Jim

This is the schematic diagram of a basic crystal set.

The tuned circuits in a crystal radio are composed of a coil (inductor) and a capacitor, usually variable, to allow tuning the circuit to the desired frequency. A crystal set can be made with a single tuned circuit (the detector coil), or with an added tuned circuit (the antenna coil) as shown in the above diagram. A dual tuned circuit is usually required for best selectivity. It is difficult to couple an antenna to a single detector coil without ruining the selectivity of the set.

The dual tuned circuit works by inductively coupling the signal from the antenna coil into the detector coil with minimal loading. The spacing between the two coils can be adjusted for proper coupling, or one coil can be rotated with respect to the other to change the coupling. In practice, the detector tuning capacitor, the antenna tuning capacitors and the amount of coupling between the two coils are all juggled until the strongest signal is obtained.

For most crystal sets a simple 365pF variable capacitor will work fine. They are available on eBay or from the Xtal Set Society. The single section capacitor is usually used for the detector, while a dual section capacitor is usually used with its sections in parallel for the antenna circuit.

The diode rectifies the received RF signal and converts it to a direct current. It makes no difference which way the diode is orientated. The 250 pF capacitor bypasses to any residual RF to ground. The .1 mF capacitor and the 5K ohm resistor in parallel serve to lower the distortion that would otherwise be present on a very strong signal.

The Q of a coil is the ratio of reactance to the ac resistance, or in effect the efficiency of the coil. Coil Q is a very important consideration when designing a crystal set. The selectivity of your completed crystal set (its ability to separate close adjacent stations) will be determined primarily by the Q of the detector coil, and to a lesser extent by the Q of the antenna coil. Additionally, a higher Q coil means higher sensitivity and stronger signals. For best performance a crystal set should have a detector coil with a minimum Q of 500 or more. The Q of the coil can be measured by a laboratory device called a Q meter.

Litz wire is best to wind high Q coils with. Because of skin effect¹ at radio frequencies the signal flows around the outside perimeter of a wire and only a thin outer layer of the wire is used. This reduces the effective size of the wire and increases the losses. Litz wire avoids the losses because it is composed a hundred or more strands of extremely fine wire, each insulated from the other. The gauge of the wire is chosen for the frequency band in use so that the entire wire can be effective. The gauge for the broadcast band is #46. It is available in many different strands for different size coils. The most popular sizes are 100, 170 and 660 strand. In my opinion, there is no excuse for not using Litz wire because of the superior performance it provides. It is available from many sources including eBay

Inductors come in many varieties. You can choose from cylindrical coils, spider coils, honeycomb coils, ferrite rods coils, and ferrite toroid coils. The old-time classic crystal set coil used a cylinder coil with varnish or silk insulated copper wire wound on it. The coil was typically about two inches in diameter and four inches long. The problem with any cylinder coil is that the closely adjacent turns have unwanted capacity (called distributed capacity) between the turns and it degrades the performance of the coil and reduces its effective Q. Distributed capacity can be lowered by slightly spacing adjacent turns of the wire, perhaps by 20 to 40 percent of the wire diameter. However, to h

he inductance of a coil is lowered as its length is increased, so the longer space-would coil will require some additional turns of wire. Adding more turns of wire adds more copper loss.

The Spider coil, the honeycomb, and other exotic variations were all developed to overcome the distributed capacity problems of cylindrical coils. They provide high Q because of their lower distributed capacity.

A Spider coil is very practical and easy to build. One that is about 4 inches outside diameter wound with 100 strand #46 Litz will have a unloaded Q of about 400, and an 8 inch one wound with 660 strand #46 Litz wire will have an unloaded Q as high as 1000. The down side is that the radio will be quite large, and the Spider coil will have to be located several inches from any metal objects to avoid lowering its Q.

Another choice is the ferrite core coil. A ferrite core in a coil has the unique ability to greatly increase its inductance. For a given inductance, this means that the coil will be much smaller with much less wire required. Since there is less wire the copper loss is less and the Q will be much higher. Ferrite toroids and ferrite rods are both available. The toroid is unique in that its field is confined to the ferrite core. Ferrite toroids are hard to wind, and hard to couple to each other because there is little external field. The ferrite rod is more practical for the home builder. Two ferrite rod coils can easily be inductively coupled the same as cylinder coils.

As a comparison, a typical cylinder coil for the broadcast band wound with #22 copper wire, 2 to 3 inches in diameter and 3-4 inches long, requires 50-100 feet of wire. Its Q will be about 200. The same coil wound on a ½ inch by 2 or 3 inch ferrite rod needs only about 9 feet of 100 strand #46 Litz wire and, depending on how it is wound and the quality of the particular ferrite core, will have a Q of between 500 and 800. A set made with the ferrite rod coil will be more than twice as selective as the cylinder coil set and the signals will be noticeably louder.

Ferrite rods are manufactured with different compositions for different frequencies. The proper ferrite for the broadcast band is #61. The rods I use are ½ inch in diameter and are available in 2, 3 and 4 inch length. For frequencies below the broadcast band you can use #33 ferrite rods and #44 Litz.

The Q we are referring to in the above paragraphs is the “unloaded” Q, or the Q that is measured when the coil and its associated capacitor have nothing else connected to them.

The performance of any coil, exotic, cylinder or ferrite cored can be directly compared by measuring its unloaded Q.

Obviously, we must connect something to our coils to make a crystal set. Anything we connect will reduce the Q somewhat. The object is to couple the antenna and the earphones with the minimum loss of Q. In the case of the single coil set, connecting the antenna to the coil itself tends to ruin its Q. The double tuned circuit overcomes this to some degree by inductively coupling the antenna tuned circuit to the detector tuned circuit. This coupling can be adjusted by spacing the coils or rotating one in relation to the other. There will be an optimum coupling point where the selectivity and sensitivity are best. There are also other ways to avoid the loss of Q and they will be described in a later section

A ferrite rod coil of sufficient inductance to resonate just below the broadcast band with your particular antenna can be used to inject the signal into the detector coil instead of the coil-capacitor combination. It will not work as well as if it were tuned, but it still gives very good performance. The usual inductance with most antennas will be from 300 to 400 uH. This would not be practical for a very short antenna or for a when the best possible reception is required.

The field around a cylinder coil and most of the other exotic types is quite large. Any metallic objects within several inches can seriously degrade the coil’s Q. It will “see” your hand and its Q will start to degrade at a distance of 5 inches or more. By contrast, a coil with a ferrite rod core has a more compact field and your hand has much less effect on its Q. It can also tolerate nearby metal objects that would lower the Q of the other types of coils.

In a set with separate antenna and detector coils the typical separation between two properly coupled 3 or 4 inch cylindrical coils could be from 4 to 12 inches which makes the crystal set quite large. With ferrite the optimum coupling occurs at an inch or two. As you can see, ferrite can be used to make a extremely high performance crystal set that is quite small

How the crystal detector is connected to the coil and the impedance of the earphones will also affect the Q reduction. In the basic crystal set the diode is connected directly to the coil. A better way is the Hobbydyne circuit which adds a 15 pF variable capacitor between the coil and the diode. This capacitor can be variable if desired, and can be adjusted for more or less selectivity. The Hobbydyne circuit shown below requires a series capacitor and a small rf choke of 10 to 27 mH to provide a dc return for the diode. The classic crystal set diode is the 1N34. However, with high Q coils a Schottky diode will usually work better.

For maximum performance the earphones must be matched to the impedance of the radio. Conventional 2000 ohm earphones work pretty well, but using a transformer can boost performance. WW2 sound powered phones are excellent, but their low 600 to 1200 ohm impedance makes a transformer mandatory. A low loss transformer with a primary impedance of 50 to 100K ohms or more to a secondary that matches your earphones will work well. You can find suitable transformers for sale at the various crystal web pages or from eBay. It should be noted that most "earbuds" and other low impedance earphones used by cellular phones and iPads are not suitable unless amplification is used in the set.

Diagram of the Hobbydyne Circuit

Diagram of a set using a FET

Adding a single FET transistor to a crystal radio can greatly improve the selectivity because it reduces the loading on the tuned circuit. The FET is configured as a source follower and has a very high input impedance. The output of the crystal diode is connected directly to the gate of the FET. In this circuit the orientation of the diode must be as shown with the cathode towards the FET. The audio output from the FET is taken from the Source lead. The FET eliminates the loading on the coil due to the diode. The Source of the FET is low impedance, and can drive most earphones directly, including sound powered phones, without the requirement for a matching transformer. It does not amplify the volume of the signal, it only matches the high impedance input from the crystal detector to the lower impedance earphones. It also works fine with low impedance earphones and ear plugs.

A Deluxe Set with a FET and a LM386 Audio Amplifier

For a superior set consider adding a LM386 amplifier which can drive any earphones or 8 ohm speaker. The LM386 amplifier shown below has a gain pot to adjust the gain anywhere from 30 to 100. The data sheet for the LM386 can be downloaded from any of the larger parts suppliers. It only draws about 5 mA quiescent current, so the battery will last a long time. I find that a gain of 30 is adequate for earphones. When using a speaker some additional gain is nice.

Because the FET and LM386 amplifier is used the gain is very high and it is possible to couple the antenna very lightly to the detector coil for superior selectivity and still have excellent sensitivity. Usually a one to three turn link around the detector coil will be sufficient, or, as described above, with the antenna coil resonant below the broadcast loosely coupled to the detector coil.

I built this set in early April, 2015 and I have posted a picture of it on the main Crystal Set page. A printed circuit board layout is available.

You can email me with questions or comments. My email address is at the bottom of the main Crystal Set page. It is a .jpg address to avoid spam, so you will have to type it into your email program.

Visit Jim's Cushman Scooter Web Site www.hobbytech.com