Crystal set frontend - why parallel LC, not serial?

[aneevuser]

I've recently realised that I'm confused about something elementary:

Why did the classic AM receiver crystal sets use a design based on an aerial feeding a tank circuit i.e. a parallel LC resonator, as opposed to a serial LC resonator?

My confusion stems from the fact that a tank circuit gives current gain at resonance, but no voltage gain. Given that a crystal set used a diode-based AM demodulator, which we need to drive with at least the forward bias of the diode (say 0.3 V for a germanium diode), wouldn't it make more sense to use a serial LC resonator with good Q, and drive the diode with the voltage across either L or C, which will be roughly Q times the aerial voltage at resonance?

What am I misunderstanding here?

[German_EE]

You now have a simple task, build two crystal sets using identical components but with one using a series resonant circuit, then report back,

If it's still available it might also be interesting to try a ZN414 circuit using a series tuned circuit.

Oh yeah, I also have no idea why parallel circuits are always used so I look forward to reading your results.

P.S. The preselector in my ham radio transceiver uses a series resonant circuit

[chris_leyson]

Simply because parallel resonant circuits are high impedance at resonance and a short antenna, compared to the wavelength, will be high impedance and better matched to a parallel resonant circuit. Later versions of crystal sets had taps on the coil to get a better match to the antenna impedance and some coils were even manufactured with a sliding contact and some crystal set even had taps for matching the detector crystal.
If you're going to use a series resonant tank then you will need to drive it with a low impedance source so some sort of impedance matching network or transformer would be required and that would defeat the simplicity of a crytal set with a parallel tank with optional impedance matching taps.

[aneevuser]

Simply because parallel resonant circuits are high impedance at resonance and a short antenna, compared to the wavelength, will be high impedance and better matched to a parallel resonant circuit.

I see. That does indeed make sense - if an aerial looks like a voltage source in series with large R, this would give dismal Q with a series LC circuit since then \$Q \propto 1/R\$ whereas it would give decent Q with a parallel LC circuit where \$Q \propto R\$

However, I'd be interested in some numbers - what would the typical impedance of say a short wire aerial be? I thought that it would only be maybe a few K ohms at most, though a quick google hasn't given me any specific figures. If so, then that still isn't a very good match to a parallel tank circuit, is it?

[aneevuser]

You now have a simple task, build two crystal sets using identical components but with one using a series resonant circuit, then report back,

I've already done this, and the results are what you'd expect - the series circuit gives a nice big signal at resonance, that drives the demodulator just fine, and the parallel circuit doesn't.

However, I've been driving the LC circuits from a signal generator, so I have control over the signal amplitude, and can get both to work by cranking up the amplitude far enough - that isn't an option when driving them from a real aerial.

[hfleming]

Like others stated, the antenna has a high impedance, >1k.  A parallel tabk won’t load the antenna at resonance, so you will get selectivity at resonace, but a series circuit will load the antenna. If you have an active antnenna with a low output impednace on the other hand, a series tank should work better.

[mark03]

I have sometimes wondered about this in the context of tuned receiving loop antennas.  Usually they are designed with a capacitor in parallel with the loop inductance, followed by a high-input-impedance preamp.  Why not put the capacitor in series and make the preamp a transimpedance stage?

[szoftveres]

However, I'd be interested in some numbers - what would the typical impedance of say a short wire aerial be? I thought that it would only be maybe a few K ohms at most, though a quick google hasn't given me any specific figures. If so, then that still isn't a very good match to a parallel tank circuit, is it?

Here's the answer http://www.kennethkuhn.com/students/crystal_radios/antenna_matching.pdf
but I'd recommend all the other articles as well http://www.kennethkuhn.com/students/crystal_radios/

Also, a short antenna is very reactive, an adjustable loading coil (e.g. 30uH - 300uH, a typical ferrite loopstick antenna coil with the ferrite rod out- and fully inserted) can make the antenna tunable and boost the receiver significantly.

Great crystal set experiment and explanation videos:

[biastee]

I see. That does indeed make sense - if an aerial looks like a voltage source in series with large R, this would give dismal Q with a series LC circuit since then \$Q \propto 1/R\$ whereas it would give decent Q with a parallel LC circuit where \$Q \propto R\$

1. Aside from impedance matching as mentioned by chris, the parallel LC also serves as the diode detector's DC ground return.



2. the shunt inductor helps to discharge static; important as low-barrier diodes are fragile.

However, I'd be interested in some numbers - what would the typical impedance of say a short wire aerial be? I thought that it would only be maybe a few K ohms at most, though a quick google hasn't given me any specific figures. If so, then that still isn't a very good match to a parallel tank circuit, is it?

The impedance will vary greatly with wire length, proximity to ground, frequency, and presence of counterpoise, etc. However, if u take IEC315-1 aerial simulation network for a 5m wire as representative, the impedance is 105 - j2181 at 1 MHz.

[vu2nan]

Crystal Radios may be broadly classified as follows:

1. Parallel-tuned, series-fed



The diode-clipped parallel-resonance voltage is output to the phones. High-impedance diodes and phones are required to match the high impedance of the parallel-resonant circuit. This configuration is suitable for reception of distant, weaker stations.

 
2. Series-tuned, shunt-fed



The diode-clipped series-resonance current is output to the phones. Low-impedance diodes and phones are required to match the low impedance of the series-resonant circuit. This configuration is suitable for reception of nearby, more powerful stations.


3. Series-tuned, shunt-fed variant



This is a variant of the above series-tuned, shunt-fed configuration. Using low-impedance diodes and phones, it delivers identical performance.

Regards,

Nandu.

[German_EE]

I was wondering when you would notice this thread  From what I remember you were the crystal set expert on eham.net. Apart from the rules being slightly different (we prefer posts with information rather than multiple links) you will find a similar group of electronics experts from a wide range of fields plus some newbies eager to learn.

[vu2nan]

I was wondering when you would notice this thread  From what I remember you were the crystal set expert on eham.net. Apart from the rules being slightly different (we prefer posts with information rather than multiple links) you will find a similar group of electronics experts from a wide range of fields plus some newbies eager to learn.

Thank you so much for the warm welcome!

Regards,

Nandu.