My Crystal Radio Set

 I've made a few Crystal Radio sets over the years. If you ask me, I'll be happy to go on forever about how they work, but there are a ton of online articles that are very good and I'd just be redoing others good work. One of the best articles I've read is: https://www.jumpjet.info/Pioneering-Wireless/eBooks/useful/2008a.pdf

The best resource that I've found for tutorials online is the Crystal Radio Set Society http://www.midnightscience.com/ They have designs, and articles by others about how to make various sets. The formulas and explanations can be found here: http://www.midnightscience.com/formulas-calculators.html#formulas1

I think one of the most useful tools is the Professor Coyle calculator pages (different website) https://crystalradio.net/professorcoyle/professorcoylecyl.shtml

You can use these pages along with code your own if you wish to design your own sets.  Both of the websites have interactive web pages that allow you to calculate the desired frequency from the inductor (L) and capacitor (C) pairs. In addition, there are various formulas for how to make your own inductors. You can use a host of different forms for making your inductors. I have not yet tried making any of the basket weave coil forms. I have made standard forms and a spider coil form. 

The spider coil form really cuts down on the inter-winding capacitance and hence self resonance. I always try to make stuff with easy to find components for the most part. I do occasionally splurge for the expensive components. In follow up articles I'll talk about measuring the properties. 

The first one I made was the Mystery Crystal Radio set:

I don't recall where I got this from. This set worked amazingly well. The two sets of coils were wound wire to wire. That is to say that turn of primary, and turn of secondary and so on until complete. I used a high impedance headphone pieces (~2000 ohms) and some stations you could hear without putting the headphone to your ear. The selectivity was better than I expected. At the time I spent a lot of time in Boston and I could tune out powerful stations like WBZ and hear others. One thing I could never figure out was how did this work without a direct ground on the aerial. I used some 22 gauge enameled wire on a 2 1/2 diameter PVC pipe. I'm always a fan of using PVC pipe as a coil form, it is relatively inexpensive and easy to get. Of course paper rolls are too. E.G. the old school Oatmeal container. 

The next one I made  I used PVC and 22 gauge wire again and housed it in a cigar box. I did not decouple the antenna other than put it through a capacitor. (Kind of a mistake, the antenna is always short and should have needed some inductance to balance it out. so that never worked. ) Overall the set worked ok but not great. By and large it looked nice but was super fundamental. 

The next one I made, I disassembled the above set and replaced the enameled wire with Litz wire. At radio frequencies (RF) current has a tendency to flow on the surface of the conductor and the RF resistance can vary as a function of frequency. This is called the skin effect. The skin effect is kind of a killer. What I have not said to date is that the amount of energy at the antenna is like micro-watts. Overall the currents and voltages are very small and any amount of resistance can make things worse. There is also this thing called Q. Since we are dealing with a resonant circuit, the width of the resonant curve is measured by Q the quality factor. (Wikipedia has a great article on Q and skin effect: https://en.wikipedia.org/wiki/Q_factor) Depending on the type of circuit you use it is formulated differently. I tend to build what is called parallel resonant circuits and therefore the greater the R the lower the Q. You want relatively high Q. Hence you want really low AC resistance. Litz wire provides this. It's kind of expensive. The most expensive one I have purchased was 660/40. This means there are 660 individually insulated wires in a bundle and are 40 gauge. This makes for a huge equivalent surface area and lowers the AC resistance. 

Well, this worked pretty well overall but the local station, 2km away, swamped a lot of the weaker stations. 

I ended up with a configuration like this: 

the three inductors are loosely coupled. the tuned circuit at the far left I tune to fundamentally short out the local station I do not wish to receive. This is sometimes called a wave trap or Purcell filter. The center is just a transformer winding. and the circuit at the right is tuned for the desired station to receive. The speaker isn't really a speaker - it is a 4000 ohm headset. 

Living in NY I found that I have a difficult time differentiating between stations.