April 25, 2007
A 3-watt red Luxeon LED at a distance of 14.82 miles (23.85 km) with downtown Salt Lake City (and the cemetary) in the foreground, as viewed from the north end of the path.
Photo by Gordon, K7HFV.
Click on the image for a full-sized version.

3-watt Luxeon seen at a distance of 14.91 Miles (23.85 km) with downtown Salt Lake City in the foreground

Figure 1:
  Relative spot sizes and shapes of the LED (left) and laser (right.)  The color and saturation of the two spots has been exaggerated to make them easier to see.
Center:  The spectra of the Laser.
Bottom:  The spectra of the LED.  Note that there was some "thickening" of both spectral lines owing to saturation of the camera's imager.
Click on the image for a larger version.
Relative spot sizes of the laser and LED
Spectrum of the Laser
Spectrum of the LED
Since our previous test several things had been happened:
Initial setup:

Each of us when to the same locations as the last time - that is, I was on the west side of the Salt Lake valley while Gordon and Ron were in the northeast corner of the valley, near the Salt Lake City cemetery - a distance of 14.82 miles (23.85 km.)  On this evening the temperature was about 55 degrees F (13 C) at the start and there was only a very slight breeze.

Interestingly, despite calmer air than last time, all of us noticed that it seemed that the city lights "flickered" more than they had done during the previous tests.  I also noted that re-focusing of the receiver's optics paid off:  The pointing of the received signal seemed "sharper" than before along with noticeably better off-axis rejection of other light sources.

Once a communications link had been established using the Luxeon and Fresnel lens system, I set up the equipment for the evening's main experiment.

The experiment - Coherent versus Noncoherent light:

The objective of this evening's experiment was to quantify the difference in scintillation between coherent light as emitted by a laser diode and noncoherent light as emitted by a standard LED:  It has been suggested that  nonuniformities (e.g. "cells" of air with different temperature, pressure, and humidity) in the air volume between two sites would result in greater apparent scintillation when using coherent light sources as compared to noncoherent light sources.

While a similar experiment was done during the last test, it wasn't really a fair comparison:  While the light from the LED was emitted by a fairly large aperture (a 250mm by 318mm Fresnel Lens) the cross-section of the beam from the Laser pointer was only a few millimeters in diameter.  Because the amount of scintillation is strongly related to the aperture of both the emitting aperture and detecting aperture, by this fact alone the Laser's signal would have experienced more scintillation.

This time, we wished to avoid this disparity so we decided to do two important things:
Because of the distance (about 15 miles) it was necessary to use good optics to collimate the beam, so I decided to interface to an 8 inch (20.3cm) reflector telescope (a Celestron C8) that I have.  The effective aperture of this telescope, taking into account blockage by the secondary mirror, is approximately 50.27 square inches (approx. 289 square cm) - a respectable size.

Several months ago, I happened to note that a local hardware store had Laser levels on a clearance sale for $3 each, so I bought several of them, knowing that the $3 price tag was a good price for the Laser alone.  Attacking the level with a variety of tools (which destroyed the level, as it turns out) yielded a self-contained Laser module that ran directly from 3 volts.

As is, the Laser module could not usefully illuminate the telescope's optics owing to its very low divergence, so I unscrewed the lens from the laser module in order to defocus the beam to a divergence suitable for interfacing with the telescope's optics.  While this significantly "de-focused" the beam, it was still far too narrow to be useful:  To remedy this, I removed the plastic lens from the Laser module, turned it around, and re-installed it, thereby increasing the amount of divergence, allowing me to more-efficiently illuminate the telescope's optics.

The optical outputs of typical Laser diodes have a property that complicates efficient beam collimation:  They produce a narrow, "fan" beam - a property illustrated in Figure 1.  Without complicating the optics, it is difficult to "capture" most of the beam (something that would require the beam to be made more rectangular) and properly illuminate the optics of the telescope, so the only other option is to only take the very "center" of the beam, discarding the majority of the optical energy:  If the beam is collimated too strongly before being applied to the telescope, the collimated beam would fully not fill the aperture and would be cut off on the sides.

For the noncoherent light source, I found a large (10mm) LED in my parts bin - a Radio Shack P/N 276-086.  This LED is rated for an optical output of 5000 mcd at 36 milliamps with a beamwidth of 30 degrees at a wavelength of 660 nm - very close to the same wavelength of the Laser as can be seen in Figure 1.

When testing, I could tell that the LED's built-in lens did not efficiently illuminate the telescope's optics (that is, the LED's beam was so wide that much light was "wasted" off to the sides) but I was satisfied, when comparing the optical output of the LED with that of the Laser, that sufficient light was available to the telescope, so I did not add any extra optics.

Interfacing the LED and Laser modules to the telescope:
Figure 2:
  The back end of the Laser module.  Center:  The front end of the LED module.
Bottom:  The Laser module installed in the telescope's eyepiece mount.

Click on the image for a larger version.
The back end of the laser module
The "business" end of the LED module
Laser module installed in the telescope's eyepiece mount

The telescope has a standard 1-1/4 inch eyepiece mount and star diagonal, so it made sense to use it to hold the Laser and LED modules.  Rummaging around my scrap metal, I found some aluminum tubing that was exactly the right size to slide into the compression-mount eyepiece holder.  To mount the LED and Laser, I found some old Thermalloy (tm) TO-5 transistor aluminum heatsinks which fit inside some gray ABS pipe that I had laying around.

The Laser module and LEDs were first glued to the TO-5 heat sink, and after this, the heat sink (with the LED or Laser attached) was then glued inside a piece of ABS pipe.  To mount the ABS section inside the aluminum tubing - and to provide a means of precise optical alignment - I drilled and tapped three 8-32 screws:  These setscrews can be seen in the pictures of Figure 2.  To optically align the LED/Laser, the screws were loosely tightened and then, using a thin-blade screwdriver, I was able to reach in and adjust the orientation of the emitter until it was properly aligned with the telescope's optics.  After this, the set screws were tightened, alignment re-checked, and then the screws and ABS piece was glued into place.

When installed (see the bottom image of Figure 2) the light output of both the Laser and LED completely illuminated the entire aperture of the telescope.  Further testing also revealed that the beam could be easily focused by the telescope to provide best possible collimation.

Driving the Laser and LED:

The only practical way to linearly modulate the Laser is to use PWM (Pulse Width Modulation) techniques, so it was natural that I use my PWM optical modulator for this task.  Because the Laser module already included the circuitry necessary to be operated directly from a 3 volt battery, I simply constructed a voltage regulator using an NPN transistor and two red LEDs to provide a regulated 3 volt source.  Because the Optical Modulator uses a current sink as its output stage, it was necessary to provide a loading resistor to provide enough load for the circuit to work, so the source voltage (from, the modulator) was paralleled with a 20 ohm resistor:  With proper adjustment of the drive current, the Laser module was properly driven with the PWM waveform.

Interfacing the LED was much easier:  A 150 ohm series resistor was used to limit the current to the LED to a peak value of about 60 milliamps while a parallel power  resistor of about 17 ohms was used.  While the 60 milliamp current exceeds the maximum rating of the LED, the average current would be half this owing to the nature of the PWM waveform.

Aiming the telescope:

In our previous outing, it was noted that it was extremely difficult to aim the Laser pointer while it was mounted on a tripod:  Without micrometer-type adjustments, the slightest touch would knock the Laser completely off-point and without a sighting device, it was difficult to get even coarse pointing.

The telescope has the obvious advantage in that it's a telescope!  With the 3-watt Luxeon in operation from the far end, I had a ready visual reference upon which I could train the telescope and with it, I could not only see the very bright red LED, but I could also see Ron's and Gordon's flashlights as they were standing near the optical transceiver.

I then removed the eyepiece and substituted the Laser module, it took only very minor adjustments (of both pointing and focusing) to peak the signal, using my own received test tone being heard via the optical link from the far end.

It was noted that, with the telescope, the apparent beamwidth of the Laser was comparable to that of the LED.  This telescope has some very fine adjustments for declination and it took about the same amount of adjustment to "pass through" the Laser's beamwidth as compared to the LED's beamwidth, with the total beamwidth being much less than one degree.

Comparisons of the signal quality of the Laser versus the LED:

According to Ron and Gordon, the perceived luminous intensity of the Laser was about the same as that of the 3 Watt Luxeon with the Fresnel lens, but it was easy to see that it was "flickering" far more than the LED had been.  While the LED in the telescope was visibly dimmer than the Laser, it was clearly less "flickery" than the Laser:  A quick listen to the test tone revealed that there was, in fact, very significant scintillation of the Laser's signal as the test tone sounded quite raspy, as the following clips demonstrate:

Important note:  Due to the MP3 audio compression, finer details of the scintillation are lost, but the audio files "sound" pretty much identical to the uncompressed PCM files to which they were originally recorded.

All of the audio clips were transmitted using the 8 inch reflector telescope and received using a 7"x10" (17.8 cm x 25.4 cm) Fresnel lens - that is, an area of approximately 70 sq. in, or 452 sq. cm.

Unfortunately, the audio clip of the test tone from the Laser experienced some clipping (overdriving) when the recording was made - but clipping occurred only on higher audio peaks.  In both clips, some "clicking" can also be heard in the background (especially on the LED clip) and this was from the strobes of aircraft flying slightly above the optical signal path.

A more "real world" comparison may be heard in the clip below.  This clip directly compares the "sound" of the Laser's signal with that of the LED.  In this clip, the Laser is tested first, and then the LED is immediately installed to assure that both tests were done with the same atmospheric conditions.

This clip has been edited to reduce its length and improve continuity:  At the end of the clip, K7RJ's comments (transmitted via LED from the end of the path where the bulk of the recording was made) were mixed in from a recording made at my end as received via the lightbeam link.  During the entire transmission (except during the 1kHz tone segment) a 4kHz "pilot" tone was transmitted along with the normal audio, but this has been notched out (except for the brief test segments) to make it more "listenable" as the 4 kHz tone is really annoying.  Other than the removal of the pilot tone, no other filtering or amplitude adjustment was done, hence the presence of the 120 Hz tone (and harmonics) from the city lights.  The rhythmic clicking of the strobes of passing aircraft can also be heard in various parts of this recording.
Figure 3:
  Waveforms of a portion of the 4 kHz test signal transmitted by the Laser.
Top-middle:  Close-up of a portion of the scintillation of the 4 kHz Laser-transmitted tone.  This portion has been "zoomed in" in both the vertical (amplitude) and horizontal (time) axes.
Bottom-middle:  Waveform of a portion of the 4 kHz test signal transmitted by LED.
Bottom:  Close-up of a portion of the scintillation of the 4 kHz LED-transmitted tone.  Note that the rate-of-change of amplitude is much slower than with the Laser as evidenced by noting the time scale along the bottom of the images.
Click on an image for a larger version.
Scintillation of 4 kHz tone transmitted via laser
Close-up of scintillation waveform of Laser-transmitted 4kHz tone
Scintllation of 4kHz tone transmitted via LED
Close-up view of scintillation of 4 kHz tone transmitted by LED

The audio clip below consists of the following segments:

A more in-depth analysis of the scintillation:

The waveforms in Figure 3 offer close-up graphical analysis of 4 kHz tones from the original uncompressed PCM recordings show the true nature of the scintillation.  These waveforms have been bandpass-filtered to from 1 kHz to about 6 kHz to remove 120 Hz energy from the city lights and the segments from which the analysis was taken have been amplitude-normalized to 16 bits full-scale.

Laser scintillation:

As can be seen from the top image of Figure 3 there is significant scintillation that occurs at a very rapid rate with a close-up being seen in in the top-middle image.  Note that the top-middle image has been "zoomed in" in terms of time to show the rapid rate-of-change and amplitude to view the depth of the scintillation:  The reference of this image is, like the others, based on a full-scale 16 bit sample.  Analysis of the original audio file reveals several things:
LED scintillation:

The bottom-middle image of Figure 3 shows a typical example of scintillation from the LED.  Analysis of the original audio file shows several things:
Observations with a 1kHz tone:

It should be mentioned that on the previous test the measurement of the scintillation using the Luxeon LED and Fresnel Lens was done using a 440 Hz tone.  It was noted at the time that the temporal resolution of this tone was too low (e.g. too few cycles) to capture some of the short term properties of the scintillation - particularly in the case of the Laser's scintillation - and this resulted in somewhat lower measured amounts of scintillation that were likely present.

In analyzing the 1 kHz tones from this test, it is more difficult to realize the apparent depth of the scintillation:  On the Laser, depths of well over 30 dB are readily apparent, but due to clipping on the original recording, the true depth of some of the faster fades cannot be accurately measured.  On the LED, the noted scintillation depth was more in line with that observed using the 4 kHz tone (about 22dB peak) because the changes were slower and were more easily captured at the lower frequency.

Observed scintillation using the Luxeon and the Fresnel Lens:

A final test was done using the Luxeon and Fresnel lens combination for transmitting and the same Fresnel as before for receiving.  For this test, the LED current was reduced so that it was visually similar in brightness to that of the LED in the telescope.  The Fresnel lens has a larger aperture than the 8 inch reflector telescope - about 117 sq. in. (759 sq. cm) - a bit more twice the area, so it was expected that lower scintillation would result.  A typical sample of the scintillation on the 4 kHz tone may be seen in Figure 4.

Measuring using the 4 kHz tone, the peak scintillation was in the 17-20dB area and the rate of change of amplitude was comparable to that observed when using the telescope and LED for transmitting.  A clip of this may be heard in the following recording:
Figure 4:
Typical scintillation of a 4 kHz tone transmitted from a 107sq in (690 sq cm) Fresnel lens.
Click on the image for a larger version.
Scintillation from Luxeon and Fresnel

As in the previous clip, the 4 kHz tone present through much of the recording has been notched out - except during the specific 4 kHz test signal.  Also present is the 120 Hz hum from urban lighting.

Comparisons with the previous test:

As mentioned before, during the previous test we used some ordinary Laser pointers and did comparisons with those and the Luxeon-Fresnel combination, and as expected, the effects of scintillation were quite severe.  Unfortunately, for this test I did not have a standard Laser pointer that could be modulated, and the Laser module installed in the telescope could not be used by itself because it no longer produced a collimated beam.

While it would have been nice to do a side-by-side comparison, it is likely that the air conditions during these two tests were quite similar, so a rough comparison between the "small aperture Laser" and the "large aperture Laser" would still be valid.

As can be heard from these recordings:
There is a distinct "fluttery" noise in the background solely due to the scintillation while the magnitude of the signal emitted by the large aperture Laser transmitter exhibited this effect to a much lower degree.  Analysis of these recordings show that the rate-of-change of amplitude of the light from the Laser pointer is higher than that of the larger aperture telescope - a property that significantly distorted the voice.

Because Ron and Gordon reported receiving badly distorted audio, I checked the audio quality of the signal that I was transmitting using a local receiver and found it to be clean.  Also, because no 4 kHz tone was transmitted at the time of this test, direct comparisons pertaining to the "dA/dT" (rate of change of amplitude over time) of the Laser pointer's signal could not be made.

Is should be noted that in the case of the Laser pointer, its light needs to travel only fraction of the distance of the total path (1 or 2 kilometers) before its normal divergence causes the beam size to exceed that of the beam size emitted by the telescope, so much of the added distortion of the Laser pointer's signal is from its traveling through only that first portion of the optical path.


As expected, the coherent light experienced more scintillation than the noncoherent light and that the rate-of-change of amplitude of the coherent light was much faster (and of greater depth) than the noncoherent light.  In this test, both light sources were collimated approximately equally and from the same size of aperture.  While much of the Laser's scintillation was due to the air turbulence, it is very likely that the severity of the scintillation was increased due to wavefront cancellation occurring due to the differing velocities in the cells of air through which the light passed.  Comment:  No attempts were made to "dodge" any nulls in atmospheric transmissivity that might have been occurring at the precise operating wavelength of the Laser.

It has occurred to us that the light being emitted by the telescope is probably not phase-coherent across the entire aperture - and this would, no doubt, increase the likelihood of scintillation.  It should be noted that achieving such phase coherence across such a large area would require extremely precise optics - something that the average experimenter would not likely be able to find or afford.  What is significant is that, using just an ordinary high-brightness LED operated at a few 10's of milliamps it was possible to get results that were superior to those obtained when using the Laser.

It also occurred to us that not every experimenter has a large reflector telescope that can be dragged around - and if they did, the bulk, weight, and fragility of this piece of equipment would greatly limit exactly where it might be taken.  While the telescope has extremely good optics, it has also been shown that very good results can be obtained by using an inexpensive plastic Fresnel lens.  While this lens is not as efficient a collimator as the telescope, one can overcome the Fresnel's inefficiencies by brute force, using an extremely high-output LED such as the Luxeon III - both techniques that quite practical to the experimentor.

Other experimentation:

Using a normal LED with minimal optics:

Just prior to the test, Ron connected a standard red LED - mounted in a small flashlight reflector - to his PWM Laser modulator just to see if it could be detected, or if I could see it with the telescope:  Because I could see their individual LED flashlights, it was hoped that the LED would be powerful enough to be detectable:  It wasn't.  This was surprising, actually, as I expected that something would be detectable by ear.  This test was thrown together at the very last minute and it probably failed owing to the fact that the reflector chosen (a cheap flashlight) was not well-suited for focusing the LED that Ron chose.  If the LED and reflector had been better-matched, I have no doubt that something would have been detected.

According to the initial analysis of the audio file using narrowband techniques, it is possible that a signal may have been detected - but it was more 90dB below the peak audio level received from the Luxeon and Fresnel lens.

Using the Luxeon emitter with its secondary lens:

Another experiment was one in which I removed the Luxeon emitter (with its secondary lens) from the Fresnel lens enclosure resulting in an effective beamwidth of around 40 degrees or so.  Ron reported that with binoculars, the 3 watt Luxeon LED (with its secondary lens) was just visible and he could hear the "peaks" of the tone via speaker.  Further analysis of the recording indicate that removal of the LED from the Fresnel reduced their received optical power by 55-60dB on average with scintillatory peaks at higher levels.

Here is a sample of the signal - but turn your volume down for the first part :
Figure 5:
  The method used to modulate the telescope.
Middle:  Waveform of the Kennecott Strobe.  Three strobe pulses (negative-going spikes) are evident.
Bottom:  The "cheap" enclosure strapped down to Ron's "elevation adjust" contraption  - photo, by Ron, K7RJ.
Click on either image for a larger version
A method of modulating a telescope...
Waveform of Kennecott Strobe
Close-up view of the elevation adjust for the "cheap" enclosure

Yelling at our gear:

After we did all of the experimentation with the LEDs and Laser, we were wondering what else to do before we tore down our gear.  Ron, out of curiosity, decided to shine a portable 1 Million Candlepower spotlight at me:  As expected, I could easily hear the thermal roar as he did so - but I noticed something else:  A "bong" has he was handling the spotlight - apparent microphonics of the filament, as the following audio file demonstrates:
Or, for an shorter version:
Not to be outdone, I decided to try a similar thing, so I installed the LED and began shouting at various parts of the telescope's anatomy - see the top image in Figure 5.  After a few experimental shouts, we determined that the star diagonal seemed to be the most sensitive part, so I screamed at that:
Shortly after this (and with a hoarse throat from the screaming) we decided that it was getting late (it was past midnight) so we packed our gear up and headed home.

Pointing the receiver elsewhere:

From where Ron and Gordon set up, they could see the very tall (1215 ft, 370 meter) smokestack of the Kennecott smelter to the west of their location.  Because of its height, it is well-marked with a set of strobe lights to provide warning to aircraft.

Being curious, they pointed the optical receiver at the smokestack and from the sound of it, they could hear what sounded like multiple flashes.  Later examination of the audio file bore this out, as seen in the middle picture of Figure 5.
As can be seen from the waveform, there are at least three flashes.  It is also apparent that the combination of optical receiver and digital recorder inverts the polarity of the received audio, so the initial impulses from the flashes are negative with a little bit of overshoot on the return.  Also apparent in the image is the presence of 120Hz energy (from city lights) and harmonics.

As mentioned before, we often heard the "clicks" of aircrafts' strobes as they flew by:  Even though these were off-axis, their strobes were intense enough to be easily detectable.  In addition to the strobing of aircraft, I could also hear a nearly constant "tick" in the background - but I never spotted the source of this noise.

More may be added to this page soon - check back again.

Return to the KA7OEI Optical Index page.
Keywords:  Lightbeam communications, light beam, lightbeam, laser beam, modulated light, optical communications, through-the-air optical communications, FSO communications, Free-Space Optical communications, LED communications, laser communications, LED, laser, light-emitting diode, lens, fresnel, fresnel lens, photodiode, photomultiplier, PMT, phototransistor, laser tube, laser diode, high power LED, luxeon, cree, phlatlight, lumileds, modulator, detector
Copyright, KA7OEI 2007-2011.  This page last updated on 20110304

Page count since 201005: