# Crystal sets - How do they work? :) (with real diodes)

Discussion in 'Radio and Wireless' started by Nanren888, Jul 24, 2018.

1. ### Nanren888

67
10
Nov 8, 2015
Looking at requirements for energy harvesting recently, I realised that I had not appreciated fully the magic in the traditional crystal set. Perhaps someone can clarify or confirm or deny for me.
.
In the simplest explanation, a perfect diode, with zero forward voltage, detects, rectifies the RF signal and delivers those RF peaks representing the amplitude to the capacitor and on to provide AM demodulation.
.
Like many here, I'm sure, I've build these using germanium diodes.
Assuming about a 0.3 volt forward drop, I'm looking for where that sort of voltage comes from.
.
My antennas were always not so very many turns on an air or ferrite core. The capacitors were usually traditional moving metal, air dielectric. I'm assuming the Q of these DIY tank circuits were never particularly high.
.
Anyone done the maths already? Can anyone confirm?
The magic to get over 0.3 volts was enough turns plus enough Q factor?
Or am I missing something?

2. ### Bluejets

3,298
652
Oct 5, 2014
Voltage is not a problem.
Antenna extracts whatever power it can via the antenna, perhaps only nanowatts.

3. ### duke37

5,173
700
Jan 9, 2011
The 0.3V for a germanium diode to conduct is only a rough approximation. I am sure someone can come up with the logarithmic equation for a more accurate result.
Capacitors should have very low loss and the circuit Q will depend on the coil. Do not get mixed up between circuit Q and loaded Q. To get a good match between aerial and earphone you have to set the tuned Q to an optimum. A low working Q can give more volume but little selectivity.

Many years ago we had a wartime battery radio, MW only and us lads wanted to listen to Dick Barton, special agent on LW. My brother made a crystal set and with a 50 yard long aerial it managed to drive four earphones. The diode was ex military, possibly from a RADAR set.

When I became rich, I bought a beaten up old Heathkit OS1 scope. The wire was still up and the scope showed 1V at the house even though we were 50 miles from Droitwich.

4. ### kellys_eye

4,286
1,145
Jun 25, 2010
If you want to do energy harvesting then purchase a house that sits under HV transmission lines.......

As per post #2, the levels of energy are not of usable quantities.

5. ### Nanren888

67
10
Nov 8, 2015
I saw voltage as the problem as if the voltage does not rise on a carrier cycle sufficiently to forward bias the diode we don't get anything out.
Granted it was a also a crystal earpiece, so the impedance is high. The "match" was as close to ooen, unloaded as we could get, I guess. Good point.
Logarithmic conduction? Also a good point. You're suggesting it's running way down in the usually neglected almost flat part of the curve near zero. Could be.
Depends on the coil? Yes, my point is they were usually by, unskilled, diy-ers so while the loss was low, wouldn't anything much in Q.
Not really a magnetic antenna expert. I sort of imagined that for any fixed, small em field, and low Q tank, even open circuit, unloaded, I'd see a maximum open circuit voltage determined by turns. No way was I seeing that getting near 0.3 volt.
Many thanks for the reminder of the log relationship: will think on that more.
Welcome any other thoughts.

6. ### Nanren888

67
10
Nov 8, 2015
Harvesting, no. Thanks though.
It was just looking at an analysis of harvesting needing to get above the diode drops for the voltage multiplier too work that made me think about the venerable crystal set.
Comment appreciated.

7. ### Nanren888

67
10
Nov 8, 2015
Actually, years ago, I saw a remote comms shed under a line that was powered from a line from the otherwise unearthed metal roof. Cheaper than a transformer. Belonged to the power authority.
Stories? Yes I'm not so young.

8. ### Bluejets

3,298
652
Oct 5, 2014
Think maybe they snuck in an u/g feeder when you were having your nanny nap.

9. ### hevans1944Hop - AC8NS

4,059
1,920
Jun 21, 2012
A diode, of any kind, conducts some current when forward biased and virtually zero current when reverse biased. That is all that is necessary for rectification of microvolt RF signals from the antenna.

10. ### Nanren888

67
10
Nov 8, 2015
@Bluejets - thanks underground cable? - No, it was legit powered from induced voltage for some time. The line was 33 or 66KV. A Comms, scada shed with nothing else round, but hill country, no cable.
.
@duke_37, hevans1944, small signal behaviour: yes. Have a new appreciation of the workings of real diodes. Actually, it's great that you can have such a simple model working so well, for so long, without needing to fall back to more accurate University engineering-physics log laws & such.
Thanks again.