Measuring light intensity setup

[Emanuel]

I am confused about the proper setup for measuring light intensity in this case of an LED lamp.
Usually when there is a certain intensity requirement the specifications include the measured unit, e.g. cd, lux, etc.
But also the test setup is mentioned in the specifications, at least at what distance the intensity must be measured.
I am supposed to design a certain light source, both white and infrared. The required intensity of the visible light is in cd and the infrared in mW/sr, but not a word about the distance at which the given intensity is required.
I have the necessary equipment, but as I see, at least the cd is quite confusing, there are many options like cd/ft, cd/m and different standards are used.
For example some LED manufacturers are providing data measured at 1ft while others at 1m. Also the test standards are using different distances, for example car lamps are tested at 10ft in US and 3.16m in Europe, headlamps at 100ft in US and 25m in Europe and warning lights at 10m.
So if there is no specification of the distance and the customer is European, what will be a correct assumption of the cd type and measuring distance?

[Doctorandus_P]

This is a confusing topic, and I am no expert.

You can start with reading:
https://en.wikipedia.org/wiki/Candela

As I understand it, it does not matter at which distance you measure the amount of Candela's, but there is no sensor that can measure this directly, as a sensor has a fixed area size, and a Candela is a cone coming from a light source. So you have to know the distance to convert one into the other. A sensor could be calibrated to measure Candelas at a specific reference distance.

It also becomes more complex as light energy is both dependent on frequency, and Candela's are a weighed value modeled after the human eye. Most light sources also do not have an uniform distribution of light output. If you want to get serious about light measurement, then also consider (at least reading about):

https://en.wikipedia.org/wiki/Integrating_sphere

[Emanuel]

I think that you are right. The sensors are calibrated to a certain distance.
I believe that the reason for specifying different measurement distances such as in car lamps has to do with the light source intensity.
While signal lamps are measured at 3.1m, headlights are tested at 25m and warning lamps at 10m.
Same thing as with the size of the integrating spheres.
So practically if a product specification doesn't mentions the measurement distance I can do the test in an existing setup with a known distance, as long as the setup is good for the intensity range.
In this case the intensities are more or less like car signal lamps, so I can measure on the existing setup I have, at 3.16m.
Now I need to find a solution to measure infrared in mW/sr, this is quite challenging as this is not illuminance but radiance.
Any idea where can I find such a device?

[Someone]

Any idea where can I find such a device?

A calibrated spectrophotometer with an integrating sphere and known entrance pupil will do all that, there are various simplifications of that for specific domains/measurements but if you want universal then you get complexity.

When you say:

I have the necessary equipment

What do you have already? What is the budget?

[Emanuel]

A calibrated spectrophotometer with an integrating sphere and known entrance pupil will do all that, there are various simplifications of that for specific domains/measurements but if you want universal then you get complexity.

What do you have already? What is the budget?

As we make automotive lamps we have a goniophotometer with all the required sensors.
This will be OK for the first part of my question, about the white light we need to make, it can be measured like any car signal light at 3.16m.
I posted the question because I was concerned about the measurement distance which is not specified.
I forgot to mention about the lamp we are supposed to make, the specifications require a photometric test with given intensities required at certain left/right and up/down points, very much the same as for car lamps.
The same is required for the infrared lamp so I assume I will need to test also with the goniophotometer but with a different sensor.
I see many radiometers such as from International Light, but they all measure in mW/cm2 and I couldn't find a formula to convert to mW/sr, probably because it is a different measurement method.
You mentioned an integrating sphere which now rings a bell, I found in the company's junk an old spectroradiometer with a small, approx. 10cm integrating sphere with SMA interface.
I have no idea if it works and for sure is not calibrated.
I also saw somewhere some larger integrating spheres, 50cm and 100 cm from an old Chinese test system.
This is what I have, I am now trying to find something about the mW/sr measurement method.   

[Someone]

Sounds like you have an abundance of equipment and lack the staff with experience. Time to get out the text books and learn.

[jonpaul]

lamp, lights and LED illumination measurement is very well studied and documented.

Suggest the fine textbooks on light measurement, and meter/ sensor manufacturer specs and applications notes

See Gossen, Minolta Tektronix meters,
EGG, Analog Devices, diodes.

Most lighting labs have the integration spheres, meters, goniometers needed.

In case your firm needs,more instruments or measurement suggest that you rent a,lab facility with the appropriate  equipment and,analysis software.

J

[Emanuel]

Most lighting labs have the integration spheres, meters, goniometers needed.

Yes, but most of them are doing tests for regular illumination products.
I can't find one with capability of measuring infrared in mW/sr with a goniophotometer.
Quite a problem.

[Martinn]

I can't find one with capability of measuring infrared in mW/sr with a goniophotometer.

Well, you need a sensor head that is calibrated to power (within the wavelength range relevant to you) and not physiological units (lm, cd, lx).

The same is required for the infrared lamp so I assume I will need to test also with the goniophotometer but with a different sensor.
I see many radiometers such as from International Light, but they all measure in mW/cm2 and I couldn't find a formula to convert to mW/sr, probably

Look up the definition of sr. You can directly (for very small angles, which are typical for goniometer sensors) convert a sr solid angle figure into a flat surface area - if you disregard the curvature from the sr definition (which is perfectly acceptable for a 3m distance and a sensor covering a few mm diameter).

[jonpaul]

There are lots of IR sources and IR meters due to use for surveillance video and gunsight illuminations.


Check for old  Tektronix  J16 Digital Photometer  and many special probes, often cheap on ebay. Some have NiCd batts and many have irridance and other probes...

Just one

https://www.ebay.com/itm/234259094249

Bon Chance

Jon

[DavidKo]

cd (lm/sr) is distance independent, same as W/sr.

Standard goniometer have lx detector. cd are only calculated based on the distance of the detector (3.14m, 25m or completely different distance depending on setup/regulation requirement). You cannot measure cd, you can only measure cd/m^2.

As mentioned below, your setup is for visible spectrum ie. photometric measurements - you do not have capability for radiometric measurements. I have doubts that you will be able to switch easily the photometer for radiometer. You need to contact your goniometer supplier.

In LED datasheets for visible spectrum are lm (measured in integration sphere) and angular distribution. You can go around it on goniometer with measurement of the whole half sphere isomap and than use your SW to see the flux in the map (be careful with used step, I will propose to use the step which have the same size as your detector - you need to calculate it from distance and size of the detector). Same think you can do if you have goniometer with radiometric detector for IR spectrum.

[Emanuel]

Check for old  Tektronix  J16 Digital Photometer  and many special probes, often cheap on ebay. Some have NiCd batts and many have irridance and other probes...

Of course, there are many radiometers capable of measuring the IR intensity, but all are in W/m2 or mW/cm2.
There is no conversion from these units to mW/sr because it is a completely different measurement setup, where the distance, angle and surface of the detector are part of the calculation.

[Emanuel]

You can go around it on goniometer with measurement of the whole half sphere isomap and than use your SW to see the flux in the map (be careful with used step, I will propose to use the step which have the same size as your detector - you need to calculate it from distance and size of the detector). Same think you can do if you have goniometer with radiometric detector for IR spectrum.

You are right. So if I understand you right I need to find a radiometric sensor, first measure the total flux in an integrating sphere and then place the same sensor in front of the goniometer and calculate the relative intensity in every point?
Or make a scan with a step based on the sensor surface and let the goniometer calculate the total flux, indeed there is such a function in the goniometer's SW.

In fact I think that I only need to find a possibility to measure the precise intensity in mW/sr in a single point.
Preparing the full point map as required by the customer is possible by measuring the relative intensity at each point.

The problem is finding such a setup or instructions how to make it.

[DavidKo]

For point measurement I expect that you simply multiply the result with distance^2. Since the power meter will give you irradiance in W/m^2.

[Ground_Loop]

Check for old  Tektronix  J16 Digital Photometer  and many special probes, often cheap on ebay. Some have NiCd batts and many have irridance and other probes...

Of course, there are many radiometers capable of measuring the IR intensity, but all are in W/m2 or mW/cm2.
There is no conversion from these units to mW/sr because it is a completely different measurement setup, where the distance, angle and surface of the detector are part of the calculation.

Isn't mW/cm^2 directly convertible to mW/sr?

[DavidKo]

Yes, you multiply it with square of distance in cm. Same as for lx->cd.

You are not able directly measure the properties of light source itself. You can only measure illuminance/irradiance and rest you need to calculate.

[Martinn]

Yes, you multiply it with square of distance in cm. Same as for lx->cd.

Being pedantic I'd say this is an approximation. cm^2 refers to a flat surface, sr to a solid angle (spherical surface). For practical use the difference certainly is neglibible (particularly for a sensor a few mm in diameter in 3 m distance), but in my understanding this is mathematically incorrect.

[TimFox]

The approximation is explained by constructing a flat surface that subtends the same solid angle as the spherical surface at a constant distance.
There will be a small trigonometric correction, which is negligible for a small solid angle.

[DavidKo]

You are right, but who cares  , I think that the accuracy of LMT goniometers is 3%  . I do not expect much better for Optronic devices. Detector size on 3.14m is ~0,5° ie. deeply in paraxial approximation. On 25m it is negligible.

[TimFox]

If it be negligible, then I do not care.

[Emanuel]

You are right, but who cares  , I think that the accuracy of LMT goniometers is 3%  . I do not expect much better for Optronic devices. Detector size on 3.14m is ~0,5° ie. deeply in paraxial approximation. On 25m it is negligible.

I think that the goniometer is more accurate than 3%. Headlights measured at 25m are required to very precise measurements.
We have the Optronik AMS 200 and the specifications mention 0.01°.

[jonpaul]

generally light measurement accuracy of 5..10% is Considered good.

1..3 %'perhaps at NIST or other nationale standards labs.

jon

[Martinn]

If it be negligible, then I do not care.

The reason why I mentioned this is that I find it dangerous to memorize the information "you can convert cd to lx by some linear factor".
This statement is perfectly reasonable for any angle (like 0.00001 sr) you'd enconter in a goniometer setup.
For large solid angles you'd get huge errors.
So what you should store in your long term memory is "for very small solid angles (like with any goniometer setup) you can (with reasonable accuracy) convert lx to cd by approximating the spherical surface area with a flat one".

[Someone]

If it be negligible, then I do not care.

The reason why I mentioned this is that I find it dangerous to memorize the information "you can convert cd to lx by some linear factor".
This statement is perfectly reasonable for any angle (like 0.00001 sr) you'd enconter in a goniometer setup.
For large solid angles you'd get huge errors.
So what you should store in your long term memory is "for very small solid angles (like with any goniometer setup) you can (with reasonable accuracy) convert lx to cd by approximating the spherical surface area with a flat one".

There I was thinking the errors you were worried about was varying etendue, which is orders of magnitude more significant in the result than spherical vs planar areas (and not calculable/correctable as they are) and easily reduced by adding a longer throw to the measurement.

[Someone]

generally light measurement accuracy of 5..10% is Considered good.

1..3 %'perhaps at NIST or other nationale standards labs.

National/Primary standards are transferred at around 1% for absolute power/intensity, there is demand for a few % absolute accuracy and its not that uncommon to find.