Crystal Set

[Murphyv310]

Hi.
I've had a little home made crystal set for some years, I think its late 20's or early 30's, its housed in a little wooden box and an ebonite front. It contains a spider wound coil, 20's tuning cap and cats whisker etc. I tried it many moons ago and it didn't work.
Yesterday it was time to investigate. I found one of the twisted tappings on the coil to be O/C. This was cleaned up and re-soldered and this restored continuity through the coil. I connected up aerial, earth and phones, a new lump of Galena in the cup and started to prod about and tune, eventually I received my local, not loud but listenable to, the next station BBC Scotland was hard to resolve. I then decided on a OA47, his made a big difference, with BBC Scot perfectly readable with the local at good volume, two other stations were flooded a bit from the local. I then decided to try an ALD110900 Zero Bias FET. As the coil had plenty of taps I was able to wire this in with no issues, again there was an improvement with better volume and selectivity, the two flooded stations with the OA47 were now readable with just a little splatter from the local. Then an odd thing happened, I wanted to measure the DC current in the phones so I connected the DVM in series with the phones and as I switched through the positions and through resistance all of a sudden the volume came way up. The meter was feeding DC via the phones to the chip, I then added a little button cell in series and did further tests and was able to receive 6 stations all were readable.
After some research one Crystal set forum says that the ALD will give regeneration when you do this or nothing. So I hooked on the scope to the coil and looking at the RF waveform there is no difference between having a battery or not so I doubt that is the reason, perhaps there is AF amplification but the current drawn is 0.85uA and the volts across the phones is 3.3mV with no signal. Looking at the DC current and Voltage through and across the phones on the strongest station is just the quiescent plus the signal. Its all very odd.
Comments please.
                       

[reset]

Trevor see if this helps.

http://radio.radiotrician.org/2018/11/z-...l-set.html

[Murphyv310]

Hi Reset.
Thanks for the info. Fortunately the Spider coil has plenty of taps and the one I settled on gave the best compromise on the ALD, funny enough taps either side made little difference with the OA47, with the Galena it was even further up the coil before it worsened.
The odd thing is with the battery, are these ALD devices not as "Zero" as we are led to believe and all I'm doing is giving it a little forward bias or it is actually turning it from a passive to an active device?
I'll pop up a diagram sometime tomorrow. This one though gives the general idea, all I have added is a 1.2v cell in series with the phones, positive to the EPAD.
   

PS the Phones are marked 6000 ohms, they are of German manufacture from the late 20's.

[Murphyv310]

Hi.
Today I decided to connect the output of the set to my Mullard 3-3 test amp and also observe the audio from the set on the scope. On the strongest station the the audio easily overloaded the amp using the ALD110900 with bias, the load resistor is 180k. Same setup with no bias but much lower output. On the strongest station with bias there is no distortion seen on the scope but less bias there is clipping on one half cycle. With a long wire and a 22pf cap to the coil I can receive 6 stations here via the amp, the furthest away being BBC Ulster, without bias it isn't readable even with the amp at full volume but fine with bias.
I'll do a video tomorrow with my results.

[Murphyv310]

Hi.
Here is a short clip of the set running through my Mullard 3-3 test amp.
https://youtu.be/myfkEysVAgY

[reset]

Hi.
Here is a short clip of the set running through my Mullard 3-3 test amp.
https://youtu.be/myfkEysVAgY

As they used to say when I was a kid.
Cool!

[Murphyv310]

And this as well.
https://youtu.be/TE-RRQjTn98

[Murphyv310]

Hi.
Using the posted diagram as reference there are some anomalies. if we put the scope via at least a X10 probe after the aerial coupling cap (10-50pf) we can see the RF carrier, (at least on a strong station) removing the phones increases the the RF level as to be expected, also removing the filter cap 1nf does the same due to RF leaking via the detector, so we are needlessly using energy in the filter cap. So lets dispense with it by adding an HF choke after the EPAD and before the phones this indeed stops the loss from the cap, a 470 pf after the choke to earth/ground then doesn't effect the RF level. This of course is also relevant to Galena or germanium detectors so we are gaining more audio for one additional component. Doing this will slightly also improve selectivity.
Using an ALD110800 which is a quad IC also allows us to move the tap or the extra winding to a lower number of turns, this then means you can then reduce turns on the secondary winding or tap to allow lower impedance phones, always parallel up the elements in these devices. Doing this today has resulted in a little more output on the horn speaker, plus of course the addition of an RF choke. We are now squeezing the most from this simple receiver.
I've now an inkling to try an FM crystal set.

[Terry]

Trevor, if using headphones, surely the winding capacity performs the RF filter function so no external filter needed?

[Murphyv310]

Yes but the leads of the phones add capacity before the inductance of the phones. My old high impedance phones have twisted cables of around four feet. I did think this was the case too but looking at the RF it's level went up with a choke and just the phones, adding a 470pf or 1nf made no noticeable difference, I then thought Ah perhaps the resistance of the choke is the reason, it was only 23ohms DC resistance whereas the phones were over 4k so I do believe we need proper RF filtering prior to the phones, rather than a cap damping residual RF. We need all the flea power we can utilize.