— David Wagner 2007/09/13 00:50

First, consider how a radio wave at some frequency (say 1MHz) will typically induce a tiny alternate current at some moderate voltage (perhaps 10-40V peak-to-peak) in an antenna. Now think of a typical resonant LC circuit driven by the antenna current. This 'tank' circuit stores more than one cycle of an exciting wave before it is charged enough to be 'full' and begin to overflow energy through the diode. The resonant circuit is oscillating at much lower voltage (<1 V), yet it has more power, and therefor much higher current, than the signal from the antenna driving it. Ideally, the carrier of the AM signal should produce a waveform peaking just at the forward voltage of the diode, so only the modulated portion overflows into the headphones. The detector should effectively skim the signal off of the carrier.

In addition, a perfect detector for a crystal radio circuit will automatically compensate for the strength of the station by having a higher forward voltage drop at higher current. Thus the carrier of a strong station is held back by a higher drop caused by the stronger (higher current) audio portion getting through the detector, while a weaker station pushing less audio current will have a lower threshold to overcome.

The capacitor is responsible for the voltage drop; that is what capacitors do. If the capacitor is too small, the voltage will get too high, and in addition to letting through the carrier of the tuned signal, other signals may overflow the tank and make the response broad as a barn. A Schottky rectifier with a very low forward voltage may require a larger capacitor and a smaller inductor to work properly as a crystal radio. To use a silicon or other diode with a higher forward voltage at the same frequency, the capacitor should be smaller, and the inductor larger.

Using LEDs as detectors allow for experimentation with varying the voltage drop. Simply shine a light (preferably another LED of the same type) at the detector to lower the effective voltage drop. With enough light, the detector might even have a negative forward voltage. A solar cell or phototransitor may provide interesting results as well.

Keep in mind, a standard broadcast AM station in the US uses three frequency ranges, around the assigned frequency. The carrier is usually within 20Hz of the assigned frequency, while the sidebands encoding the audio signal (twice) spread out over 30 kHz¹⁾ with a gap around the carrier separating the two sidebands. In the US, frequencies are allocated every 10 kHZ and the audio bandwidth is limited to 10 kHz²⁾ so the sidebands cover at most 20 kHz centered around the carrier. The energy of most audio signals are concentrated in the lower frequencies, so limiting the sideband reception to within 5 kHz of the carrier will help differentiate distant stations from nearer signals.

Although only one sideband is needed to decode the audio content, a crystal radio set should be able to tap as much of the broadcast signal power as possible, though it also needs to be able to minimize the interference when the sidebands from different stations overlap.

For parallel RCL circuits³⁾ ,

• The higher the Q, the narrower the bandwidth.
• The smaller the inductor, the larger the capacitor, and the larger the resistance, the narrower the bandwidth.

Broadcast stations typically modulate the signal an average of around 60%.⁴⁾

First, assume the content is speech, and maximize for intelligibility.

The simplest tuner has a center frequency at the carrier and a bandwidth of 8-10 kHz to capture the frequencies below 4 kHz important to understanding human speech. This corresponds to a resonant circuit Q of around 100 (at 1000 kHz), the traditional target for a typical hobby crystal radio set. The challenge is to maintain intelligibility while increasing the ability of the set to pick up and discriminate weaker signals than is typical.

If a crystal radio set uses the carrier and unneeded lower- and higher- frequency audio to forward bias the diode, the intelligible portion of the signal can pass to the headphones relatively undistorted. Half of the energy in a typical speech signal is below 500 Hz though this range contributes less than 5% to intelligibility. Although frequencies above 4 kHz contribute about 15% to intelligibility, they have only about 5% of the energy,⁵⁾ and attempting to capture this range is much more likely to be subject to interference from adjacent radio stations.

In addition, it makes sense to be able to tune only one sideband for the audio content, since two sidebands are twice as likely to be subject to interference as one. However, being able to switch in and adjust two sideband tuners would allow for a wide range of tuning possibilities. This leads to the conclusion that at least one resonant circuit in a crystal set should have a Q sufficiently low to capture the sideband encoding 500 Hz to 4-5 kHz, a bandwidth of 3.5-4.5 kHz (call it 4 kHz) requiring a Q of about 250 (at 1000 kHz). But the margin of error in these circuits makes little distinction between this range and a range including the lower frequency audio; a Q of 225 will expand the bandwidth enough to include 125 Hz and capture this additional energy.

Note these sideband tuner circuits would need to be tuned to a frequency offset from the carrier by about 2.5 kHz and thus will not tune the carrier itself. Another tuned circuit will be necessary to tune the carrier. This carrier tuner benefits from a very high Q to provide a clean demodulation signal.

The problem then becomes how to construct a set of passive filters with these characteristics. One approach is to weakly couple separate resonant circuits.

               . 
              /: 
         +--||--+
         | /   :|...................
         |    / |       /         /
         +--||--+  +--||--+  +--||--+  Big Variables
         | /    |  | /    |  | /    |
\|/      |      |  |      |  |      |
 |       |    / |  |    / |  |    / |
 |       +--||--+  +--||--+  +--||--+  Trimmers
 |       | /    |  | /    |  | /    |
 +-UU-+  +--UU--+  +-UUUU-+  +-UUUU-+
   ===|=======       ====      ====    Loose Coupling?
      |  +--UU--+  +--UU--+  +--UU--+
      |  |      |  |      |  |      |
      |  |      +--|------+--|------+---
      |  |         |         |
      |  +---------+---------+----------
      |
     ---
      -

               . 
              /: 
         +--||--+
         | /   :|...................
         |    / |       /         /
         +--||--+  +--||--+  +--||--+  Big Variables
         | /    |  | /    |  | /    |
\|/      |      |  |      |  |      |
 |       |    / |  |    / |  |    / |
 |       +--||--+  +--||--+  +--||--+  Trimmers
 |       | /    |  | /    |  | /    |
 +-UU-+  +--UU--+  +-UUUU-+  +-UUUU-+
   ===|============================    Loose Coupling?
      |                      +--UU------
      |                      |
      |                      +----------
      |
     ---
      -

The Hobbydyne appears to take the opposite approach to putting more RF energy to work. Because of its very large RF choke inductor and tiny variable or differential capacitor, the Hobbydyne looks like a very wideband tuner. It looks like it can use the broadcast energy remaining after the antenna tuner (and perhaps some traps for local blowtorches) to provide the current necessary to keep the diode at its forward voltage. If the antenna tuner has a decent Q, this is how the lower frequency audio is captured.

This is an elegant way to put to LC resonant circuits in parallel without them interfering too much with each other. The component values of each primary resonator are too small (in parallel) or too large (in series) to influence the frequency of the other much, and the two additional resonant frequencies are either too high or too low to interfere with the broadcast band tuned frequency or the audio band.

 \|/
  |         Tuggle Antenna Tuner
  |               ........
  |              /      /
  |         +--||--+--||------+
  |         | /    | /        |
  |         |      |         ---
  +--UU--+--+-UUUU-+          -
    = =          
  +-UUUU-+  +-UUUU-+
  |      |  | Tank |
  |    / |  |    / |
  +--||--+  +--||--+--------------+----+
    /       | /                   |    o
  Wave      |      +--VV--+       |     )
  Trap      |      |      |       |    o
            +-->|--+--||--+---UUUU+UU--+
                    Benny     = = = =

This is similar to this radio.

 \|/
  |         
  |         Tuggle Antenna Tuner
  |         +----------------+
  |         |                |
  |         |              / |
  |         |            ||--+
  +--UU--+--+--UU--------| 2x365p
    = =       = =      / ||--+
  +-UUUU-+  +-UUUU-+         |
  |      |  | Tank |        ---
  |    / |  |    / |         -
  +--||--+  +--||--+--||-------------+-->|--+----+
    /       | /       ||             |      |    |
  Wave      |         ||--||--+      |      |    |
  Trap      |                 |      |      |    |
            |     Hobbydyne   |  27m |      |    o
            |                 +-UUUU-+      |     )
            |                 | ====        |    o
            |                 |        100p |    |
            |                 +---------||--+    |
            |                 |             |    |
            |                 |             |    |
            +-----------------+--||--+--UUUU+UU--+
                              |      |  =======
                              +--VV--+
                               Benny

These are like this radio.

This next shows how the Hobbydyne and the Benny work the DC bias on either end of the diode. You can also see how the Hobbydyne forms another resonant circuit in parallel with the tank.

 \|/
  |
  | Wave Trap
  |      /      Tuggle Antenna Tuner
  | +--||-+               ........
  | | /   |              /      /
  | |     |  +-UUUUU---||--+--||------+
  | +-UUU-+  | = = =  /    | /        |
  |   = =    |             |         ---
  +----U-----+--UU---------+          -
                ::       /
             +--UU-----||--+-------------------+------o
             |        /    |                   |       )
             +-UUUUU-------+--||--+--||--+-UUUU+UU----o
             | = = =       |      |      | ======= 
             |       +-UUU-+      |      |
  Simplified |       | ===        |      |
  Hobbydyne  |    /  |            |      v  Benny  
             +--||---+----->|-----+-----VVV
               /                    

Here is an interesting possibility…

Standard transformer tapping…

Should have choke and benny in series?

 \|/
  |
  | Wave Trap
  |      /      Tuggle Antenna Tuner
  | +--||-+               ........
  | | /   |              /      /
  | |     |  +-UUUUU---||--+--||------+
  | +-UUU-+  | = = =  /    | /        |
  |   = =    |             |         ---
  +----U-----+--UU---------+          -
                ::       /
             +--UU-----||--+
             |        /    |
             |             |   5 :    50 :  1 turn
             +-UUUUU-------+  25 :  2500 :  1 mH
             | = = =       | 400 : 40000 : 16 Ohm
             |    /     /  |
             |  ||----||---+--||----+
             +--|    /              |
   Hobbydyne  / ||--+-->|------+    |
                    |          |    |
                    |    VVV---+    |
                    |     |    |    |
                    |     +-||-+----+    +-------o
                    |     |         |    |        )
                    |     +UUUUUUUUU+UUUU+UU-----o
                    |      ==============|==
                    +--------------------+

Shorter DC path…

Should have choke and benny in series?

 \|/
  |
  | Wave Trap
  |      /      Tuggle Antenna Tuner
  | +--||-+               ........
  | | /   |              /      /
  | |     |  +-UUUUU---||--+--||------+
  | +-UUU-+  | = = =  /    | /        |
  |   = =    |             |         ---
  +----U-----+--UU---------+          -
                ::       /
             +--UU-----||--+
             |        /    |
             |             |   5 :    50 :  1 turn
             +-UUUUU-------+  25 :  2500 :  1 mH
             | = = =       | 400 : 40000 : 16 Ohm
             |    /     /  |
             |  ||----||---+--||--+-----+
             +--|    /            |     |
   Hobbydyne  / ||--+-->|--+------+     |
                    |      |            |
                    |     VVV--+---||---+-------o
                    |          |        |        )
                    +------UUUU+UUUUUUUU+UU-----o
                           ================

Maybe use the audio transformer as the RF choke? This also eliminates many of the extra resonant circuits created by the RF bypass cap and the Hobbydyne.

Should be connected like the previous circuit? No, this should work, but the resistor needs to be a much smaller value… Perhaps a 1kΩ pot…

 \|/
  |
  | Wave Trap
  |      /      Tuggle Antenna Tuner
  | +--||-+               ........
  | | /   |              /      /
  | |     |  +-UUUUU---||--+--||------+
  | +-UUU-+  | = = =  /    | /        |
  |   = =    |             |         ---
  +----U-----+--UU---------+          -
                ::       /
             +--UU-----||--+
             |        /    |
             |             |        50 :  1 turn
             +-UUUUU-------+      2500 :  1 mH
             | = = =       |     40000 : 16 Ohm
             |    /     /  |
             |  ||----||---+--||--+-----VVV
             +--|    /            |      |
   Hobbydyne  / ||--+----->|------+--||--+------o
                    |                    |       )
                    +-----------UUUUUUUUU+UU----o
                                ============

FT50-61 Winding, A_L=69, 72 turns 32GA will fit in one layer

• 220/180 uH: 56/51 turns (Try contra-wound instead)
• 4.3/6.8 uH: 8/10
• 250 uH: 60 turns

FT37-61 Winding, A_L=55, 42 turns 32GA will fit in one layer

• 4.3/6.8/250: 9/11/67 turns (two layers)
• 4.3/6.8/97 uH: 9/11/42 turns (This should still work for the wavetrap)

T37-2 Winding, 250KHz - 10MHz, A_L=4.0, 50 turns 32GA will fit in one layer

• 4.3/6.8 uH: 33/41 turns

T50-2 Winding, 250KHz - 10MHz, A_L=4.9, 72 turns 32GA will fit in one layer

• 4.3/6.8 uH: 30/37 turns

T130-2 Winding, 250KHz - 10MHz, A_L=11.0,

• 250 uH: 151 turns

T200-2 Winding, 250KHz - 10MHz, A_L=12.0,

• 250 uH: 144 turns

• Why is the diode in a crystal radio always biased the same way?
• Could the Hobbydyne inductor be used as the primary in the audio transformer?
• Why is there no push-pull crystal radio using two detectors oppositely biased? Hey! I found one! And another about halfway down the page, and one more.

           ^        \  /  
           |      +--[]--+
        _  |      |  S1  |
        /  |      |      |
  +---||---+-->|--+--VV--+
  |  /C1   |      |  R1  |
  |        |      |      |
  +--UUUU--+      +--||--+
  |  L1    |         C2  |
  |        |             |
  |        +--|<---------+
  |                      |
  +-----------VV---------+
  |           10M         
 ---                      
  -                       

Keep in mind much of the power of the signal is in the carrier, and the typical crystal radio does not appear to use any of this power. (or half of the rest of the signal, for that matter.)

Supercap idea already done, sort of…