This page is really an extension of my amateur radio projects, since afterall, hams are allowed to use all frequencies above 300 GHz as the "highest band". That includes visible light in the THz region. I'm particularly interested in non-line-of-sight propagation at these frequencies, which seems to be possible despite the apparent oxymoron. You are free to use and adapt the circuits here provided that they are not used for commercial purposes and that you give credit where credit is due.
LaserScatter is a program designed for non-LOS laser communication. It uses your computer's soundcard to send and receive messages using long-term signal averaging. Click here to go to the laserscatter webpage.
This is a picture of my signal radiating from a 100mw IR laser (expanded!!) and scattering off of an overcast sky. The multiple traces are caused by the square wave modulator that I am using (see below). I used Argo by I2PHD and IK2CZL to receive this signal. On the transmit side I used QRSS by ON7YD to key my 18 Hz modulator through the RTS line of the serial port. The "dits" are 3 seconds in length, which is much slower than is actually required for this strength but it illustrates the concept. The signal is actually strong enough to be audible if it were at a higher modulation frequency. The recording is the signal sped up 10x. The 1200Hz "carrier" is the sound of the city lights which are also illuminating the cloud. You can also hear my local streelight with a flakey light sensor turning on and off in this recording. The CW is quieter and at an audio frequency of 150 Hz in this recording, so you may need to turn up the bass response of your speakers to hear it well. This signal is roughly 30dB above the background noise in this bandwidth, so there is plenty of SNR to give and still be copyable, especially with slower signal rates. The transmitter and receiver were located in separate windows for this test, separated by about 15 feet. The signal is at maximum when the TX and RX are pointing parrallel to eachother, which leads me to believe that this is cloudbounce and not a near-field phenomenon. I used LF frequencies beacause the SNR seems to be much better in this range using the PGP front end, and it gets me away from the spectral bleedover from 60Hz, 70 Hz trash from my monitor, 120 Hz from lights, etc. The receiver was a 4" lens and a PGP front-end with a 20mm^2 photodiode drivng an LM386 in "bass boost" operation.
Work has been slowly coming along. I've managed to acquire some laser diodes and safely connect them to my modulator. I found a pair of 120mw/780nm diodes at digikey to use in some troposcatter tests. Of course such devices have the potential to be quite dangerous, so it became necessary to create a beam expander to decrease the power density in case of accidental direct exposure. The expander I devised widens the beam to about 3" by 1" (remember, diodes emit a rectangular beam). Though still not a toy, this greatly reduces the chance for any accidents. It's constructed out of PVC pipe and seems to work very well in my few cloudbounce tests.
The expander uses the adjustable lens of a laser module to diverge the beam and a second lens to recollimate it at a larger diameter. Once takes the adjustable lens and screws it closer to the diode so that the beam is about 3/4" the diameter of the lens when placed at the larger lens's focal point. The screw joint allows precision adjustment of this distance.
The 10mw visible diode I am using is the Sharp GH06510B2A (digikey #425-1806-ND, $12) and the IR diode that I am using is the Sharp GH0781JA2C (digikey #425-1809-ND, $25). It's rated at 120mw, so this thing is quite dangerous if not respected, and it was my motivation for building the beam expander. When collimated it produces a "weak" red laser, which is probably due to some emission in the visible range. Do not be fooled! The 784nm stuff is quite invisible and *very* dangerous. A 5mw laser that is collimated to a point on the wall can be uncomfortable to view at close range--just think what another 15dB would look like....
Both of these diodes are in the 5.2mm package and require some heat sinking. Fortunately you can buy a housing kit from digikey that contains and adapter to the 10mm size, housing, and collimating lens all for $18 (digikey #38-1000-ND). To mount these goodies, I cut out a 2"-diameter PC-board "doughnut" and glued the housing to that (think of it as a 2" washer with 1/2" hole). This doughnut is squeezed into the threaded PVC adapter (see pics above) and pressed against the short section of 2" pipe to hold it in place for the expander. This 2" pipe also houses the current limiting resistor.
I found the pin-diagrams to be a little confusing on the datasheets for the above diodes. The pin layout is described as if you were looking at the window from above. If you are actually looking at the soldering side of these devices, you would make your connections like so:
O --+V
O --GND (case)
O --PD (visible), or NC (IR)
I tried to make the "laser heads" swappable in my expander. The screw joint allows one to adjust the distance between the laser and the 4" collimation lens, as there are bound to be slight differences in focal length for each wavelength. Right now I have a pair of heads for the 120mw IR diodes and a 10mw 660nm head all adjusted to about 90% I(op). Since the Ramsey PWM kit also runs on 5V I can just plug that in there too, provided I don't sink too much current from its regulator. Before I start modulating the laser with MCW or PWM, I adjust for Iop by applying a steady +5V to the diode (the TEST switch) and measuring the current while I adjust the resistor's value *down* until I reach the desired current. The multi-turned pot ensures that I don't exceed I(op) on accident--provided that you start with a high value and work down! Don't exceed I(op) for a microsecond! Once I have the pot set, I reconnect it back to the modulator and it's ready to play, be it square-wave MCW or PWM. I used a 3/4", 15-turn, 100-ohm pot (digikey #3006P-101-ND, $1.75/ea) in parallel with a 100-ohm resistor (1/2W) to give me smooth current adjustment. The modulators I am using are the "super easy modulator" you'll find down the page. The only difference here is that I am not using a stock laser pointer, but rather makimg my own out of a resistor and a bare diode.
After duct-taping a homebrew 20X telescope to the top of my expander, I'm able to observe my red laser dot on billboards up to a mile away, even with all of the city lights. It is still a dot, which means my divergence has been reduced considerably. Though nearly useless for LOS work--how would you like to aim to try to aim something that is 2ft wide at 10mi?-- I have a hunch that this will be very helpful for over-the-horizon laser tests.
front--detector end of the assembled unit
back--shows power supply (9-15V) and headhone jacks
the guts--the K3PGP front end and LM386 headphone amp
front end--closeup of PGP front end. No comments on contruction technique
mounting--the whole shebang inserted into the 4" PVC pipe
The only modification that I made to the front end was a larger bypass cap on the supply input (500uF). I found that running the PGP front end directly into a 386 amp would cause a low frequency oscillation (probably the 2N5088 amplifying transistor). This would be better solved with a buffer amp before the LM386, but this gets the job done. The PGP receiver was built "dead bug" style on a section of single-sided PC board. This allows for the photodiode to fit flush against the insulated side (you can see the holes in the center of the board where the diode leads connect to the circuit). The LM386 amp was built on a little ratshack IC protoboard. Last night I finished another LM386 amp built "dead bug" style on a much smaller piece of ground plane. I think I like that a little better, though it's not as quick as the protoboard method--the IC actually looks like a "dead bug" with it's leads sticking up in the air, in case you were wondering where the name came from. I'll try to get a picture up here.
The housing is made from two 12-oz tunafish cans, continuing in the spirit of the "tuna-tin transmitter"....Actually, they were chosen because they happen to slide snugly into a piece of 4" schedule 40 PVC pipe, which also holds a 4" glass lens. Despite the rather crude construction and relatively high-noise components, the system works very well (see recordings and measurements below). I've certainly improved my soldering skills over the last few months working on this project. Next step is to improve the "workbench"--a small table in the corner of my bedroom of my tiny apartment. Now that we have working TX and RX circuits, it's a matter of improving them and figuring out the best way to organize it all. I have a transmitter more-or-less designed that employs a sidetone, computer control for QRSS/DFCW, and a "LF" switch for long-term signal averaging under 60 Hz. It sounds rather complex, but really isn't--you can do alot with a couple switches a decade counters. It's all based on the simple design below--which works perfectly well for normal MCW operation with any baseband RX system. With all the folks working on this here in Rochester, I think we'll be be making contacts very soon!
Given that a receiver requires one to consume 24oz of tunafish, I suggest making some tuna burgers:
1 pack of saltine crackers
2 twelve-oz cans of tuna
1/2 chopped white onion
Drain the water, from the tuna and mix it with the onion in a large mixing bowl. Crush the crackers into a fine powder and add them to the tuna until you reach a consistancy that you can form into patties. Prepare as you would for hamburgers! Makes a nice addition to your club construction party!
I'm only half joking.
This is an spectral display of light bouncing off of a nearby cloud, as received in my downtown apartment. The laser was audible, as was the 120Hz buzz from city lights. As one can see the first four harmonics of the streetlights are rather loud (120, 240, 360 and 480 Hz) while higher order harmonics are not strong at all. This may suggest that a modulation frequency above 500Hz would be a better choice, since that would allow a band-pass filter to eliminate much of this interference. We are also looking into filtering the light at the receiver.
The transmitter in this recording was a 3mw laser pen chopped at approximately 410 Hz by a CPU fan--explains the broad peak in the FFT. The receiver was a K3PGP design front end wit ha 4" glass lens, driving a LM386 audio amp.
I finally got around to making a very simple laser modulator for sending modulated CW (MCW). The design is very simple, and borrows some ideas from G0MRF's modulator. A 4MHz xtal oscillator is divided by 2^13 to create a square wave output at 488 Hz (Pin 2 of the 4060 chip). Since the 4060 is a standard CMOS logic chip, it has very little current handling capability. This is fed to the gate of a MOSFET transistor. The MOSFET behaves just like the valve on your bathroom faucet: When the output of the 4060 is high, current is allowed to flow from drain to source in the transistor. This valve turns on and off with each pulse from the 4060, i.e., on and off 488 times per second. While the key is closed, the laser will key on and off at 488 Hz. This is is easily detected with a baseband light receiver, like the one below. The 1K resistor before the MOSFET is to protect against the inrush current due to the gate capacitance of the MOSFET. I don't think it would be a problem, but better safe than sorry.
1) Make sure it is *impossible* to create a short that bypasses the laser module since this would run *alot* of current through the MOSFET and your regulator. You say "DUH", but given that the battery contacts for most laser pens/keychains are way down in a tube it ain't so hard to do. I'm using insulated allegator clips make my connections--one to the spring and one to the case. (Makes me nervous.) Keep in mind that most laser pointers/keychains are positive ground, and don't forget to "premenantly press" your laser pointers button.
2) The .1 cap on the 5V supply to the 4060 should be placed as closely as possible to the chip (Dave's suggestion).
3) There is an RF oscillator in there, and I built in on one of those radio-shack IC boards that fans out like a spider (i.e., a small antenna). Probably would be wise to put this thing in a metal case so it doesn't talk to you on 75 meters.
4) This circuit simply replaces the batteries in the laser pointer. Many pointers run on 3V (two 1.5V cells), so a zener or a couple of diodes to drop the voltage to your particular pen should do the trick. A +5V regulator (7805) supplies everything.
1) Pin 2 of the 4060 gives 488 Hz (see the audio spectrum of light above). This is uncomfortably close to the 4th harmonic of our streetlights (120x4 = 480 Hz) so there is nothing wrong with selecting a different output or a different xtal for the time base. Over in Europe it's not such a big deal (the fourth harmonic is at 400 Hz), but I've found that my cloudbonced signal has a nice 8Hz beat at low signal levels in the presence of city light! I've looked at the amplitude spectrum of light, and it's much clearer above 600Hz. I'll put some of that info up here when I get a chance. I'm also looking into the possibility of using modulation frequencies below 60Hz for QRSS and DFCW experiments on non-LOS paths. All I would have to do is switch in another decade divider. More to come.....
2) I want a switch that bypasses the MOSFET to key down the laser at 100% duty cycle to aid in aiming, when we get to that point!
3) --create serial port interface for QRSS CW (replace key with a relay). DFCW will use a frequency that is a few HZ away and one can toggle between the two oscilators. I bought some 4.032 MHz xtals for this purpose.
4) This circuit can probably power a higher-power laser diode, provided a current limiting resistor is used to limit the output WELL BELOW it's maximum. Given that Ith and Iop can be a few mA apart in many diodes, this might be asking for trouble. Ideas here are welcome...
I completed a prototype receiver with optics last night. It incorporate the soup-can housing for the K3PGP front end that you can see below. The 4" glass lens is housed in standard 4" ID PVC pipe that is available at any hardware store. I found that I could line the pipe with the black flockin paper and then insert the lens. The paper creates just the right amount of tension to hold the lens in place. To adjust the distance of the lens to the detector, one can slide the flocking paper back and forth or move the soup can in and out. Once I found the right allignment I simply taped the back of the flocking paper to the inside of the pipe to keep it from moving. The pipe reducer hold the two piece together. A better way to do this would be to use two 4" coffee cans instead of soup cans since they will fit securely inside the pipe on their own, thus eliminating the use of the reducer. I can't drink coffee that fast and I had the soup cans already assembled.
prototype.jpg -- Schematic of the prototype receiver
receiver1.jpg -- Picture of the assembled unit
receiver2.jpg -- Picture of the unit showing the lens housing and the soup can feed...which looks surprisingly like an old TVRO LNB....
Again, I've just been testing it inside the apartment, but it seems to be very selective in a spacial sense--it hears what it is pointing at an nothing else. Part of this is because the lens is mounted several inches into the pipe. This keeps light from scattering off the lens at oblique angles. This may be crucial if anyone wants to use this thing in an urban/suburban environment. It makes a noticeable difference. (thanks to Tom, WB2BPE, for this suggestion). It's hardly a piece of precision equipment, as the tools involved in it's construction were, tape measure, hacksaw, and scotch tape, so there is room for improvement. Still, I can't wait to try it out once my tripod gets here!
The front-end itself does not put out quite enough audio to drive headphones directly, and in extremely dark conditions I will need to amplify the output for the soundcard itself. I'm looking for a very low noise audio amplifier. I'm thinking of something in the TL07x series, but I'm open to suggestions.
Recordings
Spectrum under dark conditions--This file was created by blocking all light from the photodiode and averaging two minutes of audio at a 44.1 KHz sample rate...which by the way, takes forever on a Pentium 166 with 24 Mb of RAM....It should represent the noise from the dark current of the diode, plus spurrious signals generated within the soundcard as well as possibly some harmonics of 60Hz that aren't completely surpressed by my 12VDC power supply. I need to do some S/N tests to figure out if the region from 20-500Hz is that much more sensitive or if there is just more noise there. It also possible that some photons are leaking into the can through the back, though there probably aren't many! Also, this circuit is prone to microphonics, so there could be some LF vibration present--the highway is 100 yards from here. The rolloff starting around 16 KHz is caused by the program filtering out energy near the nyquist frequency
Pulse-width modulation--Recording of my voice on a 3mw laser beam scattering off a wall on the other side of the room. I did this in the afternoon with sunlight creeping in through the blinds. This seriously degraded the NF of this receiver. I'm transmitting a 16KHz pulse-width modulated signal from the Ramsey LBC6-K laser transmitter. That, in combination with the fine audio provided by the electret microphone produce what you hear here. Here is a spectrogram of this signal in which you can see most of the energy is in the 16KHz carrier. You may or may not be able to hear that in the above recording, depending upon how old you are and how many rock concerts you have gone to. Here is the same signal filtered to a 3KHz passband.
QRSS CW at 3sec/dot: This is a picture of a laser diode being modulated at 45 Hz using extremely slow morse code and detected with signal averaging. Actually, this signal is strong enough to be copied by ear if we could hear 45 Hz resonably. I'm modulating laser diode with my soundcard by giving the diode 5V of DC bias while adding a 2-Volt P-P sine wave. Despite the fact that the diode came with three "button" cells for a power supply, 6V on the peak of the cycle doesn't seem to kill it. (It has no current regulation besides a 47-ohm resistor in series with the laser diode). I seems to modulate with reasonable linearity. I tried voice, but it was quite muffled, and odd harmonics were fairly substantial--no surprise there.
Why 45 Hz you ask? Given the interference from multiples of 60 and 120Hz from lights, and 60-80 Hz refresh rates on my monitors (extremly strong in the picture) there are all kinds of birdies, weak and strong, filling the spectrum in the normal audio range. There seems to be little QRM in the 20-60 Hz range, however, besides some LF microphonics that you can get by tapping the case as I did in the picture above.
I did this rather unscientific test in my apartment by shining a laser dot on a wall 20ft from the detector. I did this in mid-morning daylight with the blinds down and the apartment lights off. Clearly, the detector was still saturated under these conditions and the S/N ratios in the above example don't mean too much. Under dark conditions, that same laser dot on the wall will overload the front end of the detector on it's own.
If you have one of those HeNe or some other tube-type lasers, you really won't be able to modulate the laser in the same way as a laser diode. One way to send CW is to chop the beam at an audio rate and use a shutter to release the beam for forming the morse elements. I was wondering if this simple method was frequency-stable enough for QRSS CW or other signal-averaging detection system. Here's what I found:
Spectrogram of a fan under still air--This is an ordinary 12V computer CPU fan. Time in seconds is on the x-axis, frequency in Hz is on the y-axis.As you can see, wanders over a range of about 10Hz, or 2% of the audio frequency. This would be perfectly acceptable for normal CW operation, but probably not much good for QRSS. If one is going to use this outside in the open air, the fan will need to be sealed off from the wind. One could encase the fan in some sort of plastic or glass housing, but then you have to worry about attenuation and scattering from the material. It might be best to place the fan close to the output of the laser and put the whole unit in some sort of wind sheild, with a hole just large enough to pass the beam.
Here's a 12V fan from a computer power supply. It is considerably larger than the CPU fan and spins a little slower, thus creating a lower audio tone. It too wanders about 2% of the audio frequency. This modulation method should be just fine for using a HeNe or other tube-type laser for CW QSOs, provided that a one can find a way to form the cW elements (shutter, deflector?).....
