The AM Forum

Hola!

Need some advice...  I'm building up a loose-coupler crystal set with an antenna matching input.  I'd like to see how many distant Broadcash stations I can tune in over the winter.

Questions:   I will be tuning the complete BC band from 540 - 1700 KHz.   So I need a receiving antenna that can be tuned from the shack, no coax feed.

I have two candidate antenna systems... One, a 190' grounded tower with modest radial system -  and a 100' grounded tower that has an extensive 160M ground system, about 130 radials.  (I don't want to mess with Beverages.)

What would be the best config?...  I can run a single wire to the top of either tower (wire about 300' long) which starts at the shack and is tuned there... and electrically ties in to the top of either tower.  The electrical path starts at the bottom of the tower at zero ohms, goes vertically to the top of the tower and then the wire slopes down at a 60 degree angle to the shack.

I presently have a 75M quad at 190' quad fed with hardline that receives the entire BC band very well despite the swr mismatch and was hoping to do better with the wire/tower matched combo.


Another question: Should I instead make the system ALL vertically polarized by running an openwire match feeding the tower that can be tuned at the shack?  (centerfed vertical dipole)  Considering the better performance of verticals compared to horiz polarization below 75M, since all BC stations are vertically polarized, etc.   Or will the BC station polarization "tumble" and be random when DXing in the BC band like the higher freqs do?

Or would an old fashioned horizontal Hiram "T" top capacitance hat stretched between two towers fed with a single center fed wire work best?



Thanks.

T

   Tom,

   Sounds like a lot of fun.....

I have a couple of things to ponder.

If you can use an antenna with directivity to null a strong station on top of a weak one, then you gain an advantage.

I am experimenting with a W6LVP loop inside my garage. On the BC band at night, I see an S9+ signal every 10 Khz across the BC band. That loop has a deep null too when positioned right.

https://www.w6lvp.com/

Just thinking that a mile of wire up 200' will give you too many stations stacked on top of each other. Identifying one from the other mostly impossible. Might be much like CB Channel 19 when the skip is rolling in.

My thoughts,,,

Jim
Wd5JKO

OK Jim -

Yes, the signal congestion will be intense, especially at night, no doubt.  FB on the loops.

I was thinking I needed a high efficiency antenna because of the crystal set needing all the help it can get. But my experiments are showing that if the set is built right, there is usually plenty of signal to work with. Many of the set designs on the web are duds. There are a handful of good ones - dual stages - that have an antenna tuner loosely coupled to the  diode resonant circuit that works like gangbusters.


Another possibility is to run two antenna wires up in different directions (to two different towers as described) and use a variable phasing network to null out unwanted stations... ala like MFJ does to null out noise.  In the end, all the magic will be done with multiple stages of coils and variable capacitors - no active circuits.  

I will post the schematic of my final design... It will take some time.


T

Tom...You are giving yourself quite a challenge by eliminating beverage antennas for the BC band...I have my best luck sorting out the many sigs on the same freq with beverages....I have only 3 up right now... but there was a time when I had 6 operating terminated beverages and it really helped to sort out the sigs...However I would really be interested in your results with the verticals and especially the one with all the radials....Gud luk ...have fun...remember....nothing exceeds like excess....
.Steve

...

Tom, I've been DX'ing the BCB since the mid 1960's. You are really setting yourself up for quite a challenge. 3 years ago I bought the "PIXEL" magnetic loop with the intention of using it on 160m. I quickly realized that it performs very well even down to ~ 25khz.  Last year I used it for DXing the BCB band. I live near NYC, about 10 miles from most of the major 50 KW stations. Let's say that my noise level is not quiet. What I quickly discovered is that the deep nulls it was capable of did more to get the DX than antenna gain. Being able to null out one station and wait for another was a huge advantage. I've also used a KIWA loop for ~25 years - good antenna, but not nearly as good as the PIXEL.

BTW, the PIXEL decreased my noise floor by 4 S units - that allowed me to work 26 new countries on 160m that I otherwise would not have heard.

Here an excellent information how to construct such a loop antenna. In my opinion, the pixel is quite expensive (for me anyway)

https://pa0fri.home.xs4all.nl/Ant/Active%20antenna/Active%20receiving%20%20loop%20antenna%20eng.htm

Very cool guys -  

Thanks for the info. Yes, I've decided to build up some kind of loop as a result.   Maybe a PVC pipe frame, 4 wire turns (400 uH) 10' square mounted on a 10' tower section with a fiberglass mast, rotator. Gotta figure out the tuning remote.

Question:  Someone wrote that if a loop gets too big, its directivity degrades... is this true?   I would think 10' on the BCB is not too big, compared to a common small 2' loop.

Also, does height above ground play a role?  I could mount is on my 50' tower or use a 10'.  Obviously 10' high would be much easier to test and maintain it.  Another point is that it is vertically polarized if fed on the side vs: horiz polarized if fed at the bottom. Wonder if vert wud be best on the BC band....



BTW, [Crystal SET]  check out this guy's design and YouTube demo. He uses separate modules to move the stage coupling around. Listen to how effective his simple but amazing crystal radio is!!  He obviously optimized his design and even sold 16 on eBay.

https://www.youtube.com/watch?v=c-vtvxskcP0
https://www.youtube.com/watch?v=9J89DGnNcqY
https://www.youtube.com/watch?v=No5th3UurlI


A detailed description:

https://www.ebay.com/itm/Two-Piece-Loose-Coupler-DX-Crystal-Radio-amp-High-Gain-Audio-Amplifier-/151954045960?nma=true&si=4jtjBPxQcJHdDLJG7wTsFh6HMgw%253D&orig_cvip=true&nordt=true&rt=nc&_trksid=p2047675.l2557

SCHEMATIC Below.


Bob / W1RKW sent me a really good website of a guy who builds a lot of crystal sets.   I thought his "HobbyDyne II" was an interesting design.  Lots of ideas there.
http://www.hobbytech.com/crystalradio/crystalradio.htm



T

A magnetic wide band loop should have only 1 turn of fat wire or better tube. More turns do degrade the sensitivity due to an increased inductance. Even better a figure 8 loop of tubing. The material losses arn't important, the wire size is in order to lower the inductance. The amplifier should have a very low input impedance. Two of the general purpose amps that I did publish in the tech section with out an input resistance are excellent. Couple the outputs via a balun to the coax and set the current to approx 10-20 mA using 2N2222.
But I guess that you are not allowed to have an active antenna, so I am afraid you have to tune it and need more turns, a completely different approach. Here is a very good article about wide band magnetic loops.
http://www.lz1aq.signacor.com/docs/wsml/wideband-active-sm-loop-antenna.htm
Indeed multiturn NON shielded loops may have degraded paterns due to electric field pick-up

Hi Tom,

Crystals sets - this is the guy I think of - Dave Schmarder N2DS.  I saw him at a number of hamfests in my area a few years ago.  He told me that with his highest Q coils he was able to separate adjacent channels at the top end of the broadcast band.

http://makearadio.com/crystal/index.php

Tom -
You might want to try looking for "medium wave dxing" on Google. There is a wealth of practical information there, including audio clips of B/C DX, by individuals and some of the many world-wide organizations devoted to that phase of the hobby.
A loose coupler sounds like the way to go for a crystal set; at a club meeting several years ago, one of the members demonstrated a loose coupler (really loose!) made from PVC tubing with a foundation tube over two feet long, on which one of the two tuned coils could be slid for optimum coupling.

Keep us informed about your experiences!

Bob - NE

Thanks much for the info and reading material guys!  Yes, Tom - N2DS sure has a big collection of sets he built.

Nico / PA0NVD:  I am intrigued about the tubing loops you suggested and the wideband active small loop amplifier circuit.  I have a lot of aluminum tubing I could use for the loop conductor. If I made an active loop as the web site suggested, how big would you make the loop? The figure 8 configuration sounds like the way to go from what I read.


T

Well, well looks like LZ1AQ makes a broadband amp...   and here's a figure 8 loop made out of hardline:

http://active-antenna.eu/

And he sells the broadband amp for 89 Euros:
http://www.creative-science.org.uk/loop_lz1aq.html


http://active-antenna.eu/antennas/

T

Tom, I have a very close friend / former co-worker who lives in Japan. He builds crystal radios and recently built one that captures RF and rectifies enough to drive a very low power audio amp (IC) that he discovered. His design criteria was to have a radio that did not require any external power. If you are interested I can reach out  to him. I do know that LITZ wire was responsible for the success of the project.

Hi Tom
I should start with a figure 8 made of 2 circles of 1 square meter or so or a little bigger. The input impedance of he amp should really be low, close to a short circuit. The general purpose amp I did describe in tech section can do it. When R1 is zero, the input impedance is close to zero as well. And it costs close to zero
In order to have maximum QRN suppression, make a balanced amp from two of those amps using 2N2222. But you can start making just one and grounding the other input, making a unbalanced antenna to start. Than R5 should be 47 Ohms to match the cable
Here are the values.
R1 zero,  R2 1 kOhm, R3 470 Ohm, R4, 470 Ohm, Ropt 220 Ohm, R5 100 Ohm + VCC 12 VDC
Combine the two outputs with a high mu ring core twisted line transformer to get 50 Ohm output.
I should build the amps dead bug method at a non etched piece of PCB in order to get a perfect ground. Build small and decouple the +12VDC with a small 0.1 uF ceramic cap in parallel with a 2 - 5 uF tantalum cap
You can experiment with the value of R2 to get more or less gain.
If every works fine, you can always extend the antenna with two circles more to get a quad. That will lower the noise floor almost a factor 2 more.

Tom
  This is local to me https://www.midnightscience.net/
  Good Luck I'm sure you will have some success.

Hi Bob -


Sure that sounds good.  I am still looking for the best design. From what I read, Q is important for xtal sets and they are very unforgiving compared to active device radios. So any optimized designs using Litz wire, etc will be appreciated.  I read that some hi-performance xtal sets use coils Q=1000, so that requires some doing.  "The magic of winding Litz wire?"  


I think I will make the set work on BCB, 160M, 75M and possibly 40M too, since I'm going thru the trouble to make an active broadband loop RX system.  Might put the 2 watt 6AQ5 PW AM rig back on and make the two a pair.


Nico:    Could you give me a link to your 2N2222 design?  I didn't see anything in the tech section with a search, but then, there are 300+ pages and I probably used the wrong search keywords.

So you said the 2N2222 broadband amp can adjust its gain...  I wonder how far down in dB a one meter loop is compared to a 100' wire on the BCB?   If I were able to bring the loop up to unity compared to the wire, that would be nice.


Also, could you explain this a little more? You said:

"If every works fine, you can always extend the antenna with two circles more to get a quad. That will lower the noise floor almost a factor 2 more."

Please describe how you would build this "quad" from the beginning?



And, thank you Bob/NE and Mike/ARA for your info and links.

T

Hi Tom
The design of the 'General purpose RF amp" is in this technical forum, approx number 7 from the top. I will  attache the diagram.
The gain can be set by R2. Higher value, more gain. The limit will be where the amp's bandwidth starts to be lower due to the limited HFe of the 2N2222 or where oscillation may become a problem. If you want more gain with high bandwidth, you can just add an other amp like this at the output of the circuit. When R1 is 47 Ohms, the input impedance is approx 50 Ohms.  The gain is R2/R1. Because the loop impedance is very low (which represents R1), the gain may be very high. If you need more bandwidth at the high side, you can substitute T1, the input transistor, for a transistor with a higher Ft. But the loop may work upto 30 MHz or more as is
In the page   http://www.lz1aq.signacor.com/docs/wsml/wideband-active-sm-loop-antenna.htm  you find the quad of 4 loops. If you have a figure 8, you can add 2 loops more in a figure 8 configuration and connect them as in the above page. You get than a plane with 4 loops instead of 2.   
so it are 2 figure 8 loops in the same plane 90 degrees shifted. That will reduce the noise floor almost a factor of 2. Nice experiment later on to improve the performance a little more without changing the amps.
But you can design step by step, start with a figure 8 and a single amp, than use a double amp wit a ring core power combiner/balun in order to get a balanced loop and not a loop with one side to ground, than add an extra output amp if yo need more gain (which may not be required), than add an other figure 8 to get a quad.
Please keep us informed, nice project. Here in Costa Rica are hardly any AM transmitters anymore. In the evening you hear many south- and centro American and Caribean stations all at the same frequencies it seems. One on top of the other

Thanks again, Nico!

That homebrew broadband amp is simple enuff and will save me 89 Euros...   ;D


OK on the multi section loops.  I'll build something up soon once I decide.


BTW, check out this guy's YouTube demo of his Ham/BCB  CW/AM/SSB  crystal set.  Crude, but still amazing what such a simple set will do.  Getting some good ideas. I'll bet more than a few guys here are starting to feel the wintertime urge to build a crystal set.


https://www.youtube.com/watch?v=uU6dsJOOILQ


T

Your project started my interest as well Tom. My first need in my new home here in Costa Rica is a 40 meter antenna. When my shack is ready, I like to build me a crystal set as well. Seems a whole lot of fun.

Nico,

I just thought of a slick idea.... I wondered how the heck will I know where in the band I am tuning the crystal set?  Well, ya know the last YouTube demo with the guy doing ssb/cw with his external Hartley BFO?   Just mount one of these cheap Chinese $35 DDS 0-40MHz generators (with digital display) into the set and it will make a rock-solid ssb/cw BFO as well as give accurate digital readout to the crystal set!  (listen for the zero beat)  Since it is the only oscillator in the radio, it will make the crystal set CW/SSB stability as solid as an SDR.   Switch it off for AM until needed for freq display.  It will advance as slowly as 1 Hz steps... ;D

DDS:

https://www.ebay.com/itm/AD9850-DDS-signal-generator-module-0-40-MHZ-LCD-PC-control-sweep-function/111865608785?_trkparms=aid%3D555018%26algo%3DPL.SIM%26ao%3D1%26asc%3D20131003132420%26meid%3D3e6b16eeabc348448d2b34f27faf3d3d%26pid%3D100005%26rk%3D4%26rkt%3D12%26sd%3D321418276142%26itm%3D111865608785&_trksid=p2047675.c100005.m1851


Demo using CW/SSB: https://www.youtube.com/watch?v=uU6dsJOOILQ


Cool on the Costa Rica move.  What kind of 40M antenna are you considering?


T

Hi Tom
That is a good idea!. I have this unit   https://www.ebay.com/itm/0-55MHz-DDS-Signal-Generator-Direct-Digital-Synthesis-HAM-Radio-VFO-Wireless-MF/172385609520?hash=item2822fba730:g:k9QAAOSwal5YD
and several other units that count frequency for boat anchor receiver read-out.
Take care, they may generate some birdies when close to the coils. So you better put it in a small shielded box made from PCboard or so
I did build a Moxon for 20 meters from 4 Chinese fishing rods. For the 40 meters I did purchase a very nice fiber mast from Spiderbeam in Germany. Good price and excellent quality. It is 18 meters. I will mount a little pulley at the top and hoist a 1/2 wave vertical, slightly sloped to have the base free from the house. That gives a nice low radiation angle. I will feed it at the base with a link-coupled parallel circuit. The feed impedance is approx 3000 Ohms, so no radials needed. I just will put 2 groundstakes as a ground, that gives also some lightning protection.
Along the lower half, i will mount a non-connected wire, a half wave passive radiator for 20 meters, at approx 15 cm distance. That will break the antenna electrically in two at the 20 meter band and I get than a vertical array of two half wave radiators above each other. That results in an radiation angle almost constant from 8 degrees upto 50 degrees, so excelent for DX.
In addition, it has 4 dB more gain over a half wave vertical. The antenna will work without that wire as a full wave vertical for 20, but that gives a high radiation angle, not good for DX. A small disadvantage is that I need to make two link coupled LC circuits at the base, one for 40 and one for 20, relay switched.
An other small problem is that I like to have a motor to hoist and lower the antenna. There are frequent and violent thunderstorms here and I am at a quite open and high spot. Very attractive for lightning.... :-[

Hello Tom,
                                                                                                               
 70+ years ago, my first antenna was a short wire connected to the bed springs...worked
great.  I did not have a headset, used an ear piece from the old crank telephone set.

Enjoy the fun.......

73,
Bob, W4WSZ

Since you’re using a crystal radio, you’re going to need all the selectivity you can get. The loop antenna suggestion is good as it provides some directivity. Adding a BCB varactor for tuning will dramatically improve its performance.

I built one. The varactor diode, an NTE 618, is in series with the loop and physically mounted at the top. A bias voltage is carried through your coax and coupled to the loop at the bottom where the impedance is low. The connection is RF bypassed with a .01 cap and has two 1M Ohm DC coupling resistors in series with each lead of the coax (the diode current is minute). RF coupling is accomplished with a snap-on toroid around the loop near that connection. The secondary is 3-4 turns through the toroid with a DC blocking cap in series (to not short out the bias on the coax).

Tuning is accomplished by inserting 0-9V into the coax at the receiver end. I used a 5M pot, a 1M coupling resistor, and a .01 DC blocking cap.

It’s a lot simpler than this sounds! The pics below show the details for a 160-80M version.

Interesting, Clark!

So you have a way to tune the loop with a varactor diode instead of using a broadband circuit. Certainly worth trying out.

What I am wondering now is how much effect does the inside house wiring (and aluminum siding, if any) have on the loop patterns and ability to null -  ever tried it outside?

Also, if a loop were ground-mounted outside in the clear, how much better would it work at say 50' off the ground from BCB-40M, remotely rotated?

I decided to put my longwire 250' wire hooked to the top of the grounded 100' tower in the air soon as a reference antenna against any loops that I build.


Nico: Great on your plans for the bigger antennas.  Sounds well thought out.

Bob: FB on your bedspring crystal set. Back in the 60's I found that tapping the metal clip on a landline telephone gave the biggest signal by far... :-)   I had one of those rocket xtal sets.

T

Hi Tom
I did build an amp and connected it to a simple loop of AWG14 wire, so too thin. Just to see if the amp is stable with this input.
Attached a pic of the amp.
It is completely stable and works quite nice. I didn't try further away from the house yet. very close to the house there is a whole lot of magnetic pollution of switching power supplies, computer, WI-FI system etc, so a lot of birdies and noises.
Seems to work upto 30 MHz with the 2N2222 as expected Underneath the metal roof above the deck I did receive SSB from the USA at 14 MHz

The tuned loop works FB at any convenient height above ground. I tried it with great results on a 5’ mast with a rotator on a tripod. I doubt that there would be much improvement with additional height given that BCB signals either travel along the Earth’s surface (day) or arrive from the ionosphere (night). No need to reduce that distance by 50’ ;-)

I usually use it indoors, and despite the chicken wire within the walls, it still seems to work effectively. You can use a portable radio with signal meter to check signal strengths inside vs outside.

Hi Tom
What about using the aluminium rim of a bicycle wheel?

Is a crystal radio connected to an amplified antenna still a crystal radio....?

I don't think it is valid for worked all states challenge, but still nice to experiment with. And simple low cost wide band without tuning.

speaking of amp concoctions, being sometimes lazy I have adapted the venerable "video distribution amplifier" to RF preamp use. They are not nice like the above and sometimes need a bit of hacking to get above 10-15MHz, but they are a dozen for a dollar these days. Some were made for early VGA computer analog signals. Those have better bandwidth, but it's merely a matter of removing or downsizing a few HF-reduction caps that the dist-amps used to keep the bandwidth contained  to that required.

Nico:  Yep, an aluminum bike wheel would make a FB loop.  I'll give them a call and see if they have a pair of junk ones.

Johnny:  I think if the active matching board is set at unity gain, it would be a reasonable fudge.  

Clark: Then probably at this low frequency, height will not effect the take-off angle barely at all. The only reason to get it higher is to get it in the clear away from noise. In this case, I'll bet locating it on a small 10' tripod in the middle of the antenna field away from everything would be the best bet.

So essentially, there are four adjustable parts to the system to help select a station for best S/N:  1) The rotatable loop for signal level discrimination,  2) The antenna matching at the loop and/or antenna peaking in the radio  3) The diode detector circuit peaking   4) The variable coupling movement between the two stages.  That's quite a lot of variation.

Another thing to consider is audio. The set will do OK using suitable headphones, but a speaker will probably be too low for weak signals or pleasure listening. I plan to add a small audio chip to drive a speaker when needed.  The DXing can be done roughing it on the headphones.

I laid out 250' of #10 antenna wire today on the ground. This reference antenna will be stretched between the 50'er at the house to the top of the 100'er which is grounded in a radial field. I'll bet this system would make a great LF antenna.  I'll make a small pi-network to test it on my station receiver.  I'm curious how strong BCB signals will be compared to the 160M dipole at 190'.  We shall see.

After talking with Bob/W1RKW, for the crystal sets we both decided to start with standard cylindrical coils, single layer that are able to slide in and out of each other. Will hold off on the Litz wire for now. Another thing is substituting a Schottky diode for a germanium for less voltage drop. Must test.

T

Hard to find, but the most effective detector diode is the backward diode. Almost an order of magnitude better. Essentially it is a tunnel diode where the current reverse peak is at zero volts, It has virtually no voltage drop, in the mV range. Someone has friend in the military radar world? Or microwave labs? There they are sometimes available

I will axe around. Anyone have one of these diodes here?


BTW, thanks for posting the picture of the built broadband matching circuit board.  That will be my first go-to project once I build the loop. EZ as pie.


T

I might have some, not sure. Years ago a police officer friend told me how they confiscated police radar detectors and smashed them before disposal. I asked for and received a box full, 20 or 30 which I salvaged for parts.  They all used a pellet shaped microwave diode detector in a cavity mixer behind a feed horn. How does one measure these diodes to determine what they are?? I have a bunch of them.  Can mail them if of any use.

73s Nigel

Edit:  On further reading they might be zero bias schottky in PK37 package see link.
https://www.semigen.net/zero-bias-schottky-detector-diodes/
If so they may not work at low freqs, dont know, then this link may be of interest.
http://213.114.131.21/_pdf/DIODES/zero_bias_schot8.pdf

Hi Nigel
I don't think that are backward diodes. Most probably it are the gun diodes or the mixer diodes. There is no need for backward diodes in a radar gun.
These things are very nice to make 10 GHz transceivers for ham use. I used the "door opener" type of gun radars for transceivers about 40 years ago and we could span easily up to 30 miles using wide band FM

This guy claims better performance using a zero bias mosfet as the detector. Just FYI.
http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf

73s  Nigel

http://aldinc.com/pdf/ALD110800.pdf

Nigel:  Very cool.  A quad FET array that will run with no external power with zero voltage drop as a detector.   I must try it.

It reminds me of taking a Model-T car and adding in modern parts.  Like adding solar panels and an electric motor.  

I suppose if we were to stick with original, it would mean using a real galena catwisker.  

The article also mentions adding a toroid with 9 turns to kill IMD caused by TV and FM stations.


Between the broadband loop, FET detector,  coil designs, audio level enhancements, etc., there's a lot of optimization and experimentation fun to be had. I'll bet one of these 75M receivers will work FB on 75M AM. I wonder how many AMers have made QSOs using a crystal set on AM?

T

A MOSFET detector? Sheesh.

Here's the rig you should be building:

(http://bizarrelabs.com/images/foxhole4.gif)

:-)

Yep, just like a "Foxhole" radio. The prisoners in Eu WWII were quite ingenious.


Bob/ W1RKW has been doing some research.  Turns out the 5082-2835 Schottky diode IS the best  (0 dB loss) - but only if a tiny 10uA bias is applied. A small button calculator battery could be wired in and last a very long time.  This is as good as the MOSFET array! I don't have any info on the back diode yet, but 0 dB is as good as it gets.   (A 1N34A is about -3Db loss and most diodes are worse from there)

Here's the diode tests and other ideas.   Look at the 20 Mhz column at the bottom of the page under "Experiments With Diode Detectors:"

http://www.techlib.com/electronics/crystal.html


T

Tom,
Take a look at this high performance crystal set
https://hackaday.com/2016/09/16/high-performance-crystal-radio/

Chris uses an audio amp in most of the video but does demonstrate using audio straight from the detector in his overhead cone speakers.  Gets respectable volume from them.

The selectivity from his set is amazing.

rkw

Bob,

Wow, that really is some system..  He uses four tuning circuits total.  I like how the floating bandpass filter, with nothing physically connected to load it except coupling, is inserted in the middle.  Selectivity sounds almost as good as a tube radio with I.F tuning.  The multiple hi-Q stages are the key. And no power used at all.  (Or pick up another 3 dB by using the biased Schottky diode.)


Back to Litz wire and ABS 4" forms.   I like how the coils sit out away, in the clear, from everything for best Q, least coil pattern interference and best selectivity.  He really has optimized it to the best set I've seen to date.  When he turns that tuner coil 90 degrees, it is absolutely dead.  Clean.

It really paid to do our homework first before starting anything.

https://hackaday.com/2016/09/16/high-performance-crystal-radio/

The schematic is at 13:41  -  It is clear enough to figure out.   Note that the antenna unit tuning variable cap is two-ganged  and the sections are used individually as a series tuner and as a parallel tuner.  Designed for a long wire, in his case about 250'.   Imagine further signal discrimination when adding a rotatable loop.

I believe he is picking up a higher Q by using smaller coil inductances and larger C ratio with that huge 4-gang 500 pF sectioned cap - more than with the usual sets I've seen.. Compare his small coils (less turns) to the Plexiglas crystal set with those huge pi-wound ones.  Less L, more C, lighter coupling, less loading = higher selectivity.

Coils mounted on 10"? insulator pillars using nylon hardware. Capacitors floating on Plexi. Selectable audio transformer taps.  He squeezes every uV out of that set.  Well thought out.

And he is even tapping enough power from the bandpass filter (coupling loop) to drive his digital readout the proper way.

T

That's a good observation on the small coil windings.  I didn't catch that. That's good to know.

I like how he uses a rotatable coil on his antenna tuning  unit which is cool section by itself and is controlled by a knob so his body influence is reduced or eliminated and all the sections use 6:1 reduction gearing on the variable caps.

He did mention that he has a switch to select the detector diodes. I'm not sure what is gained by using 2 in parallel.  Why not just use the better performer unless the tuned circuits play a roll when selecting the diodes. Maybe there is a give and take between the diodes and tuning networks.  It's curious.    

One question I have, I'm going with a long wire and how critical at MW would it be if the wire kind of zig zagged. I'm thinking the straightest, longest and most vertical run would be best but my situation prevents me from doing that.  I can get length but I'd have to zig zag some amongst my natural supports.

I'll be starting my winding next weekend. Hope to have a roll of Litz by mid week.

OK Bob -

What gauge/ source of Litz are you getting?   His coils are so simple, might as well use Litz, WTF.

I'd like to try that Schottky diode and see zero loss.  I just KNOW that's where I'll end up anyway... :-)

He mentioned doing the BCB challenge - to log as many stations as possible over a week, etc.  I wonder what is the most effcient way to do this? Obviously we will have to either wait for them to ID or use a computer data base to figure it out. The loop null might help.

Yes, tuning with long rods is a great idea to minimize hand capacitance.

I was actually thinking of mounting the coils up high like that before I saw his video, in addition to experimenting with high C, less L.  Good to see it has already been proven out and saved time.

As for your antenna....   just figure you want at LEAST 1/4 wave (or even 1/8 wavelength) out there for the lowest freq to feed against ground. This will give you 20-50 ohms to play with, depending on ground efficiency.   IE, the most important thing is to get the wire stretched out even though it may not be perfectly straight.   Obviously on 900 Khz the wire will need to be 300' or so, so you will be dealing with 10-20 ohms for the lowest freqs. So bottom line is just like LF band, get as much wire out there -  as long as it is reasonably stretched out. No spider webs, even though it will add some capacitance hat effect.. If it folds back on itself too much it will simply mean increased pattern cancellation.  Bear in mind that the AM radio antenna on a car is only a few feet long and matched well -  with VERY poor ground to work against and works "decent" ... So you will have a huge advantage with even a 100' straight wire in the trees fed against a good Earth ground.

T

Tom,
I ordered 150 feet of 660/46 Litz off of EBay. i'm going for big Q.  It was the least expensive source I could find and seems to be the most readily available place for a variety of Litz.

What I have noticed over the years with DX BCB, that stations don't seem to ID as often as they should especially those that are automated.  So one could wait a while to get call letters.  Even  WTIC doesn't do the obligatory ID.  How they get away with it, I don't know.  But it is what it is.  It's part of the hunt.

OK on the length of wire and fold back.  I think I'm ok in that respect as far as short wavelength goes. Longer wavelenths might suffer but I'm thinking since the exit out of the shack is at ground level and will essentially spiral out up and out from that point to the highest point I might be OK.  I can string the wire in a nice wide open and expanding semi circle which will be close to 200 feet.  As far as ground goes, I was think of just using the electrical service ground. It's a few ground rods located near the shack and attached to the cold water pipe in the shack. We have a water well that is fed by plastic.  What do you think about an independent ground at least for the crystal set?

To aid in identifying stations, check out this website that shows the day and night patterns and power for AM stations in the U.S.:

http://www.nf8m.com/patternmaps.html (http://www.nf8m.com/patternmaps.html)

Many stations stream live (with some latency) so you can confirm what you’re hearing after narrowing the possibilities.

Now that's cheating..  :'(   S'posed to be completely self powered..

Nigel

If you don’t need a lot of current then you could also build a second crystal set with a 1N34 to use as the power source. Longwire and tuning circuit for peaking output and electrolytics for filtering.

Do that on a larger scale and you could power an ocean liner just on radio waves. Here’s proof: https://en.m.wikipedia.org/wiki/The_Big_Broadcast_of_1938 (https://en.m.wikipedia.org/wiki/The_Big_Broadcast_of_1938)

Seriously though I had the same thought.  Tune a second RX to the strongest RF signal in your location whatever that is.  Harvest the energy to power a second Epad mosfet audio amp.  Note the spec sheet applications first listing. 
And quote:
With these zero threshold devices, an analog circuit with multiple stages can be constructed to operate at extremely low supply or bias voltage levels. For example, an input amplifier stage operating at 0.2V supply voltage has been demonstrated.
unquote

Be a fun engineering exercise.  Ya I know we're driving you nuts so will shut up.

73s  Nigel

Just to make things clear, are we are allowing the introduction of modern solid state devices into this "crystal set", but we have an "unwritten" restriction that it must be self-powered?  This would lead me to believe that we might even add a nickel cadmium rechargeable battery to the mix (but lithium ion batteries are forbidden, because they are just too new).  All the power is still coming from the broadcast station, yes?   Throw in a small switching power supply and you could charge not one, but a bank of nicads.  Then you could power a FB "Hi Fi" audio amp, not a measly .2 volt device.  Let's hear that Hi Fi broadcast signal all over the house!  

Sounds like a fun engineering contest is about to unfold!

I can recall back in the '50s when I was about 8 years old, I made a crystal set, but found that an output transformer would take the impedance down to a level that would drive my headset or a small speaker very nicely.  The UTC LS-55 worked best!  Then my dad came home from RCA and gave me several 2N109 engineering sample transistors, before you could buy them in the store.  A two-stage audio amplifier was then powered by three jars with a nail and a copper strip, filled with vinegar.  I found I had enough sensitivity that I could dispense with the outdoor antenna, and just clip one side of the input tank to the finger stop on the dial telephone, using the AT&T free antenna wires.  Listened to Jean Shepard (K2ORS) on WOR after midnight every night!

Here's a screenshot in case you missed it (it lasts for just a moment) [#1].

I've also found that using two diodes for full wave rectification dramatically increases voltage. This is done by A.C. coupling the tuning components to the junction of the diodes. [#2]

This is getting very interesting!

Funny, cuz when I made the first post, I didn't know there was a REAL official WAS BCB challenge. It was my own idea at the time.  But from the various websites we see it mentioned and it likely has strict rules we will need to follow if we participate in their contest.  (Actually it's "Work as many BCB Stations as you can in a week - anywhere in the world")

I would assume no external battery power is permitted, just RF.  BUT, detecting RF and using it for power is fascinating. It would favor users that are near a big BC station and those in the rural areas would have very little to work with. I have WTIC, a 50KW station 20 miles away that could probably power an audio amp, detector and broadband antenna matcher.... :-)

So my guess is detected RF "auxiliary" power is NOT permitted in the contest due to the unfair advantage for some. But then again, location is different for everyone and will affect their overall results anyway.  We must get the rules and see....  In the end, it's just for personal fun and nobody really cares who "wins."

There's really two projects here:  One is a completely stock, conventional 1N34 detector with sensitive headphones - and the other is doing whatever you desire by hot-rodding the set to suit your drag racing urge in the unlimited class.   Receiving is unique because nobody can see or hear what you are doing, so bending the rules is just fooling yourself.

That said, I'd like to set up an experiment detecting 50 KW WTIC and charging some capacitors to see how much power can be tapped. I need to get up the new wire antenna and build the network first.
 
Clark: OK on the FCC database with field strength, etc.  That will help a lot. We need to find out what the top contesters are using. They most likely have it down to a science. Or, who knows, maybe a verbal ID is required to log it in...

T

I'm not seeing how a separate detector circuit for D.C. power would *not* be permitted. Worst case it would generate (pardon the pun) a forehead slap and a "why didn't I think of that!" moment for the competition, who you better hope isn't perusing this thread.

BTW, if you were 1 mile from a 50kW signal, you'd also be at a disadvantage as you'd have plenty of power yet a swamped receiver. A crafty design might combine both circuits using one antenna, or better, see if your diode generates it's own bias voltage since crystal sets produce both an A.C. (audio) and D.C. component.

Bah! I'd just go with the separate circuits fed from a dipole with open wire feeder connected to your "A1" and "A2" terminals. One antenna. Done. Dismiss objections as sore losermanship.

During the 1921 Transatlantic Tests, where amateur operators competed to be the first to span the Atlantic using shortwave frequencies, was it cheating that 1BCG, which was built and operated by pioneering luminaries of radio holding dozens of patents among them and included Maj Edwin Armstrong, was the resounding winner with the most strapping signal? Heck no.

Sheesh.

Here are the rules to a sample crystal set corntest.  It seems they divide the competition into many classes:

"All reception must be strictly passive, i.e. no signal power amplification before, during or after signal detection in the receiver, to include active onboard or outboard devices, such as converters, oscillators, "free power" circuits, rebroadcasters, unity gain amps, audio amplifiers, and tuners. Such devices, however, may be used to search for and acquire stations, but in order for the station to count for score, the station must be clearly heard when these devices are not in use."

Full Rules:
http://www.crystalradio.us/crystalcontests/index.htm


These guys seem to be the crystal set "ChannelMasters".   Check out some of the sets these guys have built. I was feeling kinda smug with my xtal set progress -  but not anymore.   ;D

http://www.crystalradio.us/crystalradios/2010-1.htm


T

b. Detectors with more than one rectifying junction, such as transistors, JFETs or Mosfets may be used providing they are not used in a manner that will amplify the received signal. This does not prohibit the use of multiple detectors in any series/parallel/bridge arrangement, nor does it prohibit the simultaneous use of more than one receiver and/or antennas.

Looks like you're good to go with the separate D.C. supply receiver except that you missed the contest by 9 years.

A-HA!!

The MOSFET detector and bias batteries are permitted as long as they do not produce gain:


b. Detectors with more than one rectifying junction, such as transistors, JFETs or Mosfets may be used providing they are not used in a manner that will amplify the received signal. This does not prohibit the use of multiple detectors in any series/parallel/bridge arrangement, nor does it prohibit the simultaneous use of more than one receiver and/or antennas.
c. Bias batteries are permitted to overcome inherent detector threshold levels, subject to the restrictions on signal amplification of 4.b.

They are 3885 AM HAM friendly:

(5) TWO-WAY SHORT WAVE CLASS (HAM OPERATORS ONLY)
The equipment restrictions for this class are the same as for the Short wave class, except a point is awarded for each two-way transmission. The crystal set is used as a receiver. Tube or transistor AM transmitters are allowed, with power and band usage per FCC regulations. A common Ham AM 75 meter frequency in the Southeast is 3.885 MHz.

The MOSFET detector seems to take advantage of using another crystal set so to speak as a power source but rolled up into one using the same single antenna and circuit. It's explained in the link below.  

The author claims a large external antenna is not needed.

http://www.arrl.org/files/file/Technology/tis/info/pdf/culter.pdf

Awesome on the rules, Tom.

Wow antennas bring out all the fun! Anybody can build an antenna! They all work minus some decibels or so. Hee.

I am thinking two orthogonal delta loops from the tower. Loops are quieter. With a switch or better yet - a stepper driven driven GONIOMETER at the base. Now you can electrically steer the null.

The mechanical construction involves two orthogonal stator windings, and a rotor winding.

As the rotor turns through 360 degrees two peaks and two nulls will be observed. If you add in a vertical element after, you can resolve the ambiguity and you have just reproduced a 1909 Bellini-Tosi Direction finder!

I shared this with my friend in Japan.He shared some details on the radio he built. This might be of interest to some of you.


 I already built it with purchasing special Zero bias FET ALD110900 directly from TI as per attached photos of my project radio.
This radio, it works fine and have re-generation function in it so that why very good sensitivity.
Sometimes radio start regeneration sound of Pi------- while tuning.

The original circuit which you sent me this time, using one regular bar core while I tested multi numbered core to wish higher gain..
The coil Q wise come out around 500 and it is pretty good for this type radio.
Normal coil can get only 100 to 300 even with spider coil.

Then I found another article of Q1000 coil which using special high Q core, R40C1 type from China.
Then it sudden jumps Q up to 1000 and over.
Unfortunately, Today, there are no sufficient coil Q meter, only a few old one by HP or Yokogawa, some old Heath kit so.

So I decided to build Q-meter myself and measured it.
However, my latest coil with my Q meter designed Q up to 1000 become over scaled.
Too high Q than expected.
So I remodeled my meter to up to 2000Q. And measured again.
It come out around 1500 at highest while BC band, wow! What a Nice high Q coil.
I kept searching nice Q coil and core and found someone up loaded Q2000 coil available…
It seems endless challenge to increase coil Q.…

Now my radio is built up with Q=1500 level coil and Zero threshold FET.
One single coil can very small coil inductance so that I connected 5 series for AM band with 700 PF variable capacitor.
One single coil may be used for up to 3.5MHz  w/ Coil Q over 500-1000.
Maybe my toy equipment, old variable capacitor, my poor design…

Beside I am challenging battery less radio and collect electric energy from the air, from radio signal around.
To do so, I have to reduce power consumption, the ALD110900 is good staff and I could operate radio just in 5 to 10 micro A.
IN my home, I can get around 100 micro A and around 5V from local high power AM station which 200kW and about 20km away.

Now continuing challenging to operate loud speaker.
Lately I found TI made RAIL to RAIL Op-Amp which super low power consumption of just 2micro A while idling.
I built-up it in bridged connected Amp and now operating 4 micro-A(because of 2 IC idling current) idling + actual power input by audio signal and operate in about 100-200 micro A to derive loud speaker.
The Audio Power is just around 1 mW. 
It is very tiny power, but good enough to drive high gain loud speaker with gain of 108 dB/W.

Those 200 micro A makes sometimes distorted since voltage drops much.
So, adding 1F / 5.5V gold capacitor and charged up at night and comfortably operate radio up to 300 micro A peak level.

Thank you very much for your continue support me and have a nice day, good care f you.

Sam

Hi Bob -

Thanks for the article from Sam in Japan.

It seems that Q plays a big role, obtaining 1500Q- 2000Q!   He also uses big C and small L, which we talked about for high Q.  So instead of air-core coil forms we are now talking ferrite.

Here's some on e-Bay showing  1300Q using Litz wire:
https://www.ebay.com/itm/NiZn-Ferrite-Toroid-R40C1-37x23x15-mm-for-High-Q-SW-Barron-Crystal-Radio-Coils-/151443830630


Here's some dialogue from builders about these cores and high Q in general for BCB sets:

http://theradioboard.com/rb/viewtopic.php?f=2&t=5586

I see he is tapping RF for power and using the MOSFET detector.

He mentioned using a regen loop in his set for better sensitivity.  Interesting.

It will take some time to digest all of this to make into a radio - so much has advanced.

T

Pictured - One of the nicest sets built by these guys:

I'm going to ask Sam to join this conversation. Sam is an avid ham and a retired engineer from our microwave oven factory.
Here is two more photos of his radio.

Hi Tom
He talks about regeneration and sometimes oscillation. That means that he has amplification, which is not allowed in the challenge.
There is also an other FET at Ebay. Lower cost and they claim to be better
https://www.ebay.com/itm/MOSFET-3SK143-Q-3DQ-for-Crystal-Radio-Zero-Voltage-Threshold-Detector/161600513556?hash=item25a0242614:g:YBcAAOSwPhdU3sHm:rk:11:

FB, Bob -

Yes, Sam can certainly help us with the latest applied technology.

Nico, looks like the MOSFET detector and the new cores are the ticket:

"This MOSFET 3SK143-Q is in the small surface mount package. It has the code "3DQ" printed on the surface.
Traditional crystal radios are used mineral detector or semiconductor diodes to detect radio signal. In recent years amateurs discover that some “zero-voltage-threshold” MOSFET such as ALD110900 or ALD110800 can substitute the diodes in crystal radio without use a battery, only need to add a coil between the souse gate and ground.
In recent years some Chinese amateurs discover the other kind of  MOSFET - 3SK143-Q can substitute the more expensive ALD110900 and have better performance.
The MOSFET 3SK143-Q is not “zero-voltage-threshold” but very close to, so it still can detect very weak signal and have superb selectivity in the crystal radio. Another advantage is that the output impedance of 3SK143-Q can be as low as 1 to 3 KÙ (the ALD110900 is about 25 to 50 KÙ), so you can directly connect the low impedance headphone to the crystal radio output connectors, don't need to use an impedance matching transformer. With the small Ni-Zn R40C1 Ferrite Toroids you will be possible to make a good performance pocket sized crystal radio."

T

I got a response from Sam this morning with more info …

Dear Bob-san,
I checked the forum you informed me and also found you did uploaded my text and photos.
I am sorry, found I made some mistake in it. In regard ALD110900 FET is not a TI made but Advanced Linear Devices Inc.  the OP-AMP I am talking is TI made.
In regards to that 3DQ FET, believe it or not this was a Panasonic semiconductor, an old FET for UHF tuner use sometimes ago. I did tried it too sometimes ago...

In regard micro Watts AMP, it is built on tiny conversion PCB as shown.
You can compare with tiny transistor radio Volume control.
My latest Q1500 coils with ANT and regeneration coupling design.
On the PCB, it has air power collected power supply circuit consist of 2 small brown micro inductor top left part on PCB tuned with fixed and variable ceramic capacitor then rectify by shotkey diode for local strongest station.
Those coil maker, single or double combined and also type of diode will wide vary output voltage and current even with same antenna and ground hooked up. The LED is monitor when tuning, then switched off rest of the time.
Now PCB was redesigned little bit more.

Multiple ALD FET’s can be used for higher gain reception for very weak station.
use switch to make double, triple...   But one thunder storm surge cause damaged those expensive FET's and I lately adding RF chokes to bypass grounding at the gate. Since then no damage occurred.
Above are just additional follow up for my radio project.
Anyway, it is still under going for upgrade and may be no end.....

Thank you very much for your attention and good care of you.
Good night from Nara-Japan.
 Best 73 and FB DX. TU..
Sam

Hi Tom
You really started a nice thread Tom, its becoming very interesting!!
I did some calculations and test and like to share some thoughts with you. I put my idea of a design below and appreciate comments. In my opinion he designs with the loose coupler are close to the optimum
In order to extract the energy out of the resonating circuit, the loaded Q should be approx 4 times less than the unloaded Q. When a resonance circuit has a Q of 1000 and you load it down to 500, that means that 50% of the energy is lost in the circuit, that is not what you want, you want it out of your detector.
So a single LC circuit wit a Q of 1000 loaded down to 250 will give you most of the energy with a bandwidth of approx 4 kHz.
A big problem will be adjacent transmitters, interference, so you want a better adjacent channel suppression.
The best answer will be in my opinion a bandfilter design. That gives  however a more difficult tuning especially with the high Q circuits.
A band filter design needs two circuits with the same Q in order to have the best curve.
When my shack is ready, I like to build a bandfilter receiver with one double capacitor for ease of tuning and a fine tuning for one or both circuits to get a perfect tracking
The two circuits will have an adjustable coupling to get a QxK = 1, you can peak the signal with fine tuning and the coupling.
That was done correctly with the loose coupler designs. But the coupling of ring core coils can be realized as well with small variable top capacitors.
The load of the circuits should be so that the loaded Q is approx 250  for circuits with an un-loaded Q of 1000, so an adjustable antenna coupling and adjustable load to the detector are required. The antenna coupling can be a series  capacitor for long antennas, the detector load has a lower impedance and may be realized with switched taps at the LF transformer.
In addition some parasitic circuits can be made to notch-out big disturbing signals
As you see, except the double capacitor with fine tuning, this is the loose coupler receiver!!!
The detector should be a zero bias FET, eg the 3DQ
Please comment what you think of the above Tom

Very good, Nico -

Interesting on the Q loading.  If I were to stick with coils made of Litz wire that are on 4" ABS, can I get Q=1000 and see performace approaching the ferrite core coils?  I prefer to stay with the air wound coils for a few reasons.

OK on using the dual bandpass filters with two similar Qs for best coupling and selectivity with the smaller fine-tuning cap.  I plan to put together a system similar to this crystal set in the link below. Who knows, I might add two variable bandpass filters tuned off seperate sections of the same 365 capacitor as you suggested.   The air coils make it easier to experiment with.

https://hackaday.com/2016/09/16/high-performance-crystal-radio/

What circuit do you suggest for a variable parasitic trap?   Maybe a parallel resonant circuit  in series with the antenna input area?

The 3DQ sounds good as a detector.

I am still reading and deciding what to do, but I like the layout and "experimentability" of the set linked above.  (notice how "Moon River" by Andy Williams plays... beautiful example. Listen at 8:55 in video)

Thanks again for the info and advice, OM.

T

This is my current building goal:

Sam-san,

I appreciate your taking the time to post your comments and experience. Yours is an advanced set that we will be looking closer at as we progress.

Your info about the R40C1 type cores from China is encouraging for high performance.

I will study it more and have more questions later.

Feel free to comment on any of the posts to date.

Thanks again and greetings to you in Japan.

Tom, K1JJ

Tom-san, you've reminded me of the rocket crystal radios I saw as a kid in the late 1950's; a cousin of mine in Ohio had one, but no one in my 'hood ever had one to play with.

http://www.crystalradio.net/misc/rocket/index.shtml

(Other info on this guy's site also.)

would it be possible to use 2 detector diodes in a voltage doubler arrangement expect it to work and would there be an improvement in audio output?

i'm still gathering materials but will give it a go at some point.

update: I think I found my answer. There's no free lunch.

Hi Tom
The set with the cylindrical air coils seems very nice to me. You can choose between cylindrical or spiderweb coils. I suppose that the difference isn't that big.
I guess that you can approach a Q of 1000 when wound with litz wire. It is easy to vary the coupling by turning one of the coils. The traps can be made of single LC resonance circuits coupled loosely with the antenna coil. So, if you make the position of the antenna coil fixed, you can place the detector coil at the right side rotatable and the trap coil at the left side rotatable.  The antenna coil and the detector coil will form the band filter. I should start with a distance between the coils the same as the diameter. Prevent capacitive coupling, let the cold sides of the coils face each other. A shield mate of wires at a distance of approx 1/4 - 1/8 inch connected to ground at one side is a very good manner to prevent capacitive coupling. The coupling is VERY loose for good filter aciton if the circuits are loaded correctly, only 0,2 - 0,3 % !!
You can estimate the Q by loosely coupling a generator into the antenna, e.g via 1 pF or less, peaking everything making the coupling between the detector and the antenna coils as lo as possible and measuring the 3dB bandwidth. When you couple a bit tighter, you should get two peaks, the filter is over-coupled. When both peaks are similar in amplitude, the Q of both circuits is similar.

Bob/W1RKW discovered the mother of all old buzzard radio suppliers.   This guy has cornered the market on 1920's parts. Prices are quite reasonable.

Think I'll buy a pair of 1920's headphones for the "unamplified" crystal set mode.  Take a look at the knobs offered. Unbelievable selection across the board.

http://www.oldradioparts.com/index.html


T

T, the pair on the right is very similar to the pair Gary sent me.  They're almost like new despite a layer of dust.

Tom-san,

That's another great website. Mucho crystal radio info.

The Rocket Crystal Radios:  I had one in the early 60's. I could only hear ONE station on it; WDRC 5KW that was about 1 mile away.  My older brother told be the rod adjustment at the top was the volume control... :-)    That's all it did was make WDRC louder or weaker.   So much for selectivity.

Today I got the new reference BCB receive antenna finalized and connected to the shack. I measured it as 250' of wire strung between the 50' and 100' tower tops with the 100'er grounded to a radial field. It's like an end-fed electrical 350' long wire grounded at the opposite end. Can't wait to try that out.   The loop antenna will have to wait until I build a working crystal set first.

Now that the outdoor work is finished I can play inside as the colder WX moves in..

T

Update on progress:

Good news: I tried out the new BCB 350' long wire / tower antenna.   I built up a C-L  network (1200 pF variable to ground in series with a coil) fed against ground and connected the FT-1000D. I was able to get sharp antenna tuning peaks.   I was using the 50 ohm RX input into the FT-1000D.

I compared this antenna to my two best dipoles. One was the 160M dipole at 190' fed with hardline.  The difference was 10-20 dB on most stations between 540 - 1710 KHz in favor of the long wire.  I'm sure most of this difference was due to the perfect match of the LW compared to the dipole resonant on 1850.  But still, the peak is sharp and I now have signals on AM coming in 55 over compared to much less before. Good place to start with the crystal set experiments.

It even received well on 160M and 75M for locals once matched - compared to optimized coax-fed dipoles. There is a combination of horizontal and vertical components in the various patterns.

T

This is the crystal set we will be building. It's called the Lyonodyne 17. The designer, Mike Tuggle, is well known in the radio hobby world and has been improving it since 1974, thus model 17. He's won several DXing awards with it -  and I like what I see.

Question:

Notice the two traps, trap 1 and trap 2.  These "QRM" traps are supposed to be tuned to the adjacent frequencies to suck out the big stations causing problems to the desired center frequency.  For example, listening to 1100 KHz, traps would be set to 1090 and 1110 to suck out / reduce these offending stations.  (parallel resonant circuit loosely coupled)
  
http://www.crystalradio.net/crystalsets/lyonodyne/index.shtml


Now a different radio (The Moon River" guy) we had considered uses the same circuit (parallel resonant circuit loosely coupled)  and in this case there is only one and he calls it a bandpass filter to peak the desired signal.  These circuits appear to be in the same position and both wired the same - but do opposite functions.

(Goto 3:52 in video demo to see BP filter work)
https://hackaday.com/2016/09/16/high-performance-crystal-radio/


I would think these circuits would peak the signal they are tuned to, broaden out the bandpass when tuned to different frequencies - not suck it out like a classic trap does.     Can someone can explain what is happening here ??

The only explanation I see is that the traps are dropping the offending signals across them and not feeding them forward into the diode circuit.  But then in the other radio acting as a bandpass filter, there is only one path into the detector?  The coupling arrow directions are confusing me - are they significant?


BTW, anyone have a mini UTC A-27  inter-stage audio transformer to sell?   (or equiv  -  T1 UTC A-27: input transformer --100 k-ohm primary; secondary taps, 50 to 600 ohms)


T

Hi Tom
The second receiver does not have traps. The intermediate circuit resonates at the receiving frequency forming a band filter with 3 tuned circuits in stead of 2. That will be quite sharp, but it will be more difficult to get a reasonable pass-band curve because the tuned circuit in the middle is un-loaded. and has a different Q than the antenna circuit and the detector circuit.
I should choose the other design with traps, that gives you more flexibility to notch disturbing signals wich may be the biggest problem.

Tom-san,

I looked at Mike’s schematic and I believe that the arrows are to show that there is magnetic coupling between the separate parallel resonant circuits.  I assume that (not having read  the page) the main on-channel coil is physically in the middle.  And that one trap is out past one end of the main coil and the other trap is out past the other end of the main coil so that they each can absorb power from the main coil at their resonant frequency, and hopefully the two trap coils are not very coupled to each other to avoid interaction as much as possible.

Audio transformers – how about trying a small 2.5/5/6.3 filament transformer, especially if one is a toroidal type?

Tommy-san

OK, it appears the traps simply absorb energy at their resonant freq and don't pass it on to the detector.

I imagine when tuning in a station with side QRM, it's just a matter of adjusting the trap knobs for the best S/N.  Whatever it takes to pull out the station.

Nico, from what you said, is adding a BP filter to the set a bad idea because it will not be the same Q as the others?  I was thinking of adding one (same loose coupling) based on how well the Moon River radio worked.  This would make SIX adjustable tuned circuits.... 2 ganged in the ant tuner, 2 traps, 1 detector and 1 BP.

Add in the loop antenna nulling on freq and it sounds like quite a radio.   I decided to keep it fully "no-external power."   1N34A, all tuned circuits, headphones, etc.


T

No Tom
it isn't a bad idea, just an other approach to suppress adjacent channel interference. What is done is couple much less than KxQ=1, a much too loose coupling. In order to tighten the coupling the tuned circuit has been put in between the two very loose coupled circuits. That tightens the coupling, but ONLY if the extra circuit resonates at the same frequency. But if you want a nice pass band, this method will not be the best and difficult to adjust. If adjusted well, it will give more selectivity at cost of high frequencies in the audio and will be less effective than a nothch circuit to suppress interference of adjacent channels. In addition, if will give a little more attenuation, every resonance circuit gives losses.

Tommy-san:  The 6V to 120V transformer would probably work impedance-wise.  Though I found a 100K to 750 ohm mic input transformer that will be FB for the job.  The 100K side goes across the diode and tuned circuit to keep the Q high so the headphones do not load it.  

Nico:  OK on the disadvantages of using the traps.

You mentioned that a notch filter might be better.  Is there a notch circuit that is passive, loosely coupled and simple like the BP or trap circuits?  

I'm just slowly accumulating parts and zeroing in to what I want to do.

T

Hi Tom
The traps are notch filters, by tuning the trap you can notch a disturbing signal. So the design with the two traps of Mike Tuggle seems very effective to me.

Got it now... will do - thanks!