What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Setup to Measure Nav/Strobe Intensity (candela)

Dan Langhout

Dan Langhout

Well Known Member
OK, I know there are some lighting experts out there who can give me an informed opinion on this.

I am trying to get a ballpark measurement of intensity (in candela) of an LED Position/Strobe light. Based on my understanding of the photometric units of measure (which is tenuous at best:rolleyes:) it would seem that you would measure the illuminance (lux or foot-candles) with a light meter and then multiply by the square of the distance from the sensor to the light (meters^2 or ft^2). to get the intensity value of the source (candela)?

Given that, here is a photo of my admittedly crude measuring set up:

Candela_Measuring_Setup_dimensioned.jpg


and the light meter I'm using:

Light_Meter.jpg



Note that the distance between the light meter sensor and the light is 1 foot, therefore my measurements in foot-candles should also equal the candela value (1^2=1) - right?

I take the measurements in the dark even though the meter has a relative setting to null out ambient light. I use the normal setting to measure the position light and the "peak" function to measure the strobe.

OK - Please - tell me what, if anything, is wrong with my assumptions or measuring setup. I'm not getting results even close to what I expect.
 
Last edited:
Are the steady state readings...

...of the nav light OK and the strobe reading "funny"?

I don't think that meter can do the averaging for a pulsating strobe source.

Just to check that the sensor is not saturating, move to 5 ft away and check the reading drops by 1/25.
 
Readings are both low

az_gila said:
...of the nav light OK and the strobe reading "funny"?

I don't think that meter can do the averaging for a pulsating strobe source.

Just to check that the sensor is not saturating, move to 5 ft away and check the reading drops by 1/25.
Click to expand...

Both the position light and strobe read much lower than I would expect. I have played with the sensor distance to see if the effect on the reading was "logical" (i.e. the reading increases or decreases ~ by the square of the distance). According to the manual for the meter, the "Peak" function enables the meter to capture pulses as short as 100 microseconds. Typically, I think LED strobe pulses are more like 100 milliseconds so I would expect for the meter to be able to capture the peak OK.

Obviously, what I'm trying to do here is to see if this light is in the neighborhood of meeting the FAR for a position and anti collision light (it's supposed to). Calculating the "effective" candela for the strobe is a little more complicated than just measuring the peak brightness (see FAR 23.1401) but the "peak" has to at least meet the minimum (400 cd) to even be within shouting distance of being compliant. The readings I'm getting are low by greater than a factor of 2 for the position and 8 for the strobe :eek:. I'm figuring that I must be doing something wrong here :(.
 
Time to contact...

...the manufacturer of your lights. A simple measurement test like your setup should get reasonable readings, and the shipped device should be designed with some margin over the FAA mandated minimums.

In previous VAF postings about LED lights, I've have asked if anyone has run actual tests, and I never really got a positive answer.

It all seems to be calculations based on the LED makers data sheets, which are based on a measured average, and don't often give the actual production variation range.

This paper gives a good description of the "effective" candela rating for the flashing bit.

http://www.periheliondesign.com/downloads/aircraft_beacons_using_leds.pdf

.....and you haven't even got to the directional coverage measurements yet...:)

Did you try running on 14.4 volts input?
 
az_gila said:
<snip>
In previous VAF postings about LED lights, I've have asked if anyone has run actual tests, and I never really got a positive answer.
</snip>
Click to expand...

Well . . . that's exactly what I'm trying to do here ("trust but verify" ;)) - I'm just not sure my measuring methodology or equipment is correct.


<snip>
This paper gives a good description of the "effective" candela rating for the flashing bit.

http://www.periheliondesign.com/downloads/aircraft_beacons_using_leds.pdf
</snip>
Click to expand...

I have read that paper and it gives a really good explanation. An important take-away from the paper and the FAA equation is that the actual intensity value has to be HIGHER than the minimum in order to meet the EFFECTIVE intensity value described by the equation

<snip>
.....and you haven't even got to the directional coverage measurements yet...:)
</snip>
Click to expand...

Absolutely right! :(

<snip>
Did you try running on 14.4 volts input?
</snip>
Click to expand...

I did - the lights internal power supply does an excellent job of keeping the power constant (the input current goes down as the voltage goes up).
 
Nothing wrong with your set up, but...

LED manufacturers publish lumen outputs for "ideal" junction temperatures. Reality is that when you put the LED die into anything that makes it a useful fixture, then put it into any reasonable environment, your lumens will be substantially depreciated. I would not be surprised that you are getting half of what you would expect to see since you are expecting to see what the manufacturer publishes.
The IESNA, Illuminating Engineers Society of North America, of which I have been a member of for more than 20 years, is nearly complete with new standards which will require testing of complete systems and fixtures. This is supposed to be out very soon.
All that said, illuminance is really not that important for the application of strobes and nav lights. What is important is surface brightness. LED's appear bright because they generate their light in a very tiny point. It looks bright, but does not really put out much useable light. However, the appearance of brightness is what is important for you to see someone to avoid a collision. Who cares how much light is actually leaving the fixture. If it is bright to the eye, you can see it. There are meters that test for surface brightness. They measure in foot lamberts. Unfortunately, the FAA does not test to that standard because that would be the smart thing to do.
As far as your light meter goes, the lensing system in light meters vary. A good one is set up to follow the spectrum spread that is most visible to the average human eye. LED's have trouble providing light in wide spectrums as they are essentially monochromatic, but then use a phosphor, or multiple color leds, to try to compensate for better color across a wider spectrum. Light meters may or may not represent accurately what you see with your human eye.
LED technology is changing at a very high rate. Two years from now, you will have products at least twice as bright, twice as good, and half the cost.
 
Dan Langhout said:
Both the position light and strobe read much lower than I would expect.
Click to expand...

I would expect so. The test device is going to do some filtering/averaging to get a stable reading (dv/dt). Just off the top of my head you would need a constant current source, a phototransistor, a digital storage oscilloscope, and come up with a way to calibrate everything. Its a different ballgame to quantify a light source with short duration. You could use the light meter to calibrate your test setup.
 
Last edited:
Yes...

rocketbob said:
I would expect so. The test device is going to do some filtering/averaging to get a stable reading (dv/dt). Just off the top of my head you would need a constant current source, a phototransistor, a digital storage oscilloscope, and come up with a way to calibrate everything. Its a different ballgame to quantify a light source with short duration. You could use the light meter to calibrate your test setup.
Click to expand...

....but Dan can't get the steady state reading to agree with the advertised (or is it implied?) numbers.

The meter should be reasonably accurate for that measurement mode.
 
I see you have that com radio sitting there and I assume your checking for RFI interference???

What is the result on that if you are?
 
Ahhh . . . .

<snip>
All that said, illuminance is really not that important for the application of strobes and nav lights. What is important is surface brightness. LED's appear bright because they generate their light in a very tiny point. It looks bright, but does not really put out much useable light. However, the appearance of brightness is what is important for you to see someone to avoid a collision. Who cares how much light is actually leaving the fixture. If it is bright to the eye, you can see it. There are meters that test for surface brightness. They measure in foot lamberts. Unfortunately, the FAA does not test to that standard because that would be the smart thing to do.
</snip>
Click to expand...

That makes perfect sense! This light certainly looks bright. But I assume from what you said that the FAA testing method would be more along the lines of what I am trying to do?

The light I am testing is commercially available and the manufacturer has stated that it has been tested in compliance with FAA requirements and has been submitted for TSO approval (no specific lumen or candela values are given). I could just take them at their word (and wait for the paperwork) but it seemed to me that I should be able to independently get test results that were at least in the same universe with the specs. Given that my results so far are not even close, I am assuming that there is a problem with my measurements/calculations.

<snip>
As far as your light meter goes, the lensing system in light meters vary. A good one is set up to follow the spectrum spread that is most visible to the average human eye. LED's have trouble providing light in wide spectrums as they are essentially monochromatic, but then use a phosphor, or multiple color leds, to try to compensate for better color across a wider spectrum. Light meters may or may not represent accurately what you see with your human eye.
</snip>
Click to expand...
My light meter is just a garden variety unit used to measure light levels in the workplace, etc. The specifications say that it provides "color and cosine corrected measurements". The sensor seems to be covered with a white translucent dome. Perhaps it's not suitable for this sort of measurement?
 
az_gila said:
....but Dan can't get the steady state reading to agree with the advertised (or is it implied?) numbers.

The meter should be reasonably accurate for that measurement mode.
Click to expand...

That's if you believe the meter. I can show you four different meters I have, two cheap ones of Extech quality and two Flukes (one's a scopemeter.) The Flukes will read voltages almost exactly, and they've been calibrated. The cheap ones will read several volts off in AC, and a few mV off DC. I learned not to trust non-Fluke instruments after that, for precise measurments.
 
Values to be met

az_gila said:
....but Dan can't get the steady state reading to agree with the advertised (or is it implied?) numbers.

The meter should be reasonably accurate for that measurement mode.
Click to expand...

No numbers are given for this light - just that it is compliant with FAA requirements (see my previous post).

For this application (rear position / strobe) the "straight on" intensity for the position light needs to be 20 candela. With my current setup, I measure 6.3 foot-candles at a distance of 12" which I think is the same as 6.3 candela - just a bit low :rolleyes:.
 
Good Eye!

Brantel said:
I see you have that com radio sitting there and I assume your checking for RFI interference???

What is the result on that if you are?
Click to expand...

That seems to be the good news! As far as I can tell, this thing is absolutely quiet. The light is run off of the same power supply as the COM. I have tuned to weak frequencies and even held the antenna next to the light with no apparent difference with the light turned on or off.
 
I would disagree....

JonJay said:
LED manufacturers publish lumen outputs for "ideal" junction temperatures. Reality is that when you put the LED die into anything that makes it a useful fixture, then put it into any reasonable environment, your lumens will be substantially depreciated. I would not be surprised that you are getting half of what you would expect to see since you are expecting to see what the manufacturer publishes.
Click to expand...

Yep... and it would not then meet the required specifications....

.......
All that said, illuminance is really not that important for the application of strobes and nav lights. What is important is surface brightness. LED's appear bright because they generate their light in a very tiny point. It looks bright, but does not really put out much useable light. However, the appearance of brightness is what is important for you to see someone to avoid a collision. Who cares how much light is actually leaving the fixture. If it is bright to the eye, you can see it. There are meters that test for surface brightness. They measure in foot lamberts. Unfortunately, the FAA does not test to that standard because that would be the smart thing to do.
Click to expand...

That I disagree with. It might be true for household light fixtures, but the aircraft lights are designed to be seen from 1000's of feet away. At these distances the entire light housing would like like a point source, for both LEDs and incandescent... the actual area that is emitting light would not be relevant. The light that gets to your eyeball is what counts.

For aircraft exterior lighting the FAA method is reasonable.

As far as your light meter goes, the lensing system in light meters vary. A good one is set up to follow the spectrum spread that is most visible to the average human eye. LED's have trouble providing light in wide spectrums as they are essentially monochromatic, but then use a phosphor, or multiple color leds, to try to compensate for better color across a wider spectrum. Light meters may or may not represent accurately what you see with your human eye.
LED technology is changing at a very high rate. Two years from now, you will have products at least twice as bright, twice as good, and half the cost.
Click to expand...

The light meter shown does reasonably follow the correct CIE photopic spectrum required for foot-candle measurements. As you know, this is an easy item to perform with a filter calibrated to the well known, and fixed, spectral response of a silicon photo-diode.
This was an easy task in the mid-70's when I was chief engineer of the instrument division of United Detector Technology - I'm sure things have only got better since then...:)
The harder task was getting the sensor to have a true cosine reponse for off-axis light sources - but that does not come into play here.

If you checked the light meter data sheet/user manual, you could see here...

http://www.extech.com/instruments/resources/manuals/EA30_UM.pdf

As I said before, check with the manufacturer about his claim that it meets FAA requirements....:rolleyes:
 
Last edited:
I'm going to go out on a limb here and say that I think your test setup is not correct.

When I looked into this same measurement, I remember coming to the conclusion that I needed to setup a sheet with a known reflectance at constant radius from the emitter (ie: a 1m radius semi-sphere). You would then measure the light reflected off of the sheet, compensating for its losses.

I think the problem with your approach is that your light meter sensor effectively reports the average light intensity over a circle of some diameter. Since the light source is much smaller than the sensor field of view circle, the average intensity measured will be way less. I think that the reflected light measurement setup outlined above effectively eliminates this problem, since over a small area the amount of light reflecting off the sheet is fairly uniform.

Quick edit: When you do this test, make sure to take into account not only the distance between the source and the sheet but also the distance between the sheet and the sensor. I think it would work best if the sensor is very close to the sheet and held as perpendicular as possible without shadowing. You might try a sanity check with a relatively isotropic known source like a clear 60W light bulb.

If you haven't done so already, I highly recommend reading the actual FAR requirements paying particular attention to the units specified. Then google those units and make sure you exactly understand what they mean. The distinctions between the various units describing optical flux, intensity, illuminance, etc can be subtle. Try this article for a start:
http://en.wikipedia.org/wiki/Candela#SI_photometric_light_units

Good luck,
-DC
 
Last edited:
That's a bit global...

:)
rocketbob said:
That's if you believe the meter. I can show you four different meters I have, two cheap ones of Extech quality and two Flukes (one's a scopemeter.) The Flukes will read voltages almost exactly, and they've been calibrated. The cheap ones will read several volts off in AC, and a few mV off DC. I learned not to trust non-Fluke instruments after that, for precise measurments.
Click to expand...

...in trashing meters....:)

However, the meter says about 3% or 4% accuracy. Even if that is off by a factor of 2 or 3 times, the reading obtained should be in the ballpark for the easier to measure steady state position light mode.

I also would hope that the manufacturer of the position light built some margin into his design, so that if Dan has some measurement error, the design margin would be reduced, but the reading would be close to the FAA required number.

An easy question to ask the manufacturer would be "do you perform any light output tests on all assemblies shipped"?

This steady state test is reasonably easy one to perform.
 
Gut Check Calibration?

Dan,

Maybe as a quick calibration check if you can get hold of a "certified" aircraft strobe and measure it with your existing setup and compare results to your LEDs. Use the relative numbers to see if your in the ballpark?

Gregg
 
Not a bad idea . . . . . .

fly3g said:
Dan,

Maybe as a quick calibration check if you can get hold of a "certified" aircraft strobe and measure it with your existing setup and compare results to your LEDs. Use the relative numbers to see if your in the ballpark?

Gregg
Click to expand...

. . . . . . but after doing a good bit of studying and thinking, I'm almost of the opinion that I am using the wrong instrument for this measurement. Typical light meters (mine included) are "cosine corrected" which is what you want for surface light level measurements from potentially many sources (illuminance), but probably not what you want to measure luminance levels. The value I'm interested in is luminous intensity expressed in candela. Illuminance and luminance both are functions of luminous intensity but apparently are measured differently. I have this suspicion that that the cosine correction is giving me an error in this case. I still need to think about this some more and continue to dig.
 
I don't think so....

Dan Langhout said:
. . . . . . but after doing a good bit of studying and thinking, I'm almost of the opinion that I am using the wrong instrument for this measurement. Typical light meters (mine included) are "cosine corrected" which is what you want for surface light level measurements from potentially many sources (illuminance), but probably not what you want to measure luminance levels. The value I'm interested in is luminous intensity expressed in candela. Illuminance and luminance both are functions of luminous intensity but apparently are measured differently. I have this suspicion that that the cosine correction is giving me an error in this case. I still need to think about this some more and continue to dig.
Click to expand...

Consider a commercial application of light levels on a work bench.

Measured in ft.-candles in the US.

If you had a single light on the ceiling straight above you, it would be an accurate measure of the light emitted from that bulb and falling on your work surface. Work out the "feet" part and you can deduce the "candles" (candela) bit as you mention in your first post.

All the cosine response does is allow the light that falls on the detector from a bulb over the adjacent work bench to be added to the total amount of light on your work bench.

The value added is based on the cosine of the angle to the other light source. All the cosine response does is allow all of the light sources that create light that falls on your work bench to be summed into the ft-candle measurement on your bench. This makes sense, since it's all light that you can work by, but the light overhead is weighted more than any light off to the side.

You are (should be...:)..) working in the dark - or subtract any ambient light - with a single light source and you sensor is shipped calibrated in an "end on" position.

Try a 10 ft base line and re-measure - this should make your nav lite effectively a point source, and just move your measured readings decimal point 2 over...:)
 
my 2 c...

My PERSONAL opinion is that many "meet FAA reqs" LED systems out there simply do NOT meet FAA reqs. Take a regular tail light with the required 20 watt bulb and test it in your setup. I bet it will read between 30 and 50 foot-candles (it did for me).
In terms of perceived brightness I agree the LEDs are great, but I am not sure if a very small, yet very bright spot has the same visibility at a certain distance.
Now, for NAV lights, LEDs may be OK, because you need NAV lights at night, but there is NO WAY I will put a LED-based anticollision light in my airplane. Remember, you need to be able to see anticollision lights in full daylight too.
AGAIN, just my 2 c.
 
New Measurements

az_gila said:
<snip>
You are (should be...:)..) working in the dark - or subtract any ambient light - with a single light source and you sensor is shipped calibrated in an "end on" position.

Try a 10 ft base line and re-measure - this should make your nav lite effectively a point source, and just move your measured readings decimal point 2 over...:)
</snip>
Click to expand...

OK - I set the light up oriented to shine horizontally about 4 feet above the floor and extending out over the edge of the table about 12". I then laid a measuring tape on the floor extending out to 15' from a plumb line dropped from the light. Then, with absolutely no light in the shop except the unit under test (windowless basement shop at night), I took measurements at 5' 10' and 15' - 4 measurements at each location. I used the "MAX" function of the light meter to try and capture the highest reading. I moved the sensor around in the vertical plane but attempted to keep the sensor at the proper distance.
For these test, the sensor was hand held so the distance measurements could be +- 3 to 4 inches.

I averaged the four measurements at each location and then calculated intensity in candela as average-foot-candles/distance^2.

Here's the data:

5' -- 0.35 0.36 0.36 0.35 -- 8.875 candela
10' -- 0.22 0.17 0.16 0.19 -- 18.5 candela
15' -- 0.17 0.12 0.11 0.09 -- 27.56 candela

That's a pretty large variation. The answer looks better the farther away I get but I guess I don't have much confidence in them. :(
 
I'm guessing...

Dan Langhout said:
OK - I set the light up oriented to shine horizontally about 4 feet above the floor and extending out over the edge of the table about 12". I then laid a measuring tape on the floor extending out to 15' from a plumb line dropped from the light. Then, with absolutely no light in the shop except the unit under test (windowless basement shop at night), I took measurements at 5' 10' and 15' - 4 measurements at each location. I used the "MAX" function of the light meter to try and capture the highest reading. I moved the sensor around in the vertical plane but attempted to keep the sensor at the proper distance.
For these test, the sensor was hand held so the distance measurements could be +- 3 to 4 inches.

I averaged the four measurements at each location and then calculated intensity in candela as average-foot-candles/distance^2.

Here's the data:

5' -- 0.35 0.36 0.36 0.35 -- 8.875 candela
10' -- 0.22 0.17 0.16 0.19 -- 18.5 candela
15' -- 0.17 0.12 0.11 0.09 -- 27.56 candela

That's a pretty large variation. The answer looks better the farther away I get but I guess I don't have much confidence in them. :(
Click to expand...


...that reflections off the floor are increasing the light hitting the sensor at the further distances.

Repeat the tests with a toilet paper tube internally lined with a scrap of black velvet, and point fairly accurately at the nav lite from the three distances. This should give more consistent readings.

As a very quick sanity check, set up the 15 ft reading with the sensor fixed on a tripod, and spread something on the floor between the light and the sensor (a blanket or drop cloth - something fairly large) and see if it changes the 15 ft reading measureably....

The cosine sensor will add the reflected light into your reading, upsetting the ft-candle calculations....:(
Any errors created by reflections would be in the direction you are seeing.

PS - basement - low ceiling - reflections from that side too...
 
Last edited:
New Measurements - quick redux

Gil,
I just quickly tried the toilet paper tube thing (no black velvet). Got .04, .07, and .25 at 15', 10', and 5'. Yields 9.0, 7.0 and 6.25 candela. Much better on a percentage basis but . . . . . . .:(
 
Try a longer...

Dan Langhout said:
Gil,
I just quickly tried the toilet paper tube thing (no black velvet). Got .04, .07, and .25 at 15', 10', and 5'. Yields 9.0, 7.0 and 6.25 candela. Much better on a percentage basis but . . . . . . .:(
Click to expand...

...tube.

It does seem like reflections were coming into play. The numbers should start converging on the 6 value...:)
 
Photometric Testing

Not that easy to accomplish. When a commercial fixture is tested to IES standards it is put into a completely black room, temperature controlled to a set standard, jigged up in a fixture and either rotated around the sensor or the sensor is rotated around the product. Points are taken around all axis of the fixture to establish the photometric file data. That raw data can then be taken into a variety of programs for evaluation.
The system is automatic and nobody is allowed in the room. Testing can take hours to accomplish.
I know this is way more complicated than necessary to get a rough idea of what you might get from these LED lights but there are lots of opportunities to introduce errors from reflectances from surfaces, off of you shirt, etc...
 
Yes...

JonJay said:
.....
The system is automatic and nobody is allowed in the room. Testing can take hours to accomplish.
I know this is way more complicated than necessary to get a rough idea of what you might get from these LED lights but there are lots of opportunities to introduce errors from reflectances from surfaces, off of you shirt, etc...
Click to expand...

...and the errors you mention all make the "unit under test" appear to emit more light - the black room absorbs light on every surface.

Dans measurements are showing less output than expected.

Again..."ask the manufacturer if he 100% tests the nav lights for output"...
 
JonJay said:
Not that easy to accomplish. When a commercial fixture is tested to IES standards it is put into a completely black room, temperature controlled to a set standard, jigged up in a fixture and either rotated around the sensor or the sensor is rotated around the product. Points are taken around all axis of the fixture to establish the photometric file data. That raw data can then be taken into a variety of programs for evaluation.
The system is automatic and nobody is allowed in the room. Testing can take hours to accomplish.
I know this is way more complicated than necessary to get a rough idea of what you might get from these LED lights but there are lots of opportunities to introduce errors from reflectances from surfaces, off of you shirt, etc...
Click to expand...

I'm not trying to provide certification testing here :rolleyes: - all I was really wanting to get was a warm fuzzy that this light was at least close to meeting the FAA requirements for its intended use. If my measurements were within 15%-20% of the requirements then I could easily believe that my relatively crude methods/equipment was responsible for the discrepancy. But when the numbers are off by a couple of hundred percent, then I have to figure that either I'm all wet (or the manufacturer is). Now I'm an engineer but lighting is certainly not my specialty - so I've been going with the assumption that I have completely misunderstood the problem or the techniques to at least get ballpark numbers. I'm trying to give the manufacturer the benefit of the doubt here - they claim that these will be certified/TSO'd lights although there is no paperwork yet :confused:.
 
Ok Dan

1. You're at home instead of at work with a brand new fuse kit and you're GOOFING WITH YOUR LIGHTS. :eek:

2. Turn off the lights and start working on the fuse! :D:D:D
 
If my inspector questions if my homemade nav lights meet the brightness specs, I am going to do this...

Mr. inspector stand right here and look in this direction while I turn them on:

2mg1he0.jpg


Ok now look away sir, do you have any additional questions? :D
 
I saw what you used . . . .

Brantel said:
If my inspector questions if my homemade nav lights meet the brightness specs, I am going to do this...

Mr. inspector stand right here and look in this direction while I turn them on:

Ok now look away sir, do you have any additional questions? :D
Click to expand...


. . . . for your nav lights. I suspect you left the minimum candela far behind! :D
 
Thats why the new style wingtip is great for these kinds of lights, it naturally creates a shadow on the opposite side. No problems with overlap even when the wingtips are placed back to back. That was one of my eyeball tests back when I was fooling with these things.

az_gila said:
...exceeded the maximum in the "overlap" area where another plane is meant to distinguish colors to get direction....:)
Click to expand...
 
No problem...

Dan Langhout said:
I'm not trying to provide certification testing here :rolleyes: - all I was really wanting to get was a warm fuzzy that this light was at least close to meeting the FAA requirements for its intended use. If my measurements were within 15%-20% of the requirements then I could easily believe that my relatively crude methods/equipment was responsible for the discrepancy. But when the numbers are off by a couple of hundred percent, then I have to figure that either I'm all wet (or the manufacturer is). Now I'm an engineer but lighting is certainly not my specialty - so I've been going with the assumption that I have completely misunderstood the problem or the techniques to at least get ballpark numbers. I'm trying to give the manufacturer the benefit of the doubt here - they claim that these will be certified/TSO'd lights although there is no paperwork yet :confused:.
Click to expand...

Just trying to help. If you can get a photometric file from the manufacturer, that would go a long way toward their credability. I could take that file and run it through a program to show (theoretically) what you should get on paper if that would help. The only problem is that I am not sure the standard has been finished on how the testing is to be done so if they have a file, it may not have been tested correctly anyway. The only real test would be a field test, and that is why I was trying to point out how difficult that would be to do correctly in my post. I would be interested in seeing how the FAA does the testing, or even if they do, or if it is just a hold over from the "olden days" of incandescent lamps. No other point inferred or implied.
Anyway, it seems a lot of folks have been struggling with quantifying the LED's in this application. In our industry, they are being hailed as the next coming. However, the reality of using them in real applications, cost, misleading or false claims, heat issues, are all keeping the technology from being implimented more. This is primarily in "white light" applications that bring up a whole host of other issues.
The transportation industry has been using colored LED's successfully for a long time in marker lights and more critical brake and signal lights. As Brian stated, you can just look at them and know they should work, but as Gil states, "not so fast". Not putting words in either of their mouths.
Hope you having fun with this anyway.
 
I appreciate the input!

Jon,
I didn't mean to imply that I was annoyed - far from it! ... and I am having fun! I just have to be careful not to get dragged off down any of the really interesting and fun side paths that unfortunately can make a long plane building project into an infinitely long plane building project. :)

I contacted the manufacturer - he has committed to actually testing another unit and sending it to me along with a printout of the test results from their lab (I assume this is the "photometric file" you referred to?). He also mentioned that you can't get accurate results measuring LEDs with normal light meters (apparently the spectrum of the white LEDs in particular run afoul of the normal CIE filter used in these meters?). Does that sound reasonable? At any rate, we'll see what they send me.

I haven't been able to find anything that describes how the FAA regs are tested for other than two test devices that are used by repair stations to verify that the strobes are within spec. One is the ST-4000 by American Aerospace Corp. and the other is a unit made by Goodrich Lighting (http://www.goodrich-lighting.com/catalog/Chapter06_Anti-Collision_Lights/06_0100_GoodCam.pdf that has specific provisions for testing traditional as well as LED aircraft strobes. Interestingly, the ST-4000 unit puts the sensor on the end of an 8 foot wand to control the measuring distance and then takes the data over 10 pulses. The Goodrich unit apparently uses a different equation to do the effective intensity calculation (sort of a modified Blondel-Rey) in order to correctly measure LED strobes that have a different flash pattern.
 
Last edited:
Steradian

Steradian.

FAA AC 43-13-2. Wikipedia explanation sorta makes sense to this old Retired USAF Vet.

I have Bill Dube's "Killacycle" LED Nav Light Kit from the 2004 era. The kit was complete and instructions were good including the light measurements. He gave measurements in foot candles for all the angles around the light's intended projections.

The kit met all the specs using a high end photographic light meter.

From my old Photo School days, light can be measured in several ways. The 4 that I can think of now are" Foot Candles, Candle Power,(close to the same), Lumen, Lux and Candela. Be sure what you are measuring and use the proper measuring device to measure it.

Back to Steradians; this is where placement of strobes or other position lights come into play. What I believe they are saying is; your plane has to have the minimum candela (etc) be visible at least at the minimum angles (steradians) stated in the Advisory Circular.

Dang; This sorta got outa my league in a hurry!:eek:
 
Last edited:
No problem....

Dan Langhout said:
Jon,
I didn't mean to imply that I was annoyed - far from it! ... and I am having fun! I just have to be careful not to get dragged off down any of the really interesting and fun side paths that unfortunately can make a long plane building project into an infinitely long plane building project. :)

I contacted the manufacturer - he has committed to actually testing another unit and sending it to me along with a printout of the test results from their lab (I assume this is the "photometric file" you referred to?). He also mentioned that you can't get accurate results measuring LEDs with normal light meters (apparently the spectrum of the white LEDs in particular run afoul of the normal CIE filter used in these meters?). Does that sound reasonable? At any rate, we'll see what they send me.

I haven't been able to find anything that describes how the FAA regs are tested for other than two test devices that are used by repair stations to verify that the strobes are within spec. One is the ST-4000 by American Aerospace Corp. and the other is a unit made by Goodyear Aerospace that has specific provisions for testing traditional as well as LED aircraft strobes. Interestingly, the ST-4000 unit puts the sensor on the end of an 8 foot wand to control the measuring distance and then takes the data over 10 pulses. The Goodyear unit apparently uses a different equation to do the effective intensity calculation (not Blondel-Rey) because it's "better" :eek:
Click to expand...

I suspected that the Light Meter may not perform well. Colored LED's are monochromatic. or at least they are for the simple colors we use in aviation. Light meters are designed for "white light", and if they are done correctly, across a spectrum that is reasonable for the human eye or "visible light". I could send you reams of paper on it, but that would just tie you down even more. Of course we have not even started to talk about scotopic and photopic vision, percption of color, motion, focal point, peripheral vision, or the other aspects of what makes us able to see something.
I would be curious to see the manufacturers test results. I would be even more interested in seeing their test procedure and equipment used. I really doubt they are testing it to a published standard of any type or using certified or even calibrated equipment. They are probably field testing it with their own procedures. I could be wrong, wont be the first time.
Anyway, I am annoying myself now. I have an airplane to build too, but I try to help in these forums where I feel I am qualified to do so.
I usually end up learning something too. Take care and let us know what you find. Keep on pounding rivets!!
I will try to find the FAA's regs. Had them at one time, but they are not at hand right now. They are very prescriptive, but I stand on that they are outdated for the understanding we have today about vision. Better systems can be designed with better standards for testing. MHO
 
Nah... again....

JonJay said:
I suspected that the Light Meter may not perform well. Colored LED's are monochromatic. or at least they are for the simple colors we use in aviation. Light meters are designed for "white light", and if they are done correctly, across a spectrum that is reasonable for the human eye or "visible light". I could send you reams of paper on it, but that would just tie you down even more. Of course we have not even started to talk about scotopic and photopic vision, percption of color, motion, focal point, peripheral vision, or the other aspects of what makes us able to see something.
....... MHO
Click to expand...

The CIE defines the response of the human eye, and that is what the photmeters are calibrated to with the filter in front of the silicon detector. The filter response is well calibrated, and the silicon detector response is defined by physics.

Remember, the present discussion is only for a white nav light in it's steady state, the easiest measurement - no getting into flashes...:)

The white LED is actually several seperate colors that emit together and appear to be white to the human eye. The error I think you are talking about is a second order effect and is "how well does the filtered silicon detector actually follow the human eye - as definied by the CIE - response?" This error is only a few percent for a good detector/filter combination - they have been making these exact things since the 60's and it's pretty well known by now - I created lots of curves testing different filters in the 70s. The curves were all pretty close, even then back in the dark, pre-LED ages.....:)

Dan is seeing a low output - and I bet his numbers are getting close. Since the eye does not have a linear response to brightness, his factor of three is not a noticeable as it sounds. However, it appears at this time that the white position light, in a steady state, is not up to FAA minimum requirements.

I wonder if the maker is measuring the position light while the strobe part is flashing and integrating the total light light output?

A good, detailed response from the manufacturer would be nice, even if it may be "we got a bad batch of LEDs"....

Of course we have not even started to talk about scotopic and photopic vision, percption of color, motion, focal point, peripheral vision, or the other aspects of what makes us able to see something."
Click to expand...

None of this is relavent to meeting the particular FAA requirement for position lights.... Let's stick to a simple measurement discusion of the candela produced by this particular light per the FAA definition....:)

PS - For anyone following this that might want to see the CIE curve with reference to the colors of the rainbow (what the eye sees) a nice picture is here....

http://www.marktechopto.com/products/FAQ-LEDs/difference-between-photometry-and-radiometry.cfm
 
Last edited:
az_gila said:
The CIE defines the response of the human eye, and that is what the photmeters are calibrated to with the filter in front of the silicon detector. The filter response is well calibrated, and the silicon detector response is defined by physics.

Remember, the present discussion is only for a white nav light in it's steady state, the easiest measurement - no getting into flashes...:)

The white LED is actually several seperate colors that emit together and appear to be white to the human eye. The error I think you are talking about is a second order effect and is "how well does the filtered silicon detector actually follow the human eye - as definied by the CIE - response?" This error is only a few percent for a good detector/filter combination - they have been making these exact things since the 60's and it's pretty well known by now - I created lots of curves testing different filters in the 70s. The curves were all pretty close, even then back in the dark, pre-LED ages.....:)

Dan is seeing a low output - and I bet his numbers are getting close. Since the eye does not have a linear response to brightness, his factor of three is not a noticeable as it sounds. However, it appears at this time that the white position light, in a steady state, is not up to FAA minimum requirements.

I wonder if the maker is measuring the position light while the strobe part is flashing and integrating the total light light output?

A good, detailed response from the manufacturer would be nice, even if it may be "we got a bad batch of LEDs"....



None of this is relavent to meeting the particular FAA requirement for position lights.... Let's stick to a simple measurement discusion of the candela produced by this particular light per the FAA definition....:)

PS - For anyone following this that might want to see the CIE curve with reference to the colors of the rainbow (what the eye sees) a nice picture is here....

http://www.marktechopto.com/products/FAQ-LEDs/difference-between-photometry-and-radiometry.cfm
Click to expand...

Sorry your missing the point Gil, or I am just bad at making it. Google "LED Light Measurement". You will find specific light meters designed for LED's (4 filters), and good discussions on Luminance vs Illuminance and why measuring techniques for LED's need to be different than traditional sources.
Your the meter expert, I am sure you can tell us why these meters have been designed specifically for LED's.
By the way, you need three colored LED's to make "white light".
In regard to the CIE, they have been holding advanced symposiums on this very topic, how to measure LED's. They have been holding them for a few years now and are still struggling with it. The next one is called "Physical measurement and visual and photobiological assessment". So, they think photobiological aspects are important, even if it does not relate to the FAA testing procedures. Lost on me for sure, but a lot of very smart people are still struggling with quantifying LED output and measurement. That is why I am sticking to the point that this is not easy to field verify compliance to the FAA standards.
I am sure you will find something to disagree with above, but I am getting used to it and have a tough skin. I am probably just not very good at writing my thoughts, or you just like me.:) Hopefully others are enjoying the post and learning a bit more about these ellusive LED's.
 
Last post...

JonJay said:
Sorry your missing the point Gil, or I am just bad at making it. Google "LED Light Measurement". You will find specific light meters designed for LED's (4 filters), and good discussions on Luminance vs Illuminance and why measuring techniques for LED's need to be different than traditional sources.
Your the meter expert, I am sure you can tell us why these meters have been designed specifically for LED's.
By the way, you need three colored LED's to make "white light".
In regard to the CIE, they have been holding advanced symposiums on this very topic, how to measure LED's. They have been holding them for a few years now and are still struggling with it. The next one is called "Physical measurement and visual and photobiological assessment". So, they think photobiological aspects are important, even if it does not relate to the FAA testing procedures. Lost on me for sure, but a lot of very smart people are still struggling with quantifying LED output and measurement. That is why I am sticking to the point that this is not easy to field verify compliance to the FAA standards.
I am sure you will find something to disagree with above, but I am getting used to it and have a tough skin. I am probably just not very good at writing my thoughts, or you just like me.:) Hopefully others are enjoying the post and learning a bit more about these ellusive LED's.
Click to expand...

...I believe you are delving into the LED efficiency and human perception parts for residential lighting (the "green" society) - which is very important to the LED makers.

Here, we are talking about an "indicator" (a very bright one...:)...) with one purpose only - to be seen at a long range.

I did the google search you said, and the #2 link went here... and then to a link to a recommended light meter for LEDs for measuring indicators in candela...

http://www.intl-lighttech.com/products/detectors/il700systems/sed033yled

...absolutely no difference from the meters (except for the USB interface, which hadn't been invented yet) that I was making 30 years ago. Silicon detector technology has changed little, and the physics is the same, and glass filter technolgy is similar.

The complexities come into the efficiency (total light output), and thus power savings, of LED products for commercial lighting use - a vast market.

Our steady state nav light is nowhere close to these complex human perception issues.

I think you are complicating the test that Dan is doing with other issues and second order measurement effects. A 5% error in Dan's test is nothing, but is a giant deal if you are selling LEDs on a "green" energy saving basis to light your home.

You might notice that the sensor I linked to above (specifically recommended, via your Google links, for candela measurement) has a baffled tube ahead of the sensor. This is simply a better version of the black felt lined cardboard tube I mentioned earlier, which would have a similar effect to reduce stray light.

I would like to see Dans numbers with the felt lined tube, and yes, I do believe that it does call into question a lot of the experimental LED systems being sold as "meets the FAA requirements" - which is not optional, but spelt out in the FARs.

Added...

The #3 Google link went here.. which shows Dan's setup in simple pictorial terms - from an LED web site....

http://www.led.net/pages/tech4.htm
 
Last edited:
TSO C30c and C96a . . . . .

. . . . appear to be the applicable documents for offically "compliant" position and anti collision lights respectively. There are a couple of interesting parts in both TSOs.

From C96a:

(1) Minimum Performance Standard. This technical standard order (TSO) prescribes the minimum performance standards that anticollision light systems must meet in order to be identified with the applicable TSO marking. Anticollision light systems that are to be so identified and that are manufactured on or after the date of this TSO must meet the minimum performance standard of the Society of Automotive Engineers Inc. (SAE), Aerospace Standard (AS) 8017A, ?Minimum Performance Standard for Anticollision Light Systems,? dated January 30, 1986.

. . . . . . .

e. Color and Intensity Specifications. The chromaticity coordinates contained in AS 8017A, paragraph 3.4.1 for aviation red and paragraph 3.4.2 for aviation white provide an equivalent level of safety to those specified in section .1397 of FAR Parts 23, 25, 27, and 29. For national and international standardization these coordinates have been made identical to the existing International Civil Aviation Organization requirements. This also applies to specifying intensity in terms of ?candelas? in AS 8017A as opposed to ?candles? in sections .1407 of FAR Parts 23, 25, 27 and 29. In practical application all existing anticollision light installations meet both the TSO and FAR requirements.
Click to expand...

and from C30c:

(1) Minimum Performance Standard. This technical standard order (TSO) prescribes the minimum performance standards that aircraft position lights must meet in order to be identified with the applicable TSO marking. New models of aircraft position lights that are to be so identified and that are manufactured on or after the date of this TSO must meet the minimum performance standard of the Society of Automotive Engineers, Inc. (SAE), Aerospace Standard (AS) 8037, ?Minimum Performance Standard for Aircraft Position Lights,? dated January 1986.

. . . . . . . . .

e. Color and Intensity specifications. The chromaticity coordinates contained in AS 8037, paragraph 3.3.1 for aviation red, green, and white provide an equivalent level of safety to those specified in Section .1397 of FAR Parts 23, 25, 27, and 29. For national and international standardization these coordinates have been made identical to the existing International Civil Aviation Organization requirements. This also applies to specifying intensity in terms of ?candles? in AS 8037 as opposed to ?candles? in Sections .1391 and .1395 of FAR Parts 23, 25, 27, and 29. In practical application all existing aircraft position light installations meet both the TSO and FAR requirements.
Click to expand...

Note that both of these reference SAE documents (AS 8017A and AS 8037) for the performance standards and particularly for an "alternate" specification/definition of color and intensity. I have no idea how much difference there is between these documents and the FARs since you have to buy these from SAE (~$50 each :eek:).

In addition, there is a FAA Memorandum from June 2007 that is an equivalent level of safety finding (ELOS) #ST4775WI-T-SE-1 for Learjet on a Whelen 7101102 rear position light which is an all LED unit. The issue was that the white light didn't quite meet the specific chromaticity of the FAR. Bottom line is that the FAA approved the ELOS.

The swamp just keeps getting deeper and muddier ........ :(
 
az_gila;278229 I think you are complicating the test that Dan is doing with other issues and second order measurement effects. A 5% error in Dan's test is nothing said:
http://www.led.net/pages/tech4.htm[/url]
Click to expand...
Not meaning to complicate things and this application is certainly different than the commercial lighting challenges we have. I am not an engineer or a meter expert. Thanks for delving into the measuring aspects and it sounds like the meter companies are just marketing something specific that really isn't.
Perhaps you can help this discussion further;
To get "white light" we need three colored LED's. That is not full spectrum but very specific bands fooling us into seeing white. Is the light meter technology capable of measuring the relative instensity of each color and totaling them for a combined lumen output? Does it factor the colors intensities to match the human eye or is it just a raw number?
LED's are very directional. Are the lenses, or is the tube method, capable of collecting those and averaging them?
Perhaps some of this is producing the problems in field measurement methods?
 
ELOS

Dan Langhout said:
.
In addition, there is a FAA Memorandum from June 2007 that is an equivalent level of safety finding (ELOS) #ST4775WI-T-SE-1 for Learjet on a Whelen 7101102 rear position light which is an all LED unit. The issue was that the white light didn't quite meet the specific chromaticity of the FAR. Bottom line is that the FAA approved the ELOS.

The swamp just keeps getting deeper and muddier ........ :(
Click to expand...

I did not know there was even such a thing as ELOS. At least they are being reasonable. This does move us to the point that maybe the FAR's are outdated in regard to LED's and maybe we should be moving to a new standard in regard to tests and measurements for them.
Now, who knows what that level of demonstration cost the manufacturer! Thank God we are experimental.
 
OK - more

I was going to stop - but you ask some very specific interesting questions...:)

JonJay said:
Not meaning to complicate things and this application is certainly different than the commercial lighting challenges we have. I am not an engineer or a meter expert. Thanks for delving into the measuring aspects and it sounds like the meter companies are just marketing something specific that really isn't.
Perhaps you can help this discussion further;
To get "white light" we need three colored LED's. That is not full spectrum but very specific bands fooling us into seeing white. Is the light meter technology capable of measuring the relative instensity of each color and totaling them for a combined lumen output?
Click to expand...

Yes... the meter will just sum the three separate "spikes" in the spectrum. The light output is the area under the curve, modified by the CIE scaling for every color. Spikes at the outer edges are summed (I guess integrated might be a better term) at a much reduced value than the spike in the center (green) per the CIE curve. The second order effect I mentioned is just how close does the filter/detector combination come to at each color level - i.e. how accurate is the actual curve to the CIE curve. For a light bulb, all of the colors are integrated and the area under the curve is more important than specific points. With three specific LED spikes, then the filter curve accuracy at those three colors is more important - but it should be a second order effect.

Does it factor the colors intensities to match the human eye or is it just a raw number?
Click to expand...

That's what the CIE curve does, provides a "standard" definition of the human eye. The filter/detector combo matches this curve in it's spectral response. This is required for a foot-candle measurement.

LED's are very directional. Are the lenses, or is the tube method, capable of collecting those and averaging them?
Click to expand...

Not applicable - we are looking at a single point with the detector. To get the directional response you see in the IES curves for fixtures, the detector is moved around and a plot made of the candela reading at various angles to normal in multiple planes (usually just 2 planes for the IES charts).
The candela reading does not need averaging - assuming a non-flashing source.

Perhaps some of this is producing the problems in field measurement methods?
Click to expand...

I think it's mainly a total flux output issue wich is directly connected to power efficiency (being "green")
 
Thats...

JonJay said:
I did not know there was even such a thing as ELOS. At least they are being reasonable. This does move us to the point that maybe the FAR's are outdated in regard to LED's and maybe we should be moving to a new standard in regard to tests and measurements for them.
Now, who knows what that level of demonstration cost the manufacturer! Thank God we are experimental.
Click to expand...

...because the CIE also has a much more complicated chart that defines colors.

http://en.wikipedia.org/wiki/Image:CIExy1931.svg

It sounds like the Whelen Lear Jet solution was not quite "white" per the FAA definition - which is a stumbling matter for LED use in home lighting. The eye is very sensitive to colors, and whites are not all the same - look at cool white vs. warm white bulbs. People object to looking at things with a tinted light....:)

At least the FAA recognized that the specific "shade" of white had no effect on visibility from a distance - as long as it doesn't look like aviation red or aviation green...:)

The color shade does not affect the measurements we are talking about here for Dan. The foot candle meter just takes what it gets per the eye response.

I bet the FAA just defined colors based on a bright bulb and easily obtained red and green glass colors sometime in the 30's....:D

As a side note, I think the green LEDs may be off from the FAA definition, but with no real affect to what they look like to an observer.

We may be experimental, but I believe we do need to meet the FARs for lights in this case.

PS ... the wiki color article is interesting...

http://en.wikipedia.org/wiki/CIE_1931_color_space
 
Last edited:
az_gila said:
...because the CIE also has a much more complicated chart that defines colors.

http://en.wikipedia.org/wiki/Image:CIExy1931.svg

It sounds like the Whelen Lear Jet solution was not quite "white" per the FAA definition - which is a stumbling matter for LED use in home lighting. The eye is very sensitive to colors, and whites are not all the same - look at cool white vs. warm white bulbs. People object to looking at things with a tinted light....:)

At least the FAA recognized that the specific "shade" of white had no effect on visibility from a distance - as long as it doen't look like aviation red or aviation green...:)

The color shade does not affect the measurements we are talking about here for Dan. The foot candle meter just takes what it gets per the eye response.

I bet the FAA just defined colors based on a bright bulb and easily obtained red and green glass colors sometime in the 30's....:D

As a side note, I think the green LEDs may be off from the FAA definition, but with no real affect to what they look like to an observer.

We may be experimental, but I believe we do need to meet the FARs for lights in this case.

PS ... the wiki color article is interesting...

http://en.wikipedia.org/wiki/CIE_1931_color_space
Click to expand...


Thanks for hanging in there with us. I think it is an important, fun, and interesting topic of discusssion.
I agree in regard to meeting the FARs, I am just glad we dont have to pay to prove it through a certification process. I would bet there are a lot of systems in flying airplanes out there right now that don't meet them however. I have even seen some certified products stuck on bottoms of wing tips that could not possible work. Cheers.
 
White LEDs

There is a technical note on the Philips Lumileds web site that talks about measuring LEDs. They show an example of a white high powered LED (they don't show exactly which one) and how its spectrum falls relative to the CIE curve. What was interesting is that it had a fairly high peak down around 460nm, a peak around 660nm and a fair bit of spectrum from around 600nm to 750nm that is significantly above the CIE curve. These areas are in the "corners" of the CIE curve (and of course would get attenuated appropriately by a light meter's filter)
 
The curve is the eye response...

Dan Langhout said:
There is a technical note on the Philips Lumileds web site that talks about measuring LEDs. They show an example of a white high powered LED (they don't show exactly which one) and how its spectrum falls relative to the CIE curve. What was interesting is that it had a fairly high peak down around 460nm, a peak around 660nm and a fair bit of spectrum from around 600nm to 750nm that is significantly above the CIE curve. These areas are in the "corners" of the CIE curve (and of course would get attenuated appropriately by a light meter's filter)
Click to expand...

...not the required LED output. Higher peaks on the "outer edges" of the CIE curve are needed to get the eye to respond to those "outer" colors where the eye is not as sensitive - the R and V in the ROYGBIV of the rainbow. It does not directly affect the measurement.

There is an indirect measurement effect if the filter/detector curve does not match the CIE curve, but this should be a second order effect.

I believe if you take the color chart....

http://en.wikipedia.org/wiki/Image:CIExy1931.svg

...I think the white LEDs are three point sources in the corners of the area, and how they sum together puts thm in the middle (which is white) - how they are balanced defines the "shade" of white...:)
 
Correct, and very important over time....

az_gila said:
...
... how they are balanced defines the "shade" of white...:)
Click to expand...

They have not been able to figure out how to manufacturer consistent balanced combinations. What they do to overcome this is called "bining". They check each LED White component block against a standard, whether that is warm, or cool, or in between. It is then "binned" according to where it falls. Manufacturers can then order from that "bin". The standard is pretty wide per bin and the gaps between bins pretty narrow, so if you are using an array of "white" LED's, one can fall on one side and another fall on the other where you could see the difference even in the same bin. This matters a lot in commercial lighting where multiple fixtures or multiple LED's in arrays may be used to light a wall, or valance. You can see the color difference. I would not think it matters much in this application, but....
As LED's age, the different colors age differently. There are some very tight pattents on electronics that use internal sensors, sleeper LED's, and other methods to compensate and keep the output and color temp the same. These are expensive technologies owned by a small handfull of players, like Philips, and Sylvania, and Cree. Why this might matter is, over time, if these technologies are not incorporated, we might expect to see a shift in color of our "white" LED's. Our "white" position light could start to take on a hue of color.
What I am not sure of is how many hours of burn time it would take to see the change. I would suspect that in this application it would be many, many years since we are dealing with a 50,000 hour source. I am not sure where the shift would start and be noticeable. So, I think we are safe here, but I just dont know.
 
You must log in or register to reply here.
Back
Top