An Improved Simple Crystal Set

( Updated 8 May 2003 )

I've been working for a while to come up with a simple crystal set design that offers good performance, simple operation, and is made from readily available materials.  All of the common hookups I tried worked OK during the day, but were often completely "crushed" by short-wave broadcast stations at night.  (This can be a big problem here in the North-Eastern United States.)  I believe I have a solution.  I built a couple of reasonably successful sets like this using ferrite rods, but telling anybody how to reproduce the coils is a real problem.   I'd like some of you to try this set, and tell me what you think.

Desired Characteristics

- Easily reproducible by unsophisticated builders without tweaking
- Uses readily available components
- Good performance - "Gets DX"
- Easy operation
- Works well in the face of strong short-wave signals.

The Problem

Almost all single-tuned crystal sets have limited rejection on the high side of the frequency to which they are tuned.  Furthermore, common crystal set antennas are often resonate (approximate 1/4 wavelength) in the 5-8 MHz region that contains several popular SW broadcast bands.  If you're not in a metropolitan area, it's easy for the short-wave signals to be a lot stronger than the BCB stuff you're trying to hear.

Consider the following screen photo:          ( Here's  a block diagram of the TEST SETUP.)

Frequency response of a typical single-tuned crystal set using capacitive coupling (Tuggle circuit) tuned to 1000KHz.  Horizontal scale is 1MHz per division, vertical is 10 dB per division.  The peak at the left is the zero-frequency artifact of the spectrum analyzer.  Set under test is tuned to 1MHz.  The display shows the RF response of the tank circuit, which is center tapped and loaded by a 1N34 working into a 10K resistor.  As you can see, the rejection in the crucial 5-10 MHz region is only about 30 dB.  Here in New Jersey, strong short-wave signals crash right through.  A classic double tuned circuit will do much better, but we're trying to keep things simple.

My solution

Schematically, there's nothing to it:  A 365pF cap, a 1N34,  a 2K-ohm (DC) headset, and the transformer.

This is actually a double tuned circuit.  The primary is self resonate at about 400 Khz, tightly coupled to the secondary, and tuned only by the antenna capacity.  The rejection of the primary and secondary add to give rejection of greater than 50 dB above 5 MHz.  The set uses only one variable capacitor, and has no coupling adjustments or tap switches.
Because there is both inductive and capacitive coupling between the primary and secondary winding, phasing is important.  Improper phasing will result in this sort of response. (In this case, right in the 40 -meter band!) If you hear shortwave, reverse the connections to the primary, or flip the coil over.
The above data were taken with the FET probe attached to the top of the secondary tank.  I was assuming that the detector, on the center tap, was seeing a similar signal.  (Years, ago my football coach warned what might happen if one were ASS-U-ME things.)  The first night I tested the set, short-wave propagation must have been poor.  Things seemed to work well, and I published this page.  The following night, I had 40 and 49 meter signals all over the place. The frequency response, measured right at the detector, looked like this. (Left)
The simple expedient of applying a Faraday shield to the primary winding achieves reasonable shortwave rejection.  The bump at 8MHz could be a problem if in wanders into one of the adjacent SW broadcast bands.  The "grass" around the 6 MHz line is actual short-waves signals leaking into my test setup without the benefit of an antenna. (About 23:00 local time.)

Just as radio stations are defined by their antennas, radio receivers are defined by their coils.  This one comes from Home Depot.  The form is a 4" styrene pipe coupling.  The wire started out as two-conductor #20 "thermostat wire."  It's untinned solid copper.  they make you buy 500 feet, so there's enough for 5 or 6 radios.  The secondary is 55 turns tapped near the middle.  (25 turns in this case.)
The primary is 75 turns of the same wire scramble wound around a 12-ounce beverage can, and tied to keep it neat.

The primary is wrapped in aluminum foil.  The first couple of inches of foil are wrapped in electrical tape.
The end of the foil wrapping overlapps the beginning by about an inch.  The tape prevents electrical contact between the two ends of the foil shield.  If this insulation were not there, the foil would constitute a shorted turn in the coil, and destroy it's inductive characteristics. 

This Faraday shield keeps the electrostatic field inside, but allows the magnetic field to couple to the secondary.

The primary sits in the bottom (ground end) of the secondary.  The objective is to maximize inductive coupling while limiting capacitive coupling.  (I'll ad an updated picture, with the shielded primary, soon.)

I can't take credit for this design.  This is how input coils for early tube radios were constructed when energy transfer was still important.
As I mentioned above, I started out making these transformers on ferrite rods.  Details will vary with the rod material, wire gage, etc.  The one at the right is on a 3/8" rod salvaged from a transistor radio.  The white plastic is hobby shop material, and cuts and glues nicely.  The primary, in the bobbin, should be about 120% of the secondary turns count.  The ferrite core provides tighter coupling than the air-core transformer, and the small size and physical arrangement of the windings control capacitive coupling, so the Faraday shield is not needed.

This circuit can make small sets worthwhile.