Low wattage (LED and CFL) bulbs performance

[ban-all-sheds]

I think it's another case of poor wording, rather than bad maths.

It's often poor wording arising out of bad maths.

There's nothing intrinsically wrong with percentages greater than 100% when one is talking about comparisons.

I recently read the phrase "300% cheaper". There is something very wrong with that.

 

[JohnW2]

There's nothing intrinsically wrong with percentages greater than 100% when one is talking about comparisons.

I recently read the phrase "300% cheaper". There is something very wrong with that.

Indeed there is. There are many abuses and incorrect uses of percentages, and that's an example of what is probably one of the most common. Even if one is prepared to overlook the 'wrongness' and attempt to work out what was intended, it's usually not clear. I imagine that "300% cheaper" is probably usually meant to mean that something is a either a third or a quarter of the cost of something else, but those are by no means the only possibilities!

In terms of 'correct usage', one clearly cannot talk about "X% cheaper" if X is greater than 100, since "100% cheaper" means (mathematically) that the cost is zero (just as, say, "90% cheaper" means that it costs only 10% of the cost of the comparator).

Kind Regards, John

 

[JohnW2]

Out of interest I looked up info on old 8W Philips bulbs and they were rated at 310 lm with 10 years life to 50% failure. New bulbs 2 x 1.4W at 100 lm each and 8 x 3W at 250 lm each so lumen output dropped from 3100 to 2100 yet the room is very much brighter. ... So if neither the watts used or lumen given out will show how large or how many bulbs are required to light a room how should we select replacement bulbs?

IF that the quoted light output (lumens) is correct, I can think of only two main explanations for what you have observed. Firstly, the colour of the light may differ, which could (in combination with the environment in which they're being used) leave to a different perception of the 'brightness' of the room, or even possisbly different light meter readings (depending on the meter). Secondly, the pattern of dispersion of light may differ between the lamps, again giving rise to different perceptions of brightness.

However, there is a big 'IF' at the start of that preceding paragraph and the cynic in me suspects that the quoted light outputs may not always be quite what they seem (i.e. essentially 'wrong' in some cases!) - perhaps due to the way in which they are measured.

Kind Regards, John

 

[ericmark]

Out of interest I looked up info on old 8W Philips bulbs and they were rated at 310 lm with 10 years life to 50% failure. New bulbs 2 x 1.4W at 100 lm each and 8 x 3W at 250 lm each so lumen output dropped from 3100 to 2100 yet the room is very much brighter. ... So if neither the watts used or lumen given out will show how large or how many bulbs are required to light a room how should we select replacement bulbs?

IF that the quoted light output (lumens) is correct, I can think of only two main explanations for what you have observed. Firstly, the colour of the light may differ, which could (in combination with the environment in which they're being used) leave to a different perception of the 'brightness' of the room, or even possibly different light meter readings (depending on the meter). Secondly, the pattern of dispersion of light may differ between the lamps, again giving rise to different perceptions of brightness.

However, there is a big 'IF' at the start of that preceding paragraph and the cynic in me suspects that the quoted light outputs may not always be quite what they seem (i.e. essentially 'wrong' in some cases!) - perhaps due to the way in which they are measured.

Kind Regards, John

I do not believe the Philips data the lamps warm up slowly so one is never sure as to when at full brightness however I remember using a mark/space selection to over drive a LED at Uni to increase the brightness of a standard red LED so instead of 12ma we used 24ma but only for 50% of the time and we had to measure the output making a graph etc. The light output did seem to increase to the eye but the lux meter did not record this increase rather poor results as far as I was concerned at the time.

However we were told this was a standard way to increase output from an LED and that has now started me rethinking about the results and I wonder if a Lux meter can really show the perceived light from a HF device when the mark/space ratio is not equal?

In theory lumen should reflect the human eyes ability to see but I am not sure that any measuring device can do this with unequal mark/space ratio and this would explain the difference perceived.

It would also explain the problems I had using a camera and it's readings to compare light. I had at first thought it was UV and Infra red components causing the higher than expected readings with camera from the CFL to LED now I think it may be the mark/space ratio.

 

[ban-all-sheds]

Indeed there is. There are many abuses and incorrect uses of percentages, and that's an example of what is probably one of the most common. Even if one is prepared to overlook the 'wrongness' and attempt to work out what was intended, it's usually not clear. I imagine that "300% cheaper" is probably usually meant to mean that something is a either a third or a quarter of the cost of something else, but those are by no means the only possibilities!

If my numbers come up tonight, and I have money to burn, it might be a laugh to take whoever it was to court if they refused to pay me 2 x the price of the expensive thing to take their 300% cheaper thing off their hands

 

[JohnW2]

... I remember using a mark/space selection to over drive a LED at Uni to increase the brightness of a standard red LED so instead of 12ma we used 24ma but only for 50% of the time and we had to measure the output making a graph etc. The light output did seem to increase to the eye but the lux meter did not record this increase rather poor results as far as I was concerned at the time. ... However we were told this was a standard way to increase output from an LED and that has now started me rethinking about the results and I wonder if a Lux meter can really show the perceived light from a HF device when the mark/space ratio is not equal? ... In theory lumen should reflect the human eyes ability to see but I am not sure that any measuring device can do this with unequal mark/space ratio and this would explain the difference perceived.

You make some very important points. Although, as you say, lumens relate to the brightness of light as perceived by humans (e.g. taking into account varying sensitivity of the eye to different wavelengths etc.), a problem arises when the light intensity is not constant. The human eye will tend to perceive something close to peak light intensity (the photochemical nature of the retina having the effect of introducing what one would call 'persistence' in a CRT), whereas a meter will measure some sort of average intensity - hence, as you say, brightness as perceived by humans will tend to increase (for same average light level) if the duty cycle ('mark/space' ratio) is less than 100%.

This was not much of an issue with incandescent lamps, even when dimmed by waveform chopping, since the thermal nature of the light-generating process largely 'smoothed out' any variations in light output. However, modern light sources, particularly LEDs, have the ability to literally reduce light output to zero during the 'space' parts of a sub-100% duty cycle, hence creating the potential for human/meter assessments of brightness to be very different.

Whether any of this is relevant to what you're discussing, I don't know. In other words, I don’t know whether (m)any manufacturers of LEDs employ duty cycles less than 100% (other than, perhaps, for dimming*) in order to increase perceived brightness. One might hope that they do, but this, of course, would cost a little more money than a simple 100% duty cycle.
[* in view of what we’re discussing, dimming by duty cycle reduction might be expected to be non-ideal, in terms of eye-perceived brightness, for LEDs]

... It would also explain the problems I had using a camera and it's readings to compare light. I had at first thought it was UV and Infra red components causing the higher than expected readings with camera from the CFL to LED now I think it may be the mark/space ratio.

I’m not so sure about that. Whilst the aspects of human perception we’re discussing can result in a situation in which (visible) light is perceived as brighter by a human eye than a meter suggests, I can’t see how it could explain meter readings being higher than would be expected from the perceived brightness of (visible) light. On the other hand, if (as you mention) the meter were also measuring non-visible light (UV & IR), or even just ‘over-weighting’ parts of the visible spectrum to which eyes are less sensitive, then it could well produce ‘higher than expected’ readings.

Kind Regards, John