Published 16 November 2011
Go to COMMENTS

How well do LED globes stack up?

An earlier article on compact fluorescent lamps concluded by saying we might want to wait until LED technology matures. I see now that hardware stores like Bunnings and Mitre10 are stocking LED replacements for standard incandescent globes. This is a test to see how they compare.

Which lamps have we tested

This is a test of the Viribright 11 watt cool white LED globe costing $NZ$24.95 and the Philips VisionLED 7 watt warm white globe costing NZ$35.95 both purchased from Bunnings during November 2011.

LEDglobe Philipsglobe

They are compared to a standard 60 watt incandescent bulb costing 79 cents. A compact fluorescent spiral lamp 15 watt eco brand costing NZ$7.95 is also included for comparison.

The 60 watt incandescent bulb
used as a reference for the LED tests
and the 15 watt CFL.

stdglobe CFL bulb

General information on packaging and data sheets

Viribright 11W lamp

Philips 7W VisionLED

Claims on packaging:
Uses 80% less energy. Lasts 6 years (8 hrs daily) = 17,520 hours.
11W/B22/220-240V 400lm/50-60Hz
5600K CRI 80 Beam angle 270 degrees.
Mercury Free. Replaces a 60 watt incandescent. Made in China
Claims on packaging:
7W similar to 40W*** (***Guide only. Actual light output not equivalent)
B22 (100-240V 50-60Hz)
2700K 350 lumen (50 lumen/watt).CRI >80.
Extra Long life up to 25 years** (used 2.5 hrs per day) [=22,812 hours]
Mercury and lead-free Made in China Warranty: 3 years.
Warnings and instructions:
Non-dimmable. Do not use with ANY dimmer.
Not to be installed in wet or damp location. Indoors only. Not for use in totally enclosed luminaries. Do not stare into beam directly. Stop use if product becomes dim, out of order or begins to blink.
Warnings and instructions:
Not for use with dimmers. Not for use in enclosed light fixtures. Not for use with emergency exit fixtures and emergency lights. Not for outdoor use. Turn off power before changing lamps.

Tests performed on the lamps

We have not verified CRI (colour rendering index); neither have we performed any life tests (switching on/off cycles or reduction in light output over time).

Light Output Testing

The bulbs are mounted "naked" in a test jig which allows the bulb to be rotated in marked 15 degree steps. A LX1010B lux meter is mounted into a box designed to limit unwanted light and the assembly is arranged to place the lux meter exactly 1 metre from the bulb. The test setup is carefully checked to avoid extraneous light on the meter and is also arranged to minimise light reflected from nearby surfaces especially when the bulb is rotated away from the meter. Care is taken to measure the incandescent bulb at exactly 230V since incandescents are quite sensitive to mains variation. Lux readings are plotted in Excel for each bulb after normalising to the incandescent readings. (On the scale as shown, 1.0 represents the measured luminance of the incandescent bulb on the vertical axis. In this case, the 1.0 scale therefore represents a reading of 60 lux at 1m from the bulb.

The plot of luminance versus angle is below:

HRP of lamps

This plot is actually a 2D representation of a 3D situation. You have to imagine the light bulb is at the centre of a sphere. The 60 watt incandescent bulb (red curve) is nearly an isotropic light source. That is to say it has nearly uniform light output over virtually the full sphere. (Although only 240 degrees is plotted here, it is actually only 10% down at 300 degrees after which it's light output reduces to a low value for just a few degrees right behind the base of the lamp.)

So, what can we say about these light readings?

Rating light output of lamps in lumens

These days, it is not a simple matter to compare brightness of lights based on wattage. When incandescent lamps were the only technology on the block, most people had a good feel for the brightness of a '100 watt' lamp and smaller siblings. Now we have CFL's and LED's amongst other variants of technologies where a lamp might draw only 11 watts or 15 watts to produce similar light levels to the old 60 watter. So, is there a better way? Not yet. Some say quoting lumens for the lamp is the way to go, but then you have to quote the half-luminance angle, or produce a plot like the ones above. Even then, it is not a simple comparison, especially when you have to account for differing colour temperatures (cool-white/warm white/natural white/2700K etc). However, it is a starting point.

Measuring the lumen output of a lamp is not a straight-forward matter. It is, by definition, the integrated light output over spherical space. With other than incandescent lamps then, you cannot just look at the peak lux measured, account for the distance and work back to lumens. You have to take the light output at all angles and sum these as a proportion of the whole sphere. For this exercise I have calculated a rough lumen rating using 12 measured points. This will give an approximate result but I expect it to be a good enough comparative guide.

Once the rated lumen output of the lamp is known, then the efficacy of the lamp can be calculated. This is a measure of how well the lamp converts watts of power to light. Following are the results of that exercise.

  1. Incandescent lamp 60 watt : 670 lumens. 11.2 lumens per watt.
  2. Viribright LED 11 watt : 426 lumens. 42.6 lumens per watt.
  3. Philips VisionLED 7 watt : 330 lumens. 44.2 lumens per watt.
  4. CFL eco brand 15 watt : 550 lumens. 44.4 lumens per watt.

I have used measured true wattage to calculate the efficacy figures. As you can see, the incandescent 60 watt lamp produces the most light, but does so inefficiently. The CFL provides not much less light, but does so for only 1/4 the power input. The LED lamps have a similar efficaciousness to the CFL but are lower in total light output. I have to reiterate that these figures are only an approximate guide. Not only is the light integrated using only 12 points, but the light meter itself is liable to respond differently to the differing spectrum of light produced by the three lamp technologies.

Placing lamps in fittings

Once lamps are placed in any enclosed or partly enclosed fitting such as a downlight, desk lamp or any other architectural feature, the situation changes. Most fittings tend to narrow the beam but this is not always true. The testing of lamp housings is well outside the scope of this article, however I did perform one simple test. I inserted each lamp into a simple anglepoise desk lamp and measured the light intensity 1 metre directly below, comparing this to the readings of the naked light bulb. This lampholder is a typical desk lamp, with all white powder-coated interior.

It is interesting to note that the fitting effectively focussed the light from the incandescent and the CFL, but had little impact on the LED lamps; which already had fairly narrow beams. The conclusion from this is simply to say that the fitting is important and these should be chosen both for the lamp technology and for the intended purpose.

Lamp flicker

Some people are quite sensitive to lamp flicker, mainly that from older tube style fluorescents. This test used a visible-light photodiode coupled to the Praxis measurement system to produce a plot of flicker components in the light. As a point of reference, the 100Hz flicker from the 60W incandescent bulb is used, since that produced the greatest light flicker component. Below is an overlay plot of the four lamps showing their flicker output up to 2000Hz.

flicker plot of lamps

Clearly, the 100Hz from the incandescent far exceeds the flicker from all other lamps and so I expect this is not too much of a concern, since few people complain of flicker from incandescent bulbs. The next worse is the CFL, which has dominant 100Hz, 200Hz and 300Hz components, although none of these is higher than 22% of the 100Hz flicker from the incandescent. Flicker from the Philips LED is less than 3% of the incandescent, while the viribright LED is much lower again, although it has a non-harmonic component at 132Hz. The two LED lamps can safely be said to produce a very clean, flicker-free light.

Power consumption, power factor and AC harmonics

These are all captured with a Yokogawa WT210 instrument.

LampWattsVACurrent distortionTotal PF
Incandescent 60W 58.6 58.6 3.4%1
LED Viribright 11W 10.0 20.04 156.3%0.5
LED Philips 7W 7.4 9.7 63.6%0.76
CFL 15W eco 12.4 13.4 33.5%0.92

As expected, the incandescent has a unity power factor; the 3.4% harmonic current is the distortion of the voltage at the workshop. The Viribright LED lamp is very poor, having considerable harmonics which can be seen in the plots below. The Philips LED in contrast, has much lower levels of current harmonics and is acceptable. The big surprise is this CFL, which has very low levels of harmonics and consequently a high power factor. This particular model is promoted as a high power factor (HPF) type on the packaging. It can be done.

Below are plots of harmonic currents of the four lamps. Vertical scale is in milliamps.

Incandescent harmonics Viribright harmonics Philips LED harmonics CFL harmonics

To complete the picture, below is a capture of the waveforms of the three lamps, excluding the incandescent. Note that the current scale varies, as does the timebase, but the intent is to show the shape of the current.

Although none of the
current waveforms (green traces)
resemble a sine wave, both the
Philips LED and the CFL draw
current over most of the
cycle which means they have
lower levels of harmonics.
The Viribright draws
a sharp spike of current
which contains many harmonics
extending well beyond the 19th
harmonic plotted in
the previous graphs.
harmonics waveforms

It is probably fair to say most people don't understand or care about AC harmonics caused by appliances. There is no apparent impact on operation. My concern is that these kinds of small distorting appliances are starting to proliferate and the numbers will increase rapidly as incandescent lamps are replaced with lamp technologies with electronics in them. Add to that the rapid increase of digital TV set-top boxes, phone and ipad chargers and so on, the potential for causing ever-worsening distortion of the mains is high. One distorting load, or ten will not matter, but ten in every house, in every street, over the whole country will matter. It is in the consumers interest to press for appliances with good power factor, which also then means low harmonics. It clearly can be done as evidenced by the eco-brand CFL in this test.

Sensitivity to variations of mains voltage

It is well known that the light output of an incandescent bulb goes up and down with the applied voltage. In fairness, voltage variations have to be large to perceive much change in brightness, but people do notice the 'dim' nights when the voltage gets too low. This test varies the voltage from 200V to 250V and plots the change in power consumed as well as light output. The two LED lamps varied neither in power or in light output at all. In fact they were tested down to 120V and nothing changed. The CFL did show a slight variation and the incandescent a considerable (in context) variation. Plots of these two are shown below.

variation with voltage plot

The blue line is the AC consumption and the red line is the lamp luminance. Both are normalised to 100% at 230V (local nominal voltage).

Lifetime and running costs

Both LED lamps are rated to last a long time. 12-15 years of operation is promised, assuming 4 hours use every day. What is not clear is how they would fail. Is it one day OK, next day broken, or is it a gradual loss of light, considered to be end-of-life at some unspecified percentage of the 'new' light output? It is a source of amusement to imagine an owner returning to the shop clutching his failed LED lamp and ten year old receipt, saying it didnt last fifteen years, can I have a replacement please? In any case the manufacturers warranty is for only three years and perhaps that is what one should base any calculations of payback periods.

Looking briefly at relative costs now, the incandescent bulb cost $0.79, the Viribright LED $24.95, the Philips LED $35.95 and the CFL was $7.95. You could buy 35 incandescent bulbs for one viribright and 45 for the cost of a Philips LED. So, based on life versus lamp price alone, that number of incandescents would outlast a LED globe. You could get 3 CFL's for one Viribright and 5 CFL's for the cost of a Philips LED. In this case, the lifespan vs lamp cost is more or less equitable.

Now looking at running costs. If you replaced the 60W incandescent with a Viribright LED globe, you would use 50 watts less power. Assuming it runs 4 hours every day, you would save $17.50 per annum of electricity(at 24c per kWh). This makes the payback period only 1.4 years and so it starts to look like a worthwhile proposal, because after 1.4 years you would be saving money on the investment. The Philips LED lamp is reckoned to be the same as a 40 watt incandescent, and on replacing one with the other, you would be reducing consumption by 33 watts per lamp. You would save around $11.60 per annum of electricity, making the payback period around 3 years. Only on the assumption that the lamp is going to last much longer than 3 years would this be a economic proposition.

Of course, economics is not the only issue. Some people are quite happy to pay more to help the economy or 'save the planet'. If this looks like tongue-in-cheek, it isn't meant to be, because I think there is always merit in trying to go further for less, or in this case, produce more light for the watts. It is a rather complex issue when you think about whole life costs including energy of production and impact of disposal. These matters are out of scope of this article.

Summary

So, is an 11W Viribright LED equivalent to 60W incandescent and is a 7W Philips LED equivalent to a 40W incandescent as claimed? My short answer is "not in all cases", mainly due to the much narrower beam of both LED's compared to the incandescent. However, there are situations when you want just that, so in that case, these LED lamps would make a viable replacement. I have tried both LED globes in a desk lamp and they produce a good level of pleasing light for reading and tasks. The lamp shroud conceals the whole globe so I dont get the 'laser-burn' from the Viribright globe. I expect the LED's could be used within recessed downlight fittings over working areas such as kitchen benches, but not as wide-area illumination. In fact I think recessed downlights for general lighting in large open-plan areas is ineffective and an essentially dumb idea. One of those many triumphs of form over function.

Pros and cons of each lamp in turn

Links

Comment on this article via the contact form. Comments published if relevant.
Power quality of small appliances (Related article)
Top of this page
Axino-Tech HOME
Viribright Home
eco bulb Home

From: MScaife
1 February 2012

Comment

I purchased a VIRIBright 11W from Bunnings in Oct 2011, for use as a desk lamp by my computer, runs approx 8 hours per day. Past couple of days the lamp has started to fail, when turned on in the mornings the Leds come on very dimmed and after 10 minutes or so suddenly burst to full brightness; sometimes they dont come on at all.

At $24 and lasting only 5 months, absolute crap and rubbish, advise dont purchase these, they are duds. Mike Scaife Wellington

From: Axino
13 May 2012

Comment

After some 5 months of operating the Viribright LED lamp that was bought in November 2011, it too is giving problems. The symptoms are different (so far) to those described by Mike above. This lamp occasionally blinks on and off at a slow rate for a period, then settles down, but repeats this at seemingly random intervals. I have not connected the behaviour with voltage, temperature, or time of day. This bulb is fitted into a desk lamp and operates only on average 3-4 hours per day.

I have to concur with Mike that these lamps appear quite unreliable.

From: Ian Williamson
14 May 2012

Comment

Regarding: Led and cfl tests
I must congratulate you on a very clear and concise set of tests on the Viribrights Led lamps and the comparison with the philips. I also have done some tests in particular with regard to power factor/ bad current waveforms on various CFLs and other equipment in the home or office and have lots of graphs for my efforts.

One of the main reasons that Lamp manufacturers make such poor performing electronics is the regulations do not require lamps under 25Watts to meet the normal standards of power quality.

Philips <25W CFLs are bad along with most other brands except Eco. The Philips CFLs >25W are clean. One thing I have noticed is that mixing Eco lamps and others on the same circuit seems to cause the Eco brands to fail early. I believe this is due to the "snubber" circuit in the Eco lamp (which is not present in the other brands)trying to "clean up" the waveforms of the others. I cannot prove this but after removing all non Eco CFLs from home I have had no failures. The addition of these two components (a capacitor and a resistor)cost just a few cents but in sales of millions of lamps I suppose it adds up.

Inverter heat pumps also are a significant problem with cheaper brands having very bad current waveforms and my contacts in the power companies tell me it is a significant problem. Where commercial companies can be required to "clean up" their power, home owners have not had to yet.