DX Crystal Sets
I built my first crystal set at age 10 or 11 without a lot of knowledge of how it worked. The antenna was the nearest downspout and I was thrilled to receive WLW, the 50KW giant in Cincinnati. My next try was 40+ years later for my grandson. The original objective was to get him interested in electronics, but the result was a radio building project that went on for a couple of months. My grandson got his radio and so did I.
The process of coming up with a suitable design was not straight forward. After consulting the internet, a couple of simple designs were prototyped, starting with a single tapped cylindrical coil with close-wound magnet wire, which worked, but had poor sensitivity and selectivity. Then a dual tuned version with the turns spaced to increase Q improved performance. Finally, I hit upon Dave Schmarder's excellent web site, http://makearadio.com, at which point things began to come together. Dave has spent an enormous amount of time and energy to documenting the making of crystal radios and there is a wealth of knowledge on his site.
The next major step was to incorporate Litz wire wound on spider web coil forms. The selectivity increased and many more stations could be heard. I was still not satisfied with the DX performance and it turned out to be a problem with my hearing, not the set. I guess my low level threshold of hearing has deteriorated, making it difficult if not impossible to pull out the weak DX, even with sensitive earphones. The solution was a simple audio amplifier after the detector stage. I look upon this addition as if it was a very sensitive set of headphones. Not quite a pure crystal set radio, but a solution that worked for me.
The final design is shown in the accompanying schematic diagram. Again I borrowed extensively from Dave's various designs, adding a couple of my own pieces here and there. A list of the web sites I used in the process is at the end of the article. A larger version of the schematic is at
DX Radio Schematic .
DX Radio Images
Notes on the design:
1. Antenna Tuner
The antenna tuner is a straightforward tank circuit with a ganged antenna capacitor. Trimmers were added to allow dial calibration and compensation for different antennas. To calibrate, I adjusted the coil inductance value to allow tuning at the low end. The trimmers set the upper end of the band and the 365pf trimmer adjusts for various antenna impedances to maintain dial calibration. Not shown is a bypass switch on this trimmer. The coil is on a spider web form made from HDPE (high density polyethylene), cut out on a scroll saw. A template for my spider web forms is at Spider web coil form. The cutouts within each section are probably an overkill, but every little bit helps. The wire for the coils is 175/46 Litz. A summary of the specific coil winding data is at the end of the article.
I measured the Q values of the unloaded coils with the variable capacitor used in the set using a very loosely coupled exciter coil connected to an RF signal generator and measured the response with a 100MHz scope, whose pickup was a small loop several inches from the coil. While every attempt was made to eliminate coil loading, these Q values should be taken as approximate only.
For comparison, a similar spider web coil wound with #22 magnet wire only yielded Q values in the 150 to 170 range. Quite a difference. Another spider web form made from Lexan with the 175/46 Litz wire yielded Q values from 360 to 500, indicating that the material is important, but the wire type dominates the determination of Q values. Litz wire is formed of many strands of very fine wire all insulated from each other. The result is a large conduction area at FR frequencies because of the skin effect. Resistance decreases and Q is enhanced. Litz is expensive (175/46 is about $.30/ft), but very much worth it.
The Detector circuit is a little more complicated. I opted for an inductively coupled design built as a separate unit. That way the coupling could be varied by changing the separation of the coils. 9" to 10" seems to give maximum volume with good selectivity. Greater separation increases the selectivity, but the volume decreases. The circuit started out with a tapped coil takeoff to allow impedance matching. After experimenting with a coil having multiple taps, the arrangement shown was the best compromise. Alternatively, the circuit also incorporates the Hobbydyne TM coupling arrangement with a differential capacitor and compensating trimmers. The differential cap circuit works slightly better than the tap configuration. I find that I keep one setting most of the time, so a simple coupling capacitor from the top of the coil could be used to simplify things a lot. Experimenting with various values or using a small trimmer would make the design much easier to build. The final design, with the differential coupling capacitor has very good selectivity, probably limited by the coil Q.
The actual detector is a 1n34a germanium diode selected from several to give the loudest response. A 27mH choke provides a DC return path when the capacitor coupling to the top of the coil is used and the .oo2 capacitor provides RF bypass after the diode. The 250K pot provides the audio amplifier with variable signal level to compensate for very loud stations.
An audio amplifier with a LM386 makes using readily available headphones possible. The LM386 will drive a small loudspeaker on the strongest stations from a 9v battery as well. The 10K input and .002 shunt capacitor roll off the audio at about 7.5KHz. Two gain setting are provided to give extra boost for the speaker. Almost any low impedance headphones can be used. The ones I employ most are walkman-style with approximately 30 ohms impedance. More than enough volume is available across the broadcast band with a 30m random wire antenna.
The radio was built as four units, Antenna Tuner, Detector/Amplifier and two Wave Traps. All were constructed on wooden bases with a chassis of HDPE and a garolite front panel. The HDPE chassis reduces any stray capacitance or resistance paths a wooden base would add. The main tuning capacitors were fitted with 6:1 reduction drives for ease of fine tuning. The Antenna Tuner capacitor is set back and electrically isolated to reduce hand capacity effects. Brass hardware was used throughout. The dials were fabricated from HDPE and after calibration using an RF signal generator, an overlay was printed on clear stock using a laser printer and cemented to the dial. Many of the details can be seen in the accompanying photos.
The in-line wave trap consists of a tor0idal coil, wound with Litz wire, and variable capacitor, with a by-pass switch (design again thanks to Dave).
Sources for the many of the components are listed at the end.
|Type||OD "||ID "||Turns||Inductance||Notes|
|Detector||Spider web||4.875||2.375||49||280||Tap at 26T from inside|
|Wave trap 1||Toroid||42||230||3 T take-off coil wound over main coil|
|Wave Trap 2||Spider web||4.5||1.875||52||250|
Wire for all coils is 175/46 Litz. The Spider Web coils are wound on the 5.25" form and the toroid is an Amidon FT-114A-61. Inductance is in microHenrys.
Well, how did it work? Overall I am quite pleased with the result. Over a period of a couple of weeks I was able to log 135 broadcast band stations with the 30m random wire antenna.
The radio could be improved by increasing selectivity at the upper end of the dial, perhaps using Dave's ContraCoil approach. Even with the wave traps, our local powerhouse station tends to dominate anywhere near its 700KHz frequency, but outside of that I was able to receive stations all across the band, depending upon conditions. The farthest stations were in Texas and Cuba, approximately 1020 and 1370 miles, respectively. (See map link below).
Sources and Links
Parts & Supplies
Calculators and General Information
Return to top