Top: Grandeur Peak, as seen from Ron's patio. (That's
the peak to the left of center of the picture, partly obscured by the
Bottom: Grandeur Peak as seen from Clint's back deck
from across the valley. (That's the peak, just above the vent
When we did our testing, the mountain was slightly "whiter" than the
bottom picture shows - but not quite as white as in the top picture.
Click on either image for a larger version.
You may be asking yourself, "Self, what is 'Mountain-Bounce'?"
If you've read other pages at this web site, you already know that a
of us has managed to fascinate ourselves by launching red photons at
each other, only to catch a very small number of them - calling the
entire effort "Optical Communications."
Having had good success in past experiments with optical ("lightbeam")
communications, managing to have traversed
ranging up to 173 miles and using various types of light sources,
we decided to try something else for the hell of it: Bouncing
light off a mountain.
Choosing the mountain:
This part was easy: Ron, K7RJ, simply needed to step out onto his
patio and look up (see Figure 1)
see Grandeur Peak, a relatively prominent feature
along the Wasatch Front. Being only about 2 miles from Ron's
doorstep, he had a rather commanding view of this peak!
My house, being across the valley from Ron's, meant that my view of
Grandeur peak had less "grandeur" about it, but it had the distinct
advantage that I simply needed to step out onto my back deck and look
toward the east-northeast to see it - about 12.4 miles in the distance.
The point of this was that we both had a common geographical feature at
which we could blast our red photons and then gather up as
many of them as possible.
Would this work? Based on our own experiences and of others
"cloudbounce" experiments, we figured that we had a reasonably good
shot of at least achieving a one-way (Clint transmitting to Ron)
communications. Because my path crosses the majority of the Salt
Lake Valley, however, I knew that optical signals reaching my receiver
significantly "diluted" by the scattered, upward light from the
city: Ron, being quite close to the mountain and at the
edge of the valley had a far "less-polluted" shot!
First attempt - sometime in early-mid January:
It didn't work.
Actually, we weren't too surprised about this. We chose to try,
on this occasion,
at least in part because both of us had some time to make the
attempt. Mother Nature was conspiring against us, however, as
there was some fog/smog in the valley trapped by a temperature
inversion. While Ron could see
Grandeur Peak in the dark, I could not - but we persisted anyway to, if
nothing else, become more familiar with the gear and computer programs
for when we
were able to re-try under better conditions.
For a time, it seemed as though Ron was
able to see my signal
on the computer despite the conditions,
but just as we were being amazed by this it
suddenly occurred to Ron that it he might be seeing a signal from the
which had been disabled, but not turned off - that was making its way
into the extremely
sensitive receiver electronics. He decided to power-down the
modulator completely: He did, and it was. Drat!
Another time is the charm:
On the evening of January 17th, Ron and I happened to get together,
along with his wife, Elaine (N7BDZ) and other friends, for
dinner. On the way back home I
noticed the crystal clarity of the air in the valley - a marked
contrast to the past week or two of temperature-inversion smog that had
been suffocating the area. Already being fairly late on a
Saturday night there was nothing else on the schedule so we
dragged the gear out of our respective houses and tried it again.
For this experiment, we used the same optical gear that we had used
during our previous attempt - which was also the same gear that we'd
used for the 2007
ARRL "10 GHz and up" contests
consists of a high-power LED transmitter and a sensitive optical
receiver, both using
large, plastic Fresnel lenses. This gear had already been proven
perform well, having spanned across 173 miles of Utah desert
Also, for this experiment, we used half-duplex
that is, we made no attempt to transmit and
listen at the same
time, even though the optical transceivers were perfectly capable of
such. The reason for this is that we didn't wish to
"de-sensitize" our receivers with light from our own
transmitters. At both
of our ends there were some local objects that were at least partially
lit-up by or slightly blocked the beam: On Ron's end, he was
shooting through some bare tree branches and overhead wires while
my path skimmed along the roof of a neighbor's garage and furnace vent
stack: The reflected light from these objects caused us to
some of our own signal. Not only this, there was both Rayleigh
and dust scattering to contend with and each of these effects could
"mask" a weak signal from the other end were we to attempt to transmit
while trying to receive!
Generating and detecting signals
One of the features of the optical modulator
is a built-in
generator: Programmed into this generator are a number of
discrete tones - none of which bear any harmonic relationship to the 60
Hz AC mains frequency used here. For this test, we chose the
"Middle-C" tone (about 262 Hz) - a frequency that was high enough to be
heard, but low enough in frequency to be in the "more-sensitive" range
of the optical receiver's frequency response.
Top: Ron's signal being copied at Clint's QTH. This
shows the signal disappearing after Ron
shut off the transmitter to verify.
Bottom: Ron's reception of my signal.
Using a 0.031 Hz bandwidth, I received Ron's signal with a signal-noise
ratio of about 10-15dB while Ron received my signal with about a
20-25dB signal-noise ratio.
The above pictures are screen-shots from the Spectran program.
Note that due to contrast setting and configuration differences, our
screens look different, with my screen (top) using a vertical
"waterfall" while Ron (bottom) used a horizontal one.
Click on an image for a larger version.
For detection, we used narrowband techniques utilizing the Spectran
program. Feeding the audio output from the optical receivers into
sound card input of our laptop computers, we could use this
configuration to detect extremely weak signals. For our tests, we
chose an 8kHz sampling rate and a 0.031 Hz detection bandwidth.
The high-power LED transmitter is strong enough that when operated in
the dark, one can see the beam emerging from the transmitter's lens via
Rayleigh and dust scattering
- a fact that aids in aiming the antenna as one simply points the
shaft of light at the object of interest - in this case, Grandeur
Peak. Once aimed, the receiver - which is mechanically coupled to
transmitter - is also precisely aimed at the same object and with both
having aimed at Grandeur Peak, Ron transmitted first while I "listened."
Using such a narrow detection bandwidth, it takes about a minute before
any detected signal appears on the computer's "waterfall" display, and
after several minutes of Ron's transmitting, I didn't see anything
appear. At this point Ron re-aimed his transmitter and we tried
again: To my
surprised, I soon saw a weak - but distinct - signal appear on my
waterfall display. To verify, he shut his transmitter off, and
signal subsequently disappeared see
the top image of Figure 2
Quickly, we turned the link around: After resolving a minor
configuration problem on his end, Ron soon saw a very distinct signal
appear on his computer's waterfall display (the bottom image of Figure 2)
and it, too,
disappeared as I turned my transmitter on and off. Note that
due to the sampling rate on Ron's computer being slightly off, he
received my 262 Hz signal at "249" Hz.
At this point, we decided to try to "peak" our respective signals by
slightly re-aiming our gear. Even when we weren't
changing the aiming of our
gear, we were somewhat puzzled by the rapidly-varying
levels which seemed to bounce from "as good as before" to "completely
gone" when we noticed that weather was starting to move in and the
mountain was being intermittently obscured by clouds. The final
occurred when a light rain started to fall across the Salt Lake Valley,
wiping out the optical path completely - so we gathered up our things
and went back inside.
The evening of February 5, 2009 was breezy, but clear in the Salt Lake
Valley and I could see Grandeur Peak in the moonlight.
Ideally, one would do "mountainbounce" tests sans moon, but we really
didn't have any say in the matter that night - and we were wondering if
the extra illumination was likely to cause a significant amount of
"desense" - that is, additional thermal noise from the moonlight.
Since a lunar eclipse hadn't been scheduled for that evening, we really
couldn't make an "A/B" comparison, so we plowed ahead.
With our prior experience, we knew more-or-less where to aim our
transceivers: At Grandeur Peak itself - even though recent
warm-ish weather had stripped the mountain of some of its white snow
coat. Dragging our gear outside again - Ron under his back patio
and I onto my deck - we began to alternately squirt red photons at the
mountains while the other "listened."
Again, we chose the 262-ish Hz "middle C" tone, with it being some
distance away (frequency-wise) from 60/120 Hz powerline harmonics from
the city's' lighting. Almost immediately, we detected each
others' tones using Spectran and made minor adjustments to peak our
signals: Since the optical transceivers have parallel transmit
and receive "beams", peaking one naturally peaks the other, and the red
shaft emerging from the transmitter simplified the aiming
greatly! At my end, I wielded a night-vision scope that made it a
bit easier to spot the red shaft of light in the clear air and where it
seemed to be aimed.
On this occasion, I was surprised to be able to very
easily detect Ron's signals. Switching from the narrow 0.031 Hz
bandwidth to progressively wider ones, I could still discern his signal
on the waterfall display even with a >1Hz detection bandwidth!
Switching back to 0.47Hz bandwidth offered a good compromise between
sensitivity and update speed: The narrower bandwidths offer
better effective sensitivity by virtue of selectivity, but being
narrower, any changes in signal strength are reported more-slowly and
with a greater delay: At 0.47 Hz, I could make a change and see
the results, having only to wait 3-4 seconds rather than 30 seconds or
With "good" signals visible at a 0.47 Hz bandwidth, Ron decided to send
his callsign using "QRSS" - that is, very slow Morse in which a "dit"
lasted 5 seconds (hence the designation "QRSS5) and a "dah" lasted 15
As it turns out, even though Ron's callsign is shorter than mine,
trying to maintain concentration in sending so slowly is quite a
challenge. At higher speeds, one automatically forms the letters
without thinking about how many dots or dashes there are, but at this
speed, that's no longer the case! Not only is it necessary to
count, timing the length of the dits and dahs in seconds, but one must
remember where, in the sequence, one is, as it is very easy to get
lost! Being with Ron on the telephone at this time, I decided to
simply set the phone down and avoid any temptation to distract him!
About 4 minutes later, he finished and I had his complete callsign
displayed on the screen, as can be seen in the top image of Figure
On this image, one can see, on the horizontal
waterfall display on the lower half of the image, the dots and dashes
spelling out "de K7RJ" ("de" meaning "from.") You can also
see a pair of bright lines near the top and bottom of the waterfall
display: Comparing the frequency scale on the right side of the
waterfall with the "pips" on the frequency display on the top of the
image, one notes that the frequencies of these lines are 240 Hz and 360
Hz - with an additional "pip" (not shown on the waterfall) at 480 Hz -
all of these being harmonics of the powerline-related 120Hz frequency,
with this energy coming from city lighting. The smaller pip, at
262 Hz, is Ron's signal. During this test, I used the built-in
audio filtering of Spectran to see if I could hear his signal via
ear: While I might
have been able to discern his
"key-down" periods, the results were inconclusive and I doubt that I
could have accurately detected them.
Top: The "QRSS5" signal bearing Ron's callsign, as
received by Clint.
Bottom: Clint's callsign, as received by Ron.
The differences in brightness and contrast between the images are
due both to settings on our computers and the relative strength of the
received signal. A "brighter" version of the lower image can be
Click on an image for a larger version.
After successfully completing this transmission, we decided to
double-check the peaking of each others' systems to assure the
best-possible alignment. While Ron was doing this, he suddenly
lost my signal completely and, conversely, I could no longer see
his. After several minutes of thrashing about, trying narrower
bandwidths and re-checking the clarity of the air across the valley to
the peak, I finally noticed that my transceiver was off-point:
This probably happened either from my bumping it as I went in and out
through the door between my kitchen and deck, or from a gust of wind
that nudged it very slightly: It had only been off-pointed by a
degree or two, but that was enough! After re-aligning our ends
again, we both acquired each other's signals - albeit seemingly weaker
than before - and I decided to send my callsign via QRSS.
Rather than trying to maintain the concentration required to send my
callsign at a QRSS5 rate - something that would have taken about 6
minutes - I quickly prepared, using a sound editing program, a .WAV
file with the dits and dahs generated via computer, sending "de
KA7OEI". The tones from the playback of this file would be
inputted to the optical transmitter, while I could stand back in the
warmth of my house - and while Ron continued to stay outside, in the
cold... With Ron's setting his bandwidth to 0.12 Hz, I let the
file play while Ron saw my callsign slowly worm its way across his
waterfall display, as seen in the bottom portion of Figure 3
For whatever reason, his signals from me weren't as good as they had
been earlier in the evening, but as can be seen from the picture, the
callsign is readily discernible. Again, note that the sound
card in Ron's computer was running about 6% high, causing his reported
frequency to be low by the same amount.
It was noticed that when sending via the computer, my optical
transceiver was more fully-modulated than it was with its built-in tone
generator. The reason for this was that I was running my audio
too "hot" into the modulator's input, causing it to distort to a
clipped square-wave. As it turns out, this increased the amount
of spectral energy at the modulated frequency somewhat, improving Ron's
"copy" of me slightly! If you look carefully at the bottom
picture of Figure 3
, at the far-right edge, you'll notice a
slight downward frequency shift indicated by the horizontal line moving
down as well, just after the completion of the callsign. This was
due to the QRSS keying in my audio file being generated at exactly
262.0 Hz while the modulator's internal tone generator was set to the
more musically-correct 261.6 Hz. It also points out, by virtue of
the "brightness" of the lower-frequency trace, that this signal was, in
fact, slightly weaker!
A few other experiments:
Having completed our callsign exchange, we were curious about how
different pitches of tones might be affected by the amount of thermal
noise that we were both experiencing from urban lighting. We had
originally chosen 262 Hz (middle-C) because it was a reasonable
compromise: It was high enough to be audible, but not so high
that it might drop into the higher-frequency roll-off frequency of the
receiver. Because our noise floor was limited by the ambient
light - from both the city and the moon - absolute receiver sensitivity
wasn't likely to be a problem!
In addition to the 262 Hz, we also tried 31 (a B), 41.1 (an E), and
1318.5 (also an E) to see if there was an appreciable difference in the
received signal-noise ratio. The results were inconclusive - that
is to say, one frequency range didn't seem to have any obvious
advantage of any other, although there seemed
to be a
perception that the low frequencies (31 and 41 Hz) fared slightly
better than the higher (262 and 1318 Hz) tones.
A few comments about these experiments:
"Optical Noise Floor" degradation:
At this point there are a few things that can be said about these
experiment and their results.
As can be seen from Figure 2
the signals were fairly
weak. Also, as expected, the signal-to-noise ratio that Ron
received from my signal was significantly better than that which I
received from his signal. While some of this (2dB or so) is due
to my optical transmitter's having better optical flux, most of it has
to do with the fact that I was receiving Ron's signal across a
population center that was radiating a significant "glow" of
light! When setting up the receiver, I observed that my the noise
floor of my optical receiver was about 10dB higher when pointed across
the valley than it was when I pointed it down to the ground - this
being due to the scattered light in the atmosphere.
This added noise was mostly in the form of a white-noise "hiss" -
although there was a strong component of 120 Hz energy and its
components, with one of the strongest peaks being at 360 Hz.
While our "Middle-C" tone frequency was sufficiently far away from any
of the 120 Hz harmonics, there was absolutely no escape of the
As mentioned before, Ron lives fairly close to the mountain, near the
edge of town and had relatively little inhabited area between him and
the peak. Nevertheless, he also was detecting a significant
amount of 120Hz energy and its harmonics - but this was mostly from the
fact that the snow-covered mountain was being bathed in city lights!
It should be noted that the transmitters did not
have anywhere near enough power
to impart even even a hint of a red cast to the mountain, but it could
accurately be said that anyone who was on
would, without being prompted, immediately notice either
red light source! Past
experience has shown that even amongst a sea of city lights, these LED
transmitters are conspicuous not only because they are red, but also
because they are bright, as compared with other urban light sources.
Mismatch of "spot size":
Another factor that caused a bit of degradation of the signals was
likely due to the difference in the "spot size" on the mountain.
Consider that our transmitters and receivers have about the same
beamwidth - about 1/4 of a degree. Clearly, with Ron being at
1/6th the distance, the "spot" that he was projecting onto the mountain
was much smaller than the spot that I would have been. The
converse would also be true: My transmitter would have made a
much larger red spot on the mountain than his!
This is also true of the receiver: Much of the "large" spot
produced by my transmitter was being completely missed by the
relatively "small" spot of his receiver - and the converse would be
true of my receiver and his transmitter! What this means is that
he was missing most of the light that I was shining on the mountain,
while my receiver was being "diluted" by being able to see that "extra"
part of the mountain that his transmitter didn't illuminate! (Note:
An advantage of a small spot size is that less stray light from the
city and moon would be intercepted - but this has the disadvantage of
being more difficult to aim!)
One way to "solve" this problem would have been to pick a different mountain - one that
was equidistant to us both: Our respective "spot" sizes would
have been similar in that case! The reason why we did not do this
was mentioned before: We simply chose the mountain that both of
could conveniently see from our back yards! Another possibility
would be to modify the receivers and transmitters to modify their
respective beamwidths to match the specific situation - but that
requires a bit more planning - not to mention flexibility of the
optical gear - and our gear wasn't designed for that.
The "Aspect" of the mountain:
As can be seen from the top picture in Figure 1, Ron could see
only the top portion of the Grandeur Peak mass from his location while
I could see pretty much all of it. It is possible that if he were
to choose a different location where he could see more of the upper
portion, we might have experienced better signals. Also, it might
have been possible that if he'd aimed at the lower portion that both of
us could see, that, too, might have also yielded better results!
At some point, we hope to try this again and find out!
As has been mentioned, we used the Spectran
program, useful utility written, in part, by Alberto, I2PHD. This
program includes not only the spectrum and waterfall displays, but
other features such as audio bandpass and notch filters, noise
reduction, and the capability of manually and automatically capturing
One thing that Ron noted immediately was that the 120 Hz harmonics of
the received light were "off" frequency, according to the
display. This was a result of either inaccuracies of the sampling
rate of the sound card in the computer that he was using, or some
operating-system induced offsets caused by internal sample-rate
the measured-versus-actual frequencies of the 120 Hz harmonics, he was
quickly able to determine a "correction factor" to determine at what
frequency our 262-ish Hz tone would appear - and this is why his tone
frequency is "off" in Figures 2 and 3.
Remember: If you plan to use Spectran - or any other
program - for precise frequency measurement, make sure that you measure
a known-accurate frequency so that you can discover any offsets and
correct for them!
This was, by no means, spectacular DX, nor did we really tried to
optimize our results. It was, simply, a fun experiment to do and
it required almost no effort on our part to do this, aside from lugging
gear just outside our houses!
I must confess, however, that Ron's hardship was the greater one:
While I set up my optical gear on my back deck, my audio cables easily
reached into my housewhere I operated the computer from the inside -
only needing to step outside briefly to make adjustments! Ron, on
the other hand, had to stay outside in the cold!
I'd like to thank those that
- Ron, K7RJ, at the other end.
- Elaine, N7BDZ, Ron's better half, who took the photos at Ron's end.
Top: Ron, operating "Mountain Bounce" from his patio,
with the optical transceiver pointed up toward Grandeur Peak.
(Photo by Elaine, N7BDZ)
Bottom: Clint's setup on his back deck. This photo
was taken just after we finished for the evening - when a light rain
started to fall across the valley
Click on an image for a larger version.
About the optical gear:
- Distance between Ron's house and Grandeur Peak: About 2.11
miles (3.40 km)
- Distance between Clint's house and Grandeur Peak: About
12.38 miles (19.93km)
- Total path distance: About 14.49 miles (23.33km)
- Actual distance between our two houses: 10.54 miles (16.96
- Since we weren't actually bouncing light off the peak - but
rather the sides of the mountain below the peak, the actual
distances were probably slightly different.
Equipment common to both sides of the QSO:
LED-based Optical transceiver used at Ron's end:
- The LED was amplitude modulated with a current-linear modulator
with a resting current of 1.1 amps. Details of the
modulator are here: LED_linear_modulator.html
- The transmit LED in both cases was a Red Luxeon III emitter
module (Lumileds M/N: LXHL-PD09) epoxied to a heat sink.
- The optical receivers were my "version 3" design, described
with both receivers using BPW34 photodiodes.
- Audio interface units, incorporating audio amplifiers, audio
recorder interface, audible S-meter, and a few other features were used
- details are here: optical_comm_audio_interface_device.html
- Both transceivers have separate and identical TX and RX lenses
LED-based Optical transceiver used at Clint's end:
- This enclosure is described in detail here: Optical_enclosure_first_version.html
- Lens size: Unmounted, the Fresnel Lenses are 250mm x 318mm
and have a focal length of 318mm. The mounting frames vignette
the lenses by about 10mm in each dimension, so the available lens area
is about 240mm x 308mm. Each lens is protected by a sheet of
Plexiglas and the front surface has been coated with a protective
polymer to prevent scratching and moisture accumulation.
- For optimal far-field optical flux density, a glass PCX
(Plano-ConveX) lens is used in front of the LED to appropriately
illuminate the Fresnel, the LED-Lens distance being set empirically for
- The enclosure is described
- Lens size: Unmounted, the Fresnel lenses are 404mm x 430mm
and have a focal length of 229 mm. The mounting frames vignette
the lens by about 10mm in each dimension, so the available lens area is
about 394mm x 420mm. Each lens is protected by a sheet of
Plexiglas and the front surface has been coated with a protective
polymer to prevent scratching and moisture accumulation.
- For optimal far-field optical flux density, an optical acrylic
DCX (Double-ConveX) lens was reground to an aspherical shape to provide
optimal illumination of the Fresnel. This turned out to be
necessary owing to the very short focal length of the lens that made it
difficult to efficiently illuminate the lens. After adjustment,
this LED/Lens combination produces about 25% higher far-field flux than
the other assembly, with an almost identical half-power beamwidth.
to the KA7OEI Optical communications Index page.
If you have questions or comments concerning the contents
page, feel free to contact me using the information at this URL.
Keywords: Lightbeam communications, light
laser beam, modulated light, optical communications, through-the-air
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
This page and contents copyright
2009. Last update: 20091222