LEDs vs CFLs

[John D v2.0]

Eric Phillips are very good for warranty, I tweeted them when an LED I bought went back to flashing constantly after 2-3 years. They got in touch and sent me a replacement, all the way from Netherland. The new one was a better model too!
And I think the problem with LEDs is often the cri. Low cri and high colour temperature can look very bright, but doesn't really light things up well. Remember the old sodium street lights, they were very efficient but you couldn't read a colour underground map or anything!
Better quality LEDs have less lumens per watt but the light they do produce is better for seeing things.

Not sure what that is about pulsed light seeming brighter, LEDs should be dimmed by pwm as they don't dim linearly well, but that won't make them seem brighter than the average time they're on. If the frequency is too low you'd see the brighter light flashing though. Your eyes are non linear too, they adjust for light levels down to a certain point.

 

[bernardgreen]

LEDs should be dimmed by pwm as they don't dim linearly well

Sorry John but that is not totally correct. The brightness of light from an LED element is proportional to the current driven though it and the response time of the LED element is fast enough that pulsed current will produce pulses of light.

PWM ( Pulse Width Modulation ) of an LED element appears to dim the LED only because almost all "white" LED lamps are an LED element producing ultra violet light which activates a phosphor layer to produce white light (or with different phosphor, the colour required ). The phosphor has persistance and thus the pulses of UV light are converted by the phosphor into smoother ( not pulsing ) white light.

With non phosphor LED lamps PFM ( Pulse Frequency Modulation ) can provide reasonably linear dimming using the persistance of the retina in the eye to smoother out the pulses of light for the brain to process.

 

[JohnW2]

Not sure what that is about pulsed light seeming brighter ....

Eyes/brain tend to perceive the peak brightness of pulsed light. If the light is pulsed with a 1:1 mark:space ratio, then the peak brightness will be about double what it would have been without pulsing for the same average (LED element) current/power and light output. If the frequency of pulsing is high enough, then, as bernard has said, the persistence of the retina will make it appear to have a constant brightness, close to the peak.

Quite apart from dimming, I think that this 'trick' is often used, particularly in cheap LED lamps/bulbs - pushing the current capabilities of the LED element(s) close to their limit (hence reduced reliability/life?) in order to result in greater perceived brightness.

Kind Regards, John

 

[ericmark]

With a LED replacement for a fluorescent tube very likely you have a switched mode unit, however with a bulb in the main they seem to us a simple capacitor to limit current, I noted the fan used in my living rooms seems to be rotating slowly, if the 10 bulbs had switch mode regulation that would be very unlikely, it is far more likely there is a simple capacitor, however I have not noted any problem with cameras.

As to what time the camera uses to measure light I don't know. But in the main in doors I am using 1/20th second, at this speed I have not seen banding, but you make a good point.

 

[JohnW2]

As to what time the camera uses to measure light I don't know. But in the main in doors I am using 1/20th second, at this speed I have not seen banding, but you make a good point.

As Bernard has said, with LEDs (or, indeed fluorescent tubes and CFLs) it's not just the persistence of the retina (i.e. our perception) that makes pulsed light appear fairly continuous - there is often a phosphor layer which also has persistence - and that will result in the light actually being (i.e. not just perceived as being) less 'pulsed'. That being the case, you might not see a lot of banding even if you used a very short exposure time.

Kind Regards, John

 

[dishman]

I replaced Philips CFL with first 3W and then 5W LED lamps, both seemed brighter than the CFL so in my experience the theory and practice don't match.

I always assumed that this was probably to do with the (generally) higher quality of light emitted by most LEDs. Probably something to with a generally higher CRI. As you say there are a lot of poor quality CFL. Some CFLs have a muddy quality of light, and I have found they are often under rated in terms of equivalence. So a 60w equivalent is probably really equal to a 40w incandescent.

Many LEDs often have a more intense and focused light which again probably makes it seem brighter than an equivalent CFL.

There is a such a range of LEDs it is hard to keep up sometimes.

Personally, I like the filament ones. But how do these compare with the (I assume) directional ones? The filament ones have near a near 360 degree emission of light. The directional ones are around 250 degrees or less.

https://www.bltdirect.com/osram-par...clear-es-gls-100w-alternative-very-warm-white

https://www.any-lamp.co.uk/philips-corepro-ledbulb-e27-a60-13-5w-840-matt-replaces-100w


In addition the directional ones seem to have or require heat sinks, but the filaments do not. Is it purely due to a better design for heat dissipation?

So, for now, a good cfl like the philips tornado, that has a good quality light, is very bright and has a 360degree emission (important when you have up-lighting chandelier type) fits the bill for me.

I think the reality of modern lighting is that it has been fully commercialised in the sense that there is a huge plethora of choice now. You have to be far more specific about choosing the correct type of fitting, placement, bulb and type of bulb for the lighting task at hand. There is, especially with LEDs, far more rubbish to wade through too.

 

[aptsys]

Eyes/brain tend to perceive the peak brightness of pulsed light. If the light is pulsed with a 1:1 mark:space ratio, then the peak brightness will be about double what it would have been without pulsing for the same average (LED element) current/power and light output.

This is only true at one frequency and depends on the wavelength of the light emitted. From what I remember from when I did my PhD, the frequency was always between 40Hz and 90Hz, so useless for lighting.

 

[bernardgreen]

40Hz and 90Hz, so useless for lighting.

There is a trade off between the switching frequency ( high to avoid strobing ) and the heat produced in the semiconductor junctions of the switching device when it is in the linear stage between fully ON and fully OFF. Too much heat and the device fails. Early transistors could not switch rapidly and thus the linear period was long so low frequency switching was necessary to reduce the time the device was in the linear stage. Modern transistors switch much faster fropm OFF to ON and from ON to OFF and thus the switching frequency can be much higher.

 

[aptsys]

There is a trade off between the switching frequency ( high to avoid strobing ) and the heat produced in the semiconductor junctions of the switching device when it is in the linear stage between fully ON and fully OFF. Too much heat and the device fails. Early transistors could not switch rapidly and thus the linear period was long so low frequency switching was necessary to reduce the time the device was in the linear stage. Modern transistors switch much faster fropm OFF to ON and from ON to OFF and thus the switching frequency can be much higher.

I'm talking about the eye response

 

[JohnW2]

This is only true at one frequency and depends on the wavelength of the light emitted. From what I remember from when I did my PhD, the frequency was always between 40Hz and 90Hz, so useless for lighting.

It will obviously only be a one pulse frequency and one wavelength (if any) that the perceived brightness will be be exactly double the 'average' brightness. However, if I recall correctly, the perceived brightness of pulsed light will (within reason), if the frequency of pulsing is great enough for the pulsatile nature not to be apparent, always be perceived as brighter than the actual 'average' brightness. However, as always, my recollection may be wrong!

Kind Regards, John/

 

[bernardgreen]

24 Hz will appear as constant brightness provided the light source is not moving across the field of vision. The frame rate of projected films in the cinema is (was) 24 Hz with no persistance on the screen and almost no flicker.

It is a different matter if the light is from a spot source that is visible and moving across the field of vision. The light source can appear to be trailing a line of pearls ( spots of light ) as it moves across the field of vision. This happens with the amber warning lights at road works, these are pulsed to get a bright light with much reduced power and hence much longer battery life. Because the pulsed light is very bright it creates a row of long persistance "dots" in the retina.

 

[JohnW2]

Because the pulsed light is very bright it creates a row of long persistance "dots" in the retina.

"Very bright light" is a different matter, because (whether pulsed or not) that can result in changes in the retinal receptor cells (such as temporary depletion of photochemical pigments) which can result in very long 'persistance' (many seconds, sometimes longer), far longer than the 'usual persistance' at more normal light levels.

Even 'usual persistance' is far from entirely retinal - some 'smoothing' (and a lot of other sophisticated processing) happens in the brain.

Kind Regards, John

 

[bernardgreen]

(such as temporary depletion of photochemical pigments)

That results in an after image in the complimentary ( inverse ) colours to the original bright image. The pearls are the same colour as the light an short duration ( less than a second )

 

[JohnW2]

That results in an after image in the complimentary ( inverse ) colours to the original bright image. The pearls are the same colour as the light an short duration ( less than a second )

Indeed so.

Kind Regards, John

 

[ericmark]

I remember in University being given a project to see if pulsing the LED would make it brighter without it burning out. The whole project was flawed I realise now as we were measuring the output with a meter which would have had some smoothing built in so even if it was brighter, the meter would not record it. It would have needed at least an oscilloscope so time could be stretched.

However what we did find is the cheap red LED could be damaged through over current and although it still had an output after the over current it was lower than before. I would guess the same is true with white LED's used for lighting.

This means if the LED is current regulated using a simple capacitor any over voltage may not actually cause them to fail, but could cause them to have a lower light output. We all know the fluorescent output drops with age, but not sure if the LED output drops due to age or over voltage, it does seem to drop but hard to work out why, also using a dimmer, or PIR with no neutral it must cause a slight volt drop, using a capacitor this would mean reduced output, but using a pulse width modulated driver it would auto correct for volt drop. I have a replacement for a quartz halogen tube marked 85 ~ 250 volt, I would guess a dimmer switch would not effect the output, it would auto correct what the dimmer does. It is marked as non dimmable.

If we take a fluorescent tube and compare to a LED very little in it, however we also have to include the LED driver and fluorescent tube ballast, the performance of a wire wound ballast is very different to a HF electronic unit, and a resistor or capacitor as a current limiter is very different to a PWM driver. If one compares a single lamp with a single lamp yes you can show which is best, however there are so many variables one can't really do a comparison.

All we can say is in general terms LED's are slightly better output wise to fluorescent, faster to start, more likely to work with cheap dimmers, and last a little longer, however it is so little that it does not really warrant swapping to LED, just replace with LED when they fail. In order that bulbs matched I have actually taken out working CFL and tungsten and replaced with LED, what I now have is a storage problem specially with GU10 bulbs where they have been swapped for LED and old unit kept as it still works.