Caving - How many Lumen for a light is enough?
- Thread starter Bloke28
- Start date
The standard way is to use an 'integrating sphere' - essentially a large hollow ball painted on the inside with fancy ultra-white paint with a hole for a light to be poked inside, a hole for a lightmeter to be attached, and a baffle between the two holes to prevent a direct light path. The idea is that the light the lightmeter receives is essentially independent of the beamshape of the light, and proportional only to the actual total light output.Pitlamp said:
Properly calibrated professional ones are expensive, but amateur models can be made which are reasonably accurate (good enough to tell if a stated light output is likely to be roughly correct or a work of marketing fiction).
DavidGibson
Member
bograt said:
There isnt one! They are measuring different quantities. It would like trying to convert acres to inches or pints to feet.
The lumen is simply a measure of power. It can be converted to watts, for example. The candela (or candlepower) is a measure of how concentrated the light beam is. If you have the same number of lumens in a tighter beam, the candela rating will be higher. Technically speaking, the candela is equivalent to a lumen per steradian. The steradian is a measure of three-dimensional angle.
This means that for photography, where you want a particular bit of scene to be illuminated at a certain level, it is the candela rating that is more important. (Or, actually, the candela-second rating of you want to get technical). But for moving around, a tight beam is not necessarily good, so you might want a lower candela rating but a higher lumen rating, which would infer a widebeam lamp.
Someone asked how you would measure the output of a lamp in lumens. I think the easiest way is to measure the output in candelas and to make a rough estimate of the beam angle in steradians. Multiply the two and you have the lumens. You can measure the candela rating by using a photographic lightmeter. The difficulty is all in the numbers - you need to be able to convert from photographic exposure to candelas, and from degrees to steradians. One could ask ... is it all worth it? If you DO want to do it, you might find helpful an article of mine in the CREG journal. Its called Determining the Photographic Guide Number of an LED. The article describes how, given the lumen rating of an LED you can work out a photographic guide number. You would be working the opposite way around, starting with guide number (or exposure) and working back to get lumens.
DavidGibson
Member
Photometric quantities are nowhere near as easy to deal with as metres, kilograms and seconds. For example - if you simply use a "PV cell" how will you know the efficiency of the conversion of incoming light to electrical power? How will you know whether it has the right response to different wavelengths of light? (Part of the definition of the lumen is that the light power is weighted according to the human eye sensitivity at different wavelengths). And - if you are going to measure the output power from a PV cell... well, that depends on the load you give it. If its open-circuit you will have volts and no current (hence no power); if you short-circuit it you will have current and no voltage (hence no power) and if you set the load anywhere inbetween you can have whatever power figure you like - up to a maximum. So, in effect, the output power bears no easily-used relation to the incident power... when it comes to using the cell for measurement, as opposed to simply power generation. In fact, it is the short-circuit current that is proportional to the incoming light intensity, so that's what you would need to measure.bograt said:
And you mustnt forget that you wouldnt be measuring lumens, you would be measuring "lux", that is lumens per square metre, of course. And how would you calibrate the system? For metres, kilograms and seconds, there are easily established references. E.g. you can buy a ruler, or a set of weights, or a clock. There is no easily established reference for photometric units - even for professional engineers - unless you use a commercial lightmeter and hope that the manufacturer calibrated it properly.
Fortunately, none of this matters all that much because the human eye responds to the logarithm of the intensity of the light. Photographic exposure - like the decibel scale for audio work - is logarithmic and so you dont have to be particularly accurate in your measurement. A quarter of a "stop" either way is not a LARGE change in photographic exposure, nor in perceived brightness even though, in lumens, it might look quite a large number. A quarter of a stop is about 20%, so youre not going to need four-figure accuracy in your measurements. 1000 lumens is not going to look THAT different to 800 or 1200 - although detectable of course if you switch between the two. But the beam angle may have a larger contribution to perceived "lamp brightness" than the lumen rating.
Actually, I may have gone OTT, there, with my list of difficulties. You can make comparisons easily enough, and for calibration you could always assume that a 700 lm lamp is what it says it is.
Nevertheless, I think that simply using a "bare" PV cell might not be easy for accurate measurements but a simple photocell connected to an ammeter and calibrated with reference to a particular caving lamp might be all that is required for "ballpark" figures.
But I think the easiest way is to use a photographic lightmeter - because it is calibrated, and weighted for human eye response. To measure beam angle you can use a camera. If you photograph a lightbeam shone onto an even surface, that will allow you to get a good feel for beam angle. You could go further and measure of the overall lumen figure simply by counting all the pixels and adding them together (or asking someone to write a short program to do that for you). Its actually a bit more complicated than that, if the beam is wide, but you'll be within the 20% figure I have above. But how will you know the reflectance of the surface in order to calibrate the system? And how do you make sure the camera does not apply any automatic gain control? These are a lot of major headaches in trying to make any sort of reliable and repeatable photometric measurement
I dont mean to sound negative. Just pointing out that its as well to be aware of the limitations (and of course I cannot resist talking about the subject).
Big Jim
Member
Some sound advice on the subject...... https://www.youtube.com/watch?v=Vt8aAy_8Ub4
That does introduce uncertainty regarding the LED efficiency bin.
The only easy current measurements are typically tailcap measurements.
In the case of a linear driver and high currents, a typical cheap DMM may introduce resistance large enough to reduce the output current.
In the case of buck and/or boost drivers, a tailcap current measurement may be convertible into an LED current estimate if the circuit efficiency is known, but if the initial issue is one of doubt regarding manufacturer claims for output, then claims of LED bin and driver efficiency would presumably also be in some doubt.
The only easy current measurements are typically tailcap measurements.
In the case of a linear driver and high currents, a typical cheap DMM may introduce resistance large enough to reduce the output current.
In the case of buck and/or boost drivers, a tailcap current measurement may be convertible into an LED current estimate if the circuit efficiency is known, but if the initial issue is one of doubt regarding manufacturer claims for output, then claims of LED bin and driver efficiency would presumably also be in some doubt.
royfellows
Well-known member
and of course output will vary with temperature as well, look at Cree's @25 and @85
Like trying to work out the length of a piece of string
Like trying to work out the length of a piece of string
Though in fairness, a current reading can still be useful.
In the case of a linear driver, it can show actual emitter current or some possibly lower but close figure, and in the case of a buck and/or boost driver, it can show input power and at least put an upper bound on possible LED lumens if assuming a highest-bin LED (or if the manufacturer states the bin used) and a maximally efficient driver.
Of course, lumens coming out of the LED and lumens coming out of the light may be significantly different - a poor spot optic or a 'naked' LED recessed in a light can result in meaningful light loss, as could a badly-scratched front lens.
In the case of a linear driver, it can show actual emitter current or some possibly lower but close figure, and in the case of a buck and/or boost driver, it can show input power and at least put an upper bound on possible LED lumens if assuming a highest-bin LED (or if the manufacturer states the bin used) and a maximally efficient driver.
Of course, lumens coming out of the LED and lumens coming out of the light may be significantly different - a poor spot optic or a 'naked' LED recessed in a light can result in meaningful light loss, as could a badly-scratched front lens.
ChrisJC
Well-known member
All true. But you will get a reasonable estimate, much more easily than trying to actually measure it.
I wonder how many lamp lumen ratings are actually measured, rather than calculated from max current / max lumens rating on the LEDs.
Certainly on mine, I calculate the output as 12x 943lumens = 11200lm, which is the rated output of the LEDs at 3A (which I can accurately measure).
Chris.
I wonder how many lamp lumen ratings are actually measured, rather than calculated from max current / max lumens rating on the LEDs.
Certainly on mine, I calculate the output as 12x 943lumens = 11200lm, which is the rated output of the LEDs at 3A (which I can accurately measure).
Chris.
bograt
Active member
Sideways thinking--- If the OP is thinking about output from lamps, how about constructing a hood over the headpiece and building a series of reflectors, prisms, lenses to concentrate the light into a small orifice which can be measured by a light sensitive device - LDR + circuit, PV.cell - or whatever?
Obviously calibration would have be a problem in order to standardise this to "universal" units, but I suggest it may work for comparison purposes?
Is there a 'Base Standard' for calibration of Lumens?--(One for you Dave--)
This would offer the information we are looking for i.e. light available from caving lamps.
Obviously calibration would have be a problem in order to standardise this to "universal" units, but I suggest it may work for comparison purposes?
Is there a 'Base Standard' for calibration of Lumens?--(One for you Dave--)
This would offer the information we are looking for i.e. light available from caving lamps.
That's the job an integrating sphere does.bograt said:
With a highly reflective matte white coating on the inside, the idea is that whatever direction a particular light ray takes between source and sensor (ie whatever angle it comes out of the light source) it causes an effectively similar excitation of the sensor, and therefore the beamshape of a light source doesn't affect the reading, and the reading is simply proportional to total light output (ie proportional to lumens).
I'm tempted to home-make one and see how bad it is. Presumably one simple test is to point a narrow-beam light in different directions and see how that affects the reading - the less the reading changes as the light moves, the better job the sphere is doing of consistently integrating the light.
A fairly decent one would be a bare LED of known output 'bin', well heatsinked, running at known current and not running at extreme power level. Or even better, a succession of such LEDs which would tend to average out variation within a given output bin.bograt said:
For what kind of price?
I'd guess that the cost of anything truly standard would be high, and arguably that's a secondary issue.
Unless there was something to do the integrating sufficiently well in the first place, having any kind of calibration standard of even home-made levels of accuracy may be moot unless maybe only comparing lights on flood-only settings.
I'd guess that the cost of anything truly standard would be high, and arguably that's a secondary issue.
Unless there was something to do the integrating sufficiently well in the first place, having any kind of calibration standard of even home-made levels of accuracy may be moot unless maybe only comparing lights on flood-only settings.
Pitlamp
Well-known member
Ah - yes, sorry, I never intended to hijack this thread - but thanks for making the comment in such a pleasant and non confrontational way Bograt
A very big thank you to all those who have posted valuable information above.
Potholer: I suspect I may see you on 5th September for a regular function which you and I often attend. Will you be there? Maybe you and I can usefully chat about this?
A very big thank you to all those who have posted valuable information above.
Potholer: I suspect I may see you on 5th September for a regular function which you and I often attend. Will you be there? Maybe you and I can usefully chat about this?
