Whelen Model(s) 9035801/2/3/4


My Most Satisfying Analysis to date, the Whelen Models 9035801/2/3/4 was analyzed based on this model’s Installation guide available on Whelen’s website. Whelen has done a nice job in making the installation guide, by providing the necessary details and good drawings.

Here is a photo I had found on the internet, which got me very interested in this light model. To be very honest, had it not been for his photo, I may not have started this blog. Thanks to the source, though now untraceable.

Whelen Model 9035801/2/3/4 installed on a Cessna Citation. Note the LED array.

Part being Analyzed

Manufacturer: Whelen

LED Light model Number: 9035801/2/3/4

Description of Model: LED Anti-Collision/LED Position Light Assembly

Applicable FAR Part : 23

Documents Referred for Analysis:

1. Whelen Installation Guide for Model(s) 9035801, 9035802, 9035803, 9035804

2. CREE XR-E LED Datasheet

3. CREE XR-E Lumen Maintenance Technical Article

4. FAA Part 23.1401

Tools Used for Analysis:

1. Microsoft Excel 2007

Microsoft Excel is a powerful tool with which you may perform analysis of limited complexity. Theoretically it is too powerful to be unbounded in its applications, but practical limitations force designers to switch over to more convenient tools. I personally use Excel for a quick evaluation of optical designs which have either minimal or no secondary optics. 

Introduction

Whelen’s Model 9035801/2/3/4 is a combined navigation light and anti-collision assembly for aircraft wingtips. Of the four models listed above, two are intended to be mounted on the right wingtip, while the other two are intended for the left wingtip. Only one unit goes on each wingtip.

This wingtip unit includes both a strobe light system and a navigation light system. However, in this analysis, we shall be investigating only the strobe light system.

Terms and Definitions Used throughout this Analysis:

Axis : The “axis” of our interest in this discussion is the aircraft axis that passes along the aircraft fuselage from the nose to the tail.

LED : Unless otherwise referred to, “LED” shall refer specifically to the White XR-E LED manufactured by CREE.

Forward: Refers to that direction, parallel to the aircraft axis, that points in the same direction as the aircraft’s nose.

Rear: Refers to that direction, parallel to the aircraft axis, that points in the same direction as the aircraft’s tail.

Outboard : Refers to the direction perpendicular to the aircraft axis, that points away from the aircraft’s fuselage.

Unit description

The unit comprises of three printed circuit boards, possibly made of FR4 material. The forward facing board is inclined at an angle of 105° to the aircraft axis, the outward facing board is mounted parallel to the aircraft axis, while the rear facing board is mounted at an angle of 75° to the aircraft axis.

Image 02: The Light Unit, held vertical. Note the LEDs on board A and B, outlined in yellow. These are white XR-E LEDs for the strobe portion of the integrated light. The LEDs outlined in green are for the forward navigation/position light. These are RED LEDs for the left wing unit, and Green LEDs for the right wing unit.

Each printed circuit board has 12 LEDs exclusively for the strobe light function. In addition, the forward facing board has 5 colored LEDs for the forward navigation/position light system. 5 of these colored LEDs are installed on the forward facing board, and a single LED is mounted on a separate small board on the base plate, parallel to the aircraft axis. In total, there are 6 colored LEDs for the forward navigation/position light system. These colored LEDs are RED for the model that goes onto the left wingtip, and Green for the model that goes on the right wingtip. The rear facing board has 2 White LEDs, possibly similar to the LEDs used for the strobe function, for the rear navigation/position light system. With one light installed on each wingtip, a complete navigation/position and anti-collision light system is realized. Image 03:The LEDs outlined in Yellow are the White CREE XR-E LEDs for the strobe. Those outlined in green are colored CREE XR-E LEDs for the forward navigation light. Only Boards A and B are visible here.

Image 03: Board Angles for the Whelen Light. Note that the light is being seen from above, that is, the top view, when this is installed on the right wingtip.

Since this unit has been observed on Business Jets, we shall consider only FAR Part 23 rules that apply to this category.

Method of analysis (Brief)

1. Analyze the Whelen Installation Guide and note the parameters of interest.

2. Use the FAA provided Blondel Rey Equation to determine the Blondel Rey factor to arrive at the effective intensity.

3. Characterize the LED based on graphs available in the datasheet to determine the current at which each LED is being driven.

4. Characterize the LED based on graphs available in the datasheet to determine the multiplicative factor of the luminous flux of each LED, at the current which was determined in (4).

5. Choose a LED part number for this analysis, and determine the actual luminous flux from the LED, at the current at which it is being driven, using the multiplicative factor determined in (5).

6. Establish the relationship between the luminous flux and the temperature of the LED junction.

7. Tabularize the spatial intensity distribution of the LED, based on the relevant graph available in the LED datasheet.

8. Perform mathematical calculations to determine the spatial luminous intensity distribution of the LED, at a given current, based on the table generated in (8).

9. Approximate the three LED Boards (A, B C) as dimensionless point sources of light with the cumulative intensity and intensity distribution of all 36 LEDs, respecting the orientation of each board, and arrive at the horizontal intensity distribution of the light unit. Apply the necessary luminous flux degradations based on assumed temperatures of evaluation and collectively plot the horizontal intensity distribution of the light under varying thermal and lifetime conditions. Compare this with the FAR requirements.

10. Plot the vertical intensity distribution of the Whelen light by considering only Board B. Apply the necessary luminous flux degradations based on assumed temperatures of evaluation and collectively plot the vertical intensity distribution (at horizontal angle=90° from the aircraft axis) of the light under varying thermal and lifetime conditions. Compare this with the FAR requirements.

11. Come to conclusions about the light’s performance and expected useful life of the anti-collision light.

Detailed Analysis:

1. Analyze the Whelen Installation Guide and note the parameters of interest. Use reasonable assumptions to determine the power at which each LED is being operated.

Table 01 : Given data and Derived data for this analysis.

The number of LEDs, the intensity distribution of the LED, and the amount of light from the LED can help us arrive at the intensity.

The amount of light given out by an LED is a function of the current being passed through the LED. This current, multiplied by the voltage developed across the LED, will determine the power being dissipated by each LED. This power, multiplied by the total number of LEDs for the strobe system, when added to the electrical loss of the power supply, gives us the total power drawn by the strobe light, or the current drawn by the strobe light at a particular voltage.

We’ll work in just the reverse of what was described above to arrive at the current through the LED.

In Table 01, in the Derived Data Section, the Input power is derived from the Installation guide, which details the operating voltage (28Volts) and the current drawn by the strobe (4.2 Amps). Assuming a driver efficiency (The efficiency of the electronic circuitry that strives to maintain a constant current through the LEDs) of 90%, the actual power reaching the LEDs may be assumed to be 105.84W. Dividing this by 36 LEDs (3 boards of 12 LEDs each), We arrive at a peak power of 2.94W per LED. This is the power dissipated by the LED when the strobe is “flashing”.

2. Use the FAA provided Blondel Rey Equation to determine the Blondel Rey factor to arrive at the effective intensity.

Equation : Blondel Rey Equation for Effective Intensity of Flashing lights

Pattern 01: The Flash pattern of the light, as a function of time. (250mS ON time, 45 flashes per minute)

To the human eye, a flashing light appears to be of lower intensity when compared to the same light glowing steadily. The FAA recognizes one human eye model that best describes the “effective” intensity of a flashing light. This equation, known as the Blondel Rey equation, may be found contained within FAA part 23.1401. We shall use this Blondel Rey equation to arrive at the “effective” intensity of the strobe light, as perceived by a human.

The strobe light is kept “on” for a duration of 0.25s (250mS). Since the same current will be maintained through the LED, we can confidently assume that the Intensity of the LED remains invariant during the “on” period. As a result, the numerator in the equation integrates to I[t2-t1] = 0.25I. The denominator reduces to 0.2+[t2-t1] = 0.45. Dividing the numerator by denominator gives us 0.5555I. In this, the 0.5555 multiplier may be considered the Blondel Rey factor. In other words, the effective intensity of the strobe light is just a little over half the measured intensity of the same light, kept glowing continuously at the same power level.

3. Characterize the LED based on graphs available in the datasheet to determine the current at which each LED is being driven.

Table 02 : Table : Voltage, Power, Current relationship. Graph: Current vs Power

The CREE XR-E datasheet gives us the Current-Voltage relationship of the LED. For every 0.1V, the current through the LED is noted against the corresponding voltage. Multiplying the corresponding voltage and current gives us the power being dissipated by the LED. Hence, a relationship between power and current is arrived at. Using Excel, we can determine the equation of the current-power relationship for this particular LED. With this equation, for the input power of 2.94W, the current through the LED is determined to be 816mA. Whelen may have set it to around 800mA or 820mA, but let’s stick with this value of 816mA.

4. Characterize the LED based on graphs available in the datasheet to determine the multiplicative factor of the luminous flux of each LED, at the current which was determined in (3).

Table 03: Relative Luminous FLux vs Current through the LED

Luminous flux is the quantity of light passing through the LED’s lens, which is also the total amount of usable light that the LED gives off. The luminous flux varies with the current through the LED, in a non-linear fashion. The CREE XR-E datasheet provides a graph of this relationship.

The luminous flux of the LED, at a current of 350mA, is considered to be the 100% yardstick. The luminous flux at any other current level is expressed as a multiple of the flux at 350mA. Since the manufacturer offers CREE XR-E LEDs with varying levels of flux at 350mA, its most sensible to provide only the relative luminous flux across the applied current range.

The equation for this relationship, for the CREE XR-E has been established in table 03. using this equation, it is determined that at a current of 816mA, the luminous flux through the LED is roughly 1.9 times the luminous flux at 350mA.

5. Choose a LED part number for this analysis, and determine the actual luminous flux from the LED, at the current at which it is being driven, using the multiplicative factor determined in (4).

Table 04 : Selecting the Highest Flux Bin LED

CREE offers 4 flux bins for the cool white XR-E LED. Most aircraft lights use Cool White as the a good portion of the color spectrum falls within the FAA’s acceptable definition of “Aviation White”. This shall be touched upon in the LEDs section. Of the four  flux bins, the Q5 flux bin guarantees a minimum of 107 lumens per LED at 350mA, at 25°C, at the beginning of the LED’s life. The highest flux bin LEDs are the costliest, but here we’re assuming that Whelen has chosen this to make their lights bright.

Using the value of 107 lumens (the unit of luminous flux) per LED, we multiply it by the normalised value of luminous flux at 816mA, which was determined in (4). 107 lumens X 1.9 gives us roughly 203 lumens, at a die temperature of 25°C at the beginning of the life of the LED.

6. Establish the relationship between the luminous flux and the temperature of the LED junction.

Table 05: Light Output Degradation with Temperature

If you’ve noticed, as soon as we began talking about the actual luminous flux (lumens) being given out by each LED, we were cautious enough to specify the temperature and the age of the LED. Conventional sources of light rely on heat to give off light. However, the light output of LEDs degrades with time and temperature. In this section, we shall concentrate only on the temperature relationship.

Using the graphical relationship provided in the datasheet, between LED temperature and the relative luminous flux at each temperature point, we can arrive at an equation (with the help of Excel) that relates these two parameters. The graph starts at 25°C, and stops at 150°C. Taking the LED die (also referred to as the “junction”) beyond 150°C may permanently damage the LED. Now what is unclear about this relationship is whether the LED performance is better below 25°C or not. Common sense would respond in the affirmative, but a common sense of caution would prevent us from jumping to our own conclusions about an LED which we don’t manufacture.

We shall use this relationship later in the analysis to determine how well the light performs under varying thermal conditions.

7. Tabularize the spatial intensity distribution of the LED, based on the relevant graph available in the LED datasheet.

Table 06 : Normalised Spatial Distribution (table on the left). Actual spatial luminous intensity distribution for the 107lumen CREE XR-E LED, at 350mA, 25degrees C, 0hrs (table on the right)

The LED datasheet provides us with the spatial intensity distribution of the LED. It’s as simple as measuring the light intensity at varying angles from the LED’s normal. This angular intensity distribution is rotationally symmetric about the LED’s normal.

If we are aware of the intensity of the LED along the 0° line (normal to the LED), we can determine the intensity at various angular points by the use of this provided spatial intensity curve. This will be helpful for us in determining the intensity of the whole strobe unit at a few angles of interest.

8. Perform mathematical calculations to determine the spatial luminous intensity distribution of the LED, at a given current, based on the table generated in (7).

The spatial intensity distribution can help us determine the peak intensity of the LED. Conceptually, this involves the virtual construction of the rings of a hemisphere, determining the luminous flux that passes through each of these rings based on the spatial intensity distribution, and then determining the intensity along the LED’s normal. This procedure shall be detailed in the LED section.

Table 06 shows on the table to the right of the spatial distribution curve, the spatial luminous intensity in candela, at a current of 350mA, at a die temperature of 25°C, at the beginning of the LED’s life.

9. Approximate the three LED Boards (A, B C) as dimensionless point sources of light with the cumulative intensity and intensity distribution of all 36 LEDs, respecting the orientation of each board, and arrive at the horizontal intensity distribution of the light unit. Apply the necessary luminous flux degradations based on assumed temperatures of evaluation and collectively plot the horizontal intensity distribution of the light under varying thermal and lifetime conditions. Compare this with FAR requirements.

Table 07 : The Effective intensity of one Strobe Light unit, for angles from 0 deg to 180deg, referenced to the aircraft axis. Any light beyond the angles of 0 to 180 is disregarded

This is the part we have all been waiting for.

The strobe light consists of 3 boards, each with 12 LEDs for the strobe function. We shall assume each board to be one single LED, with 12 times the intensity of each LED that is driven at 816mA, at different die temperatures and ages. We shall respect the angles of the LED boards, and make each assumed single LED point in the direction of their boards. We then tabulate the intensity of each LED at angles referenced to the aircraft’s axis.

Refer Table 07. The intensities of each board are added to the intensities of other boards for the same angles referenced to the aircraft axis. This ensures that the intensity contributions from each board is taken into consideration, at every horizontal angle.

Since the LED does not emit any light beyond 90° on either side of its normal axis, the intensities in each column drop to zero for any angle that beyond 90° from the LED’s axis. Also, the proximity of LEDs means that beyond a certain angle, the light given out by LEDs that are obstructed by the physical presence of the adjacent LEDs on the same board. For this reason, for angles beyond 80 degrees, the intensity contribution per board is only from 3 LEDs, instead of 12 LEDs. (The centre board has only 3 LEDs per column. Towards the extremes, the other 9 LEDs are obstructed by the edge string of 3 LEDs).

The summed up contribution from each LED board is multiplied by the Blondel Rey factor to account for the human eye response to a flashing source. The intensity is tabulated in the grey column, under effective intensity, at T=25°C and 0hrs of LED life (consider it a LED turned on only now, during the analysis). The other columns tabulate the spatial intensity of the unit under elevated junction temperatures, by multiplying the effective intensity at T=25°C with the intensity fraction determined in (6).

The last column takes into account the light intensity at a point when the luminous output has fallen to 70% of the intensity at T=25°C and Operating time=0hrs. Above this, we determine a junction temperature at which the intensity will be further degraded, to just meet the FAA minimum requirement. This temperature was determined by trial and error. First, we identify the point in the intensity distribution curve that has the least intensity. Next, we degrade that intensity to 70%, and note the intensity. We then divide the FAA minimum requirement (400cd) by the degraded minimum intensity. That ratio is how much more the light can degrade before falling below the FAA minimum. We find on the temperature-luminous flux graph which temperature results in that same degradation ratio. This temperature is the maximum permissible temperature, at a 70% degraded output due to life, when the light can still be “legal”.

Lets walk the talk. At angles referenced to the aircraft axis of 0° and 180°, the intensity is minimum at 663cd, at T=25°C and Life =0hrs of operation. At a particular age of the LED (in terms of operating hours), the light output will fall to 70% of the initial. This will occur at a time (in operating hours) which is dependent upon the LED current, LED junction temperature, and the ambient air temperature (around the LED). At this point, the light output falls to 663 X 0.7 = 464cd. So how much more can we further degrade the light output before we hit the FAA minimum of 400cd? 400cd/464cd = 0.862.

By looking at the relationship between temperature and the luminous flux degradation, this ratio of 0.862 occurs at a die temperature of 78.5°C.

10. Plot the vertical intensity distribution of the Whelen light by considering only Board B. Apply the necessary luminous flux degradations based on assumed temperatures of evaluation and collectively plot the vertical intensity distribution (at horizontal angle=90° from the aircraft axis) of the light under varying thermal and lifetime conditions. Compare this with FAR requirements.

Table 08 : Only Board B's contribution to the intensity in the vertical plane at a point of observation 90 degrees to the aircraft axis along the horizontal plane.

The same concept that was used in (9) is used in evaluating the vertical performance of the light. However, for the sake of simplicity, only the contribution of Board B’s contribution to the intensity is considered. Board A and C will have their effect on increasing the intensity at the 90° angle to the aircraft axis along the horizontal plane, but even without these boards, it can be seen that the light performs just fine.

11. Come to conclusions about the light’s performance and expected useful life of the anti-collision light.

Here are the plots of the aircraft light’s intensity pattern, along the horizontal plane (at a vertical angle of 0°), and the vertical plane (at a horizontal angle of 90°). Note the FAA minimums, and the performance of the light under various temperature / age conditions.

Pattern 02: The Horizonatal Intensity pattern of the light (At an observed vertical angle=0°), under various conditions. Note that the outer lens and its effects have not been accounted for.

Pattern 03 : Vertical Intensity Pattern of the Light (at horizontal angle = 90°) under various conditions. NOte that the outer lens and its effects have not been accounted for.

The Cessna Citation X’s Chapter on Lighting states, about White Anti-collision lights, “These lights are of very high intensity and can be disturbing to other airplanes and ground personnel if they are used during ground operation. They should be turned on just prior to takeoff roll and secured shortly after landing.”

Considering that most SOPs (Standard Operating Procedures) require the anti-collision lights be turned on when entering the runway (for takeoff) and turned off while exiting the runway (after landing), the strobe lights will be operated on ground for a very short period of time. A major chunk of its time will be up in the air, where temperatures are low, allowing for a cooling of the strobe lights. With this in mind, it may seem that the LED strobe lights may not operate in sever conditions. But yes, nothing much can be said till a proper thermal analysis is conducted, or the actual unit is measured in simulated conditions.

Please note that this analysis does not consider the effect of the outer lens on the intensity of the strobe unit.

Take a look at Pattern 02, the horizontal intensity distribution of the strobe light. Assuming on a hot day, the temperature inside the integrated light unit soars to 70°C. When the strobe light is operated, the thermal resistance between the LED junction and the LED’s case (which the LED datasheet specifies as 8°C/W : Refer Table 01) will drop a temperature of 23.5°C (See table 01). This means that if the case of the LED is X°C, the junction (die) inside the LED will actually be at (X+23.5)°C when operated at that power of 2.94W. So if the strobes were turned on at a unit temperature of 70°C, the junction within the LED will shoot to 93.5°C. That’s when the light’s light output (sans the outer lens) will be as intense as the values in the 100°C column in Table 07. The minimum value is 535ecd, and assuming that 80% of the light is lost in through the glass, the minimum intensity of the unit turns out to be 535 X 0.8 = 428cd. The lights are still above the minimum FAA intensities as detailed in FAR 23.1401. As the temperature rises, the light output will fall, but by then the aircraft would be airborne, and in an environment where cooling shouldn’t be a problem.

Another good thing which may make someone not fret too much about the thermal performance of the light is the flash rate. At 45 flashes per minute, and an “on” time of 250mS, the duty cycle of the strobe light is 250mS/(1.33s) = 0.1875. With this duty cycle, the average power of each LED is 2.94W*0.1875 = 0.55W. This is a figure that feels good, from a thermal perspective.

Now let’s take a look at the CREE XR-E’s lumen maintenance document. Contained in that document is the variation of the L70 lifetime of the LED (The number of hours of operation of the LED, under the mentioned parameters, which will degrade the light output to 70% of the intensity at the beginnning of life of the LED) versus the parameters : LED Junction Temperature, LED operating current, and the ambient air temperature. Refer the graph below.

Graph 01: Variation of the LED's L70 Life against the Junction Temperature, and Drive Current for various Ambient Air Temperatures.

Of these graphs, we purposely choose a current that is higher than the operating current. Here we choose 1000mA as the operating current, just to make our life estimations a bit conservative. Also, we had previously arrived at the junction temperature of 78.5°C to ensure that at the L70 life, the LED would be just meeting the FAA minimum requirements. Here, let’s find that L70 value, however conservative.

Lets search for a graph with a blue line (the blue line represents the 1000mA operating current) that is on the right hand side of the 80°C (remember, 78.5°C ~ 80°C). The graph with the Tair = 55°C is what we’re looking for. The L70 line terminates at 85°C; the graph does not stretch left any further. At this 85°C junction temperature, at a drive current of 1000mA, the expected L70 life of the LED is around 58,000hrs.

You may cry foul, but all we’re doing is to make our estimate so conservative that there will be no way in which an operator may ever find his lights flashing at an intensity below the FAA minimums. Also note that this estimate assumes:

1. A continuous operation of the LED at 1000mA (the Whelen light drives a lesser current through the LEDs)

2. A continuous junction temperature of 85°C (Which means the board temperature is around 85°C -23.5°C (see table 01) = 61.5°C. This may be on the higher side for an aircraft at its cruise altitude of more than 30,000ft; but nevertheless, lets remain this conservative without the actual thermal performance of the unit).

3. A continuous ambient air temperature of 55°C.

Now, the LEDs glow for 250mS, with a flash every 1.33 seconds (See pattern 01). This means that the actual duty cycle of the LEDs is 0.25/1.33= 0.1875.

In other words, if we are to account for the life of the LED with this duty cycle, we can expect the LEDs, under the above 3 operating conditions, to fall to 70% of their brightness at a life of:

58,000/0.1875 = 309,333hrs.

To be very honest, the MTBF (Mean Time Between Failures) of the electronic components that drive the light may not boast a figure more than 80,000hrs. IN which case, even before the LEDs may die out, the other electronic circuitry may fail.

In other words, an operator may safely assume their Whelen lights to last not more than 80,000 hrs. Most good companies peg a value lower than this just for the sake of reliability. So if you use a pair of these lights,

1. Get an intensity check done every 1000hrs.

2. Change your lights after 50,000hrs of unit operation.

If you were able to digest this analysis, drop in a mail at SolidStateLights[at]gmail.com. Would love to hear from you. Or Tweet me, aka “solidstatelight”.

Good luck, and Good night.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s