Energy Matters Forums

[zzsstt]

After the previous LED lights thread, I decided to do some testing, as it has been a while since I have done so.

Due to time constraints, I have not yet arranged any LED tube replacements (that will come) but I purchased some MR16 LED "globes". These were very well priced from a Australian eBay seller, and I had heard good things about them from another forum member.

On first unwrapping the gobes, they do look reasonable, in fact the design is quite pleasing. These particular units were marked as "warm white", which is nominally the same as the halogens they were to replace. As the first units were installed (looking very neat) and switched on, the initial impression was that they were really quite bright. But there is no way they are "warm white". I have a selection of light sources to compare with, and these LED's are very blue which makes a visual comparison quite difficult - they look bright because the bluer tinge makes everything look stark or harsh, giving the impression of brightness.

So, roll out the light meter. The 20W 35 degree halogens have been used intermittently (task lighting) in a kitchen for over a year, and measurements were taken without cleaning (i.e. coated in grease!). In this comparison the light meter reveals that the LED's are producing between 30%-50% less light. The light meter is calibrated to the CIE curve, with the highest sensitivity at the 550nm (green) area, and should not significantly bias in favour of either light.

So whilst the far bluer output of the LED's looks bright, they're not actually producing as much light. When working under the lights this is also apparent. It's very interesting to put an LED in the adjacent fitting to a halogen. From across the room the colour difference is very noticeable and the bluer LED light loooks impressive. But under the light the lower level of illumination is quite noticeable, and attempting fine work under the LED's demonstrates the lower levels of real light.

So, at this stage the conclusion is that the "warm white" LED's at 5W (I haven't checked, that is the figure stamped on the light) use about 75% less power than the 20W halogens, and produce a much bluer but lower intensity light. Individually they are not really a replacement for even the 20W energy saving halogen in a task lighting situation, but replacing 2 for 1 still gives a nominal 50% power saving and probably (I haven't tried it yet) enough light.

At about $14 each, they compare to the 20W Osrams at about $9 (variable). For task lighting, given a 2:1 ratio that's at least a $15 price penalty for the equivalent light output. At 2x5W, there is a 10W power saving, which at 23c/kwh would be 0.23c/hour, recovering the $15 in about 6500hours, or about 4.5 years given 4 hours a day of usage.

Unfortunately, having replaced 6 20W halogens with these LED's, the first one failed after about 2 minutes. The globes were on a track driven by a single (expensive, electronic) power supply,so the fact that only a single unit failed would indicate that the issue was with the light not the power supply. As yet I have not spoken to the suppliers, so I don't know what the outcome will be, but it is not very encouraging that a lamp requiring 4 hours/day for 4.5 years should fail after 2 minutes. Hopefully that was a fluke, and the rest will continue to work for many years.

Conclusions?

1/ I'd say they require at least 1.5:1 or preferably 2:1 as replacements for 20W IRC halogens.
2/ The light produced is far bluer than the "warm white" tag suggests
3/ As a result of the colour, they appear to produce far more light than they actually do
4/ For task lighting (short durations) they are unlikely ever to repay their cost

As most halogen illuminated rooms are massively over-lit, these units are probably capable of producing power savings when used to replace 50W halogens. However their lower output means that if the room was correctly lit by 50W'ers, they will simply not produce enough light if used on a 1:1 basis. There is still a question mark (at the very least) over their longevity, which is a concern given the capital investment and the relatively small power savings. Although they may save 50% of the power, it's 50% of a small (20W) usage.

Note that the Osram 20W IRC halogen is advertised as producing the same light as a standard 35W halogen, whilst the 35W version is equvalent to a 50W halogen. This extrapolates to the LEDs (at 30%-50% less light than the 20W IRC) being at most 50% the brightness of a 50W halogen.

[karlajensen]

That was a tough read my friend.

The LED that failed because of the power supply and the other ones will fail soon, its something to do with the switching frequency of a switchmode supply - dedicated LED drivers are required.

Alternative is GU10 fittings which obviously must have their own drivers built in and hence no problems.

I bought some 5W jobs -probably from the same bloke, does he claim as bright as the 9W ones?

My conclusion with the 9W lights were that in fact they used between 5-6W each so side by side with a 5W bulb the difference was close to indescernable -basically due to the slight difference in colour between the two, the 9W ones being yellower than the new "warm" globes.

-God forbid the "white" globes must be whiter than the sun!

I paid $80 for 6x this time round but the first lot of dimmable GU10's I got were more like $30 each.

My dilema now is lot 2 -the 5W jobs aren't dimmable but the 9W ones are and hence the question, am I better off with the dimmable ones in the living area or the kitchen where thay are now......
more tinkering, cant mix'n match as the non dimming ones hated the dimmers in the circuit even at full on the LED's flashed, 1s on 1s off so I disconnected the dimmers.

Living room in my case went from 300W of Halogens who's use was forbidden except in exceptional circumstances to 30W LED's whose light output is plenty and in comparison to the 20W fluro uplight we were using (which broke hence the downlights getting a treat ) far better.

So if you did the math on 4 hrs a day there, no doubt there would be some significant savings.
4hrsx 55W (5 W for the power supplies) x 6 (lamps) =1.32KWhrs
vs 4hrs x 5W x 6 (lamps) = 0.12KWhrs
difference per day = $0.2628 /day and thus 305 days to pay back which is obviously less than a year!

WIN WIN WIN -aside from the fact I never used the old ones

Now if I didn't use my normal sparkie mate who does things for me in exchange for beer I'm guessing the exercise would have cost three times as much and hence a 2-3 year payback for most people is reasonable

[Gordon-Loomberah]

Living room in my case went from 300W of Halogens who's use was forbidden except in exceptional circumstances

Special circumstances... like turning the living room into a sauna you'd need

[zzsstt]

The LED that failed because of the power supply and the other ones will fail soon, its something to do with the switching frequency of a switchmode supply - dedicated LED drivers are required.

I shall investigate this further, and probably invest in an LED driver for the lights. By the way, the supplier agreed instantly to replace the failed unit and made no mention of either power supplies or the chance of other units failing.

Alternative is GU10 fittings which obviously must have their own drivers built in and hence no problems.

This particular case is an MR16 track light system.

I bought some 5W jobs -probably from the same bloke, does he claim as bright as the 9W ones?

It is indeed the same seller.

My conclusion with the 9W lights were that in fact they used between 5-6W each so side by side with a 5W bulb the difference was close to indescernable -basically due to the slight difference in colour between the two, the 9W ones being yellower than the new "warm" globes.

-God forbid the "white" globes must be whiter than the sun!

Indeed!

So if you did the math on 4 hrs a day there, no doubt there would be some significant savings.
4hrsx 55W (5 W for the power supplies) x 6 (lamps) =1.32KWhrs
vs 4hrs x 5W x 6 (lamps) = 0.12KWhrs
difference per day = $0.2628 /day and thus 305 days to pay back which is obviously less than a year!

That is certainly the case, but in my case the room is adequately lit by a T5 fluoro, and the MR16's are simply task lighting (actually a track over the kitchen bench) and are only used during intricate food preparation. That limits their use to perhaps an hour or two at a time, and on many days they are not used at all, which in turn makes payback a very long term proposition!

HOWEVER. Having said all the above, tonight I worked under these LEDs for the first time. I must say that they are not in any way offensive, and they do produce enough light (I still have all my fingers!). The very "white" (bluer than "warm white" should be) light gives a weird "I'm on a stage" feeling by virtue of the starkness of the illumination, but it's not really problematic. And I must say that, as I said above, they do look quite cool - the built-in heatsink (?) fins flow nicely in to the track fittings and look very "tech". They certainly do not produce as much raw light as the 20W halogens, but they look good and don't cost too much more than the normal MR16 halogens. If they last for a few years I'll be reasonably happy with them!

It is certainly not the case that they produce "the same light as a 50W halogen" as some suppliers (not this one!) have claimed, but they do produce enough light. They do offer energy savings, reasonably large when compared to a typically over-lit room of MR16s or GU10's, but fairly minor compared to a T5 fluoro. We'll have to wait and see what their longevity is!

Certainly they are a far better offering than the "$100 each and very little light output" units that I looked at 12 to 18 months ago.

[zzsstt]

All of which brings me to the topic of LED drivers! Which one? Because this is track lighting I need a constant voltage 12V unit, rated at perhaps 60W to 100W. There's a guy on eBay selling a 100W indoor unit, and various people selling IP66 exterior drivers for garden lighting. But very few make mention of the efficiency of the driver. PC power supplies these days are often rated at 85% efficient across the majority of their output range. Does that mean that the LED driver I saw that was rated at 83% efficient is par for the course?

[Sandivee]

A lot of the Taiwan based LED chip manufacturers seem to think that warm white is 3500K.

One thing you will notice about the 5W lamps from ebay are that they have a very narrow beam angle which makes them seem brighter than most of the other cheaper LED MR16 lamps. The 3x3W CREE MR16 whilst only operating at ~5W actually has a wider beam angle and probably close to a 25-35W halogen, but obviously can still replace a 50W if you don't need intense light.

Drivers are certainly a good choice to replace, but probably worth upgrading to a better quality lamp at the same time also. It'll be hard to find something thats above 90% efficient, so your somewhat stuck with that.

Either way, better than a halogen IMO.

James

[Photon]

Has anyone had any experience with Bridgelux LED products?

http://bridgelux.com/
http://www.candlepowerforums.com/vb/sho ... LED-Review

THey seem to have the opposite problem of conventional leds, in that they provide a flood of light with no focused beam.

[LeonF]

How to solve the hot emitting problem of MR16?
Most led lights have a suitable heatsink whilst to the diameter of a led halogen replacement bulb is around 50mm, even there is actually not height restriction for the ceiling, we had lengthen the aluminum heatsink to 80mm, total length 112mm for the 13.5w GU10. Still too hot around 70c that we can not promise it will last for 50k+hours... Anyone who has experience of the the tempt of the Osram or Philips led halogen bulb approx 10w????

Leon

[Tracker]

All of which brings me to the topic of LED drivers! Which one? ?

The guy at Oatley Electronics, is doing a lot with LED lights and supplies..

I wonder what he can supply. ... I know that he also concentrates on "Drivers"..

I saw a 20W chip there the other day... LED the size of a stamp.
problem is the need for a good heat-sink.. I live in hope to get a great and compact central lighting alternative that gives a pleasing WARM light.

The annoying thing is that I keep waiting, because the change is so fast and extensive, that I don't want to committ, because I just KNOW, that the perfect choice will appear tomorrow.

its something to do with the switching frequency of a switchmode supply - dedicated LED drivers are required.

My gut feeling is that this will not be right.. I can't see the frequency having a great impact.. I would be more interested in power stability..
..
.

[dmaunder]

Ok, I am new to all of this, but learning fast. I have just leased an office 140 sqm with around 24 Halogen down lights. The ceiling is plaster, with no manhole and for all intents and purposes inaccessible, so removing/replacing the transformers is not a realistic option. What are my options?