of unspecified CRI, and a non dimming power supply. Meh.

And I agree with your concern about compliance testing, IME, that alone could eat up their entire $50k budget from kickstarter...

 

[LEDninja]

There was a thread on a Toshiba E-core bulb a while ago.
http://www.candlepowerforums.com/vb...7-Toshiba-E-CORE-LED-Bulb-Can-t-Figure-it-out!

In Japan 8.7 watts and outputs 810 lumens. That is 93 lm/W.
in the USA 7.8 watts and outputs only 350 lumens. That is only 44.9 lm/W.

The reply in post #12 is quite interesting:
The model with high efficacy only sold in Japanese market where there is no safety codes required, maybe soon in the future but not now. The models sold in US market must pass the safety codes of FCC or UL. In order to pass the safety requirement, there are some components in the model for Japanese market must be replaced but that makes power efficacy drop. That [email protected] model has EMS issue and its power factor is only around 0.6. I am a LED light bulb designer. We have the same problem. In order to pass the safety codes like FCC or CE (for European market), we have to push the power consumption to 10 watts to generate 800 lm but it passes the FCC and CE. This is why the efficacy is so different in these two models.

NanoLightmight be able to get away with selling a few bulbs online but to do a major distribution via hardware stores etc. they will need UL or equivalent (CSA, ETL). They could run into the same issues as Toshiba resulting in a drop of efficiency to half of what they now claim.
There is a rule that requires people protected from touching the electrical parts. Not sure if using the circuit board as the case will pass this requirement.

 

[SemiMan]

An "LED light bulb designer" in Hamilton, Canada .... CRS? ....

FCC is compliance testing for RF emissions/susceptibility, it is not a safety test.

I can't see safety approvals, i.e. UL, CSA, ETL having much impact on efficiency, at least not more than a few percent, 10 percent max. All that impacts is clearance distance, isolation, etc. Really it should have little if any. FCC/DOC could add another few percent as you will need to deal with conducted emissions.

EnergyStar enforces power factor and yes that impacts peak efficiency.

Highly compatible dimming can also impact efficiency depending on the implementation, but generally that is impacting efficiency when dimmed, not on full.

That said, a design that takes into account all these design factors is going to have maybe 10% more losses than just a simple high voltage buck regulator.

Again, there are reference designs from several IC companies for power supplies for dimmable bulbs that would pass safety and compliance testing that are >80% and some a hitting high 80s now. That is a 10-20% TOTAL drop in efficiency.

Most of the issues with bulb efficiency we see, at least for 2700-3000K bulbs is:

- They are 2700 to 3000K and do not use red leds
- They are > 80 CRI and coupled with above means wasted spectrum (not adding much to lumens)
- Consistency ... i.e. can't use the highest bins

Take a "realistic" LED from 12 months ago shipping in high volume (time to do Energystar testing on the bulb). Now look at a realistic high volume, realistic cost 2700K-3000K bin that was 80 CRI minimum. Now look at the light efficiency at 85C keeping in mind the drive current is unlikely 350mA, but likely higher (relative, maybe you use small LEDs). That LED was probably 85-95 lumens/watt at temperature. If you start with 95 lumens/watt and take off 20% for electrical losses and 10% for optical losses (they multiply), then you get 68 lumens/watt .... or about 800 lumens from 12 watts. Sound familiar?

Sure we can buy super efficient, lower CRI cool white LEDs in high color temps at a premium price which is quite suitable for a low volume flashlight, but when you are building millions of bulbs, you have different design and purchasing restrictions. Safety and compliance testing compatibility will cause a bit of a hit, but realistically, not


Nanolights efficiency comes from:
- No optical losses (bare leds)
- Assumed low electrical losses (still likely 10%)
- Highest binned, cool, low CRI LEDS

That is not practical in many markets for sale, though if you are paying $0.30+ for electricity because it comes from diesel, you are less concerned about light quality and more about efficiency. You may want dimming though.


In terms of cost of compliance testing ... say $7,000 for North American safety testing, another $5000 to tack on Europe. $10K for emissions testing. $3-4K for Energystar. They can cover a large market for $20K ... assuming they pass. More passes = more cost.

Semiman

 

[Optical Inferno]

An "LED light bulb designer" in Hamilton, Canada .... CRS? ....
Semiman

CRS is in Welland Ontario but close to Hamilton I guess...

Also, notice the amount of LEDs. 30 from my count. They are likely under driving them, thus boosting the lm/W from each LED and reducing the thermal issues (but not eliminating).

1200lm from 30 LEDs = 40lm/LED

I'm leaning towards there likely being only about 100-150mA going through say a 1W'ish LED like an Edison Edixeon or Seoul LED.

Anyone else think this??

 

[SemiMan]

I think that is a reasonable assumption for the 12W. The 10W looks to have closer to 11 LEDs.

I agree Seoul or Edison.

 

[DaMeatMan]

This thing looks very much "home made" and i wouldn't spend more then a buck or two on something like this. Crappy "DIY" looking designs like this really don't help much in terms of bringing LED tech into the homes of the average Joe. To put it simply, think of the electric car. Now what are the first couple of images that pop into your mind? If you said something like "goofy looking", "ugly", "dorky", "sissy mobile" etc.. then you get what i mean. Great technology that is hindered by the general publics perception of the technology. Flood the market with junk like this, and the general publics perception of LED lighting quickly becomes "expensive", "weird looking", "ugly", and they hold onto the older technologies that they know and love despite the obvious disadvantages and lack of efficiency.

We need companies to spend more time thinking about overall design, elegance, things of that nature. At least at that point, you end up with a product that costs more, but at least the damm thing looks like it should cost more. This leads to faster adoption of the tech, provides the consumer with something "expensive or attractive looking", and becomes trendy and hip. Think iPhone... despite the fact that i hate Apple and everything they stand for, those folks are hardware design geniouses.. or at least they were when Steve was around.

What do you guy's think?

 

[jtr1962]

I'm sure they'll have some kind of plastic globe diffuser on the finished product if for no other reason than to keep the LEDs from being damaged. As for electric cars, the ones which have a serious following like the EV1 or the various Tesla models hardly look goofy or ugly, at least to my eyes. If anything, I think products where form follows function look and work well. Something designed to slip through the air at high speeds like a car should look and be streamlined, not boxy like many of the cars we see these days. Appearance is fine to a point, but when something is designed highly suboptimally just for appearance sake, then perhaps the public's perception of what that something should look like is really what needs to be changed. Nothing wrong with polishing the rough edges if you don't compromise function much. I'm sure they could keep the basic concept intact here while improving a bit on appearances. That said, it's only a matter of time before we have LEDs with efficiencies of 150 lm/W or more at high power levels. When that happens, everyone will be shrinking the size of their heatsinks. The NanoLight might fill the ultrahigh efficiency niche for a few years, but in the end increasing LED efficiency will make the concept obsolete.

 

[LEDAdd1ct]

but in the end increasing LED efficiency will make the concept obsolete.

It (increasing efficiency) will also, hopefully, put LEDs in every home, at dirt cheap prices.

I pine for the day when we can get $5.00, 1600 lumen LED bulbs with 95+ CRI that draw ten watts at any CCT you please.

 

[LEDninja]

I'm sure they'll have some kind of plastic globe diffuser on the finished product if for no other reason than to keep the LEDs from being damaged.

The NanoLight still needs to get rid of 6 watts of heat. A plastic globe diffuser would trap that heat. Unless they got holes all over the diffuser.

The NanoLight might fill the ultrahigh efficiency niche for a few years, but in the end increasing LED efficiency will make the concept obsolete.

I thought increasing LED efficiency will help NanoLight. When traditional LED bulbs go from 94 lm/w (L-prize) to 133 lm/w, the NanoLight would use the same improvement in LED efficiency to go from 133 lm/w to 200 lm/w.

 

[LEDAdd1ct]

This is why I cannot justify purchasing and retaining LED bulbs right now.

There is too much headroom for LEDs to improve in efficiency.

As fun as they are to play with, in a few years what seems "special" now at 133 lumens per watt will be mainstream,
and the LED bulbs for sale now will be power hungry dinosaurs.

 

[jtr1962]

The NanoLight still needs to get rid of 6 watts of heat. A plastic globe diffuser would trap that heat. Unless they got holes all over the diffuser.

They could use remote phosphor which would put about 2 of those 6 watts of heat onto the globe.

I thought increasing LED efficiency will help NanoLight. When traditional LED bulbs go from 94 lm/w (L-prize) to 133 lm/w, the NanoLight would use the same improvement in LED efficiency to go from 133 lm/w to 200 lm/w.

Not necessarily. I've noticed that as LEDs get more efficient the efficiency increase underdriving them is less in terms of percentage. Once "regular" LED bulbs are at 133 lm/W something like the Nanolight might only get you 150 or 160 lm/W. Also, at some point heat generation of LEDs will reach the stage where the heatsinking of the socket is adequate. This will probably occur when we get around 200 lm/W.

 

[idleprocess]

This is why I cannot justify purchasing and retaining LED bulbs right now.

There is too much headroom for LEDs to improve in efficiency.

It depends on what you're replacing, how you use it, and where it is in its life cycle. An oft-used incandescent would be an ideal candidate - the ROI will be swift in terms of your electrical bill. A mid-life CFL or recent-gen LED bulb is a poor candidate. A rarely-used incandescent is probably the worst candidate.

Just because something better comes out net week/month/year that's better doesn't invalidate the expense - it's a light bulb, not some $1 million piece of capital equipment that you were depending on remaining bleeding-edge for 2 years to sustain a competitive advantage.

 

[LEDAdd1ct]

What you say is, of course, true.

75 Watt Incandescent, Four Hours a Day

Initial Outlay = Zero (already own it)

$12.01 per year at four hours per day


Switch 75 Watt LED Bulb, Four Hours a Day

Initial Outlay = $50.00

$3.20 per year at four hours per day

Payback Time

Year One

Incandescent Costs $12.01

Switch 75 Costs $3.20

Amount Saved = $12.01 - $3.20 = $8.81

Year Four

Amount Saved = $35.24

Year Six

Amount Saved = $52.86

Assuming I use it more than four hours a day some days, I would say the break even point is around five years.

To replace the three most commonly used 100 watt incandescents with Switch 100 bulbs would cost $66.00 * 3 = $200.00 with tax

That's a lot of money to invest in a technology right now that is undergoing rapid change.

It looks like the Switch100, at 21 watts power consumption and $66.00 initial outlay will also take about five years to pay off.

Most of these bulbs carry five year warranties.

I hope what I'm writing makes sense.

I think this Nanolight LED Bulb has a real chance of changing things and speeding adoption.

However, I am with AnAppleSnail on this: Edison sockets need to go. They are holding things back. We are trying to fit a round peg into a square hole instead of creating novel solutions designed from the ground up to work with LEDs.

 

[idleprocess]

A Switch bulb isn't the best comparison since they're priced like the LED bulbs of 3+ years ago. I regularly see Philips and Sylvania units for around $20; the various decent-quality store-exclusive brands go for $20 or less with similar - but not not as good - warranties.

Agree that the Edison socket is a huge limitation. Perhaps the Zhaga standard will result in some nifty interchangeable modular lights that work well and are affordable ... but I'm not optimistic. The average homeowner is allergic to electrical work, whether they get up on a ladder and wire-nut it themselves in 15-30 minutes or cough up for a professional to do it, so it's going to be difficult to remove the Edison socket in residential settings.

 

[BM_lol]

133lm/W claim is impressive in claim.
I'll believe the results when they get DOE to confirm those bulb specs.

 

[LEDAdd1ct]

We have seen too many products over the years make a big news splash and go nowhere.

It is not my intention to play cynic, but, yes, I'll believe it when I can buy it, at advertised lumens/watt and at affordable prices.

 

[LEDninja]

I think the price would be reasonable.

There is no heatsink. There is no case (they are using the circuit board as the case). The driver seems very simple (otherwise the pf would be better). They are trying to find someone in China to put it together for them.

In addition to one of the few bulbs that replaces 100w incandescent, the design criteria seem to be low cost through simplicity.

 

[SemiMan]

FYI, CRI IS about 70, and then have not responded to any of my questions about safety certifications, independent lab tests, etc.