Finding Piezoelectric Crystal Frequency for Coherer experiment?

[mishima]

Hi there, I've been fascinated by using simple coherers to pick up the electromagnetic signal from a spark (in imitation of early radio experiments, for high school education). I am using a push button ignitor (piezoelectric crystal) attached to a simple dipole antenna for the transmitter. When I push the ignitor, the coherer (which is 2 or 3 meters away) turns on an LED on a 3V power supply.

I also have a software defined radio (SDR) that can pick up a wide range of frequencies (1kHz to 2 GHz, continuous). I would like to use this in tandem with my coherer, to get a measure of the signal strength simultaneously (in an effort to answer the question, what is the minimum signal strength required to get a response from the coherer).

I understand the signal from a spark is broad and basically interference across a range of frequencies, but I was curious if there is one frequency that might (for some reason) be better for transmitting a spark than another.

 

[berkeman]

Your transmitting antenna will likely be the main resonance in the transmission. How long is it? Is it a long vertical wire over a ground plane?

 

[anorlunda]

Your SDR should be an excellent tool to see the frequency spectrum put out by the spark. You can watch it change in real time as you vary parameters such as antenna length and position.

Anecdotally, I got new hearing aids this week. They work fine. However, I found that the piezo sparking igniter in my gas stove drives the aids crazy. It sounds like a building fire alarm in my ears when it is sparking.

 

[Baluncore]

I understand the signal from a spark is broad and basically interference across a range of frequencies, but I was curious if there is one frequency that might (for some reason) be better for transmitting a spark than another.

The frequency will be determined by the wavelength of the dipole and coupling network.
The spark is an impulse that initiates ringing of the dipole.
How do you couple the spark to the dipole?
A lower frequency from a longer dipole will usually transfer more energy.
You will need to experiment.

 

[mishima]

Your transmitting antenna will likely be the main resonance in the transmission. How long is it? Is it a long vertical wire over a ground plane?

The transmitting antenna is also just a (rather crude) dipole from materials I had on hand (20 gauge copper wire and wooden dowel split at the center). Its 1 meter in total (two 0.5 m elements) so 149.896 MHz.

The frequency will be determined by the wavelength of the dipole and coupling network.
The spark is an impulse that initiates ringing of the dipole.
A lower frequency from a longer dipole will usually transfer more energy.
You will need to experiment.

Interesting, I do have some VLF antennas (~25kHz) on hand...although they are the magnetic loop type not dipoles.

 

[tech99]

Hi there, I've been fascinated by using simple coherers to pick up the electromagnetic signal from a spark (in imitation of early radio experiments, for high school education). I am using a push button ignitor (piezoelectric crystal) attached to a simple dipole antenna for the transmitter. When I push the ignitor, the coherer (which is 2 or 3 meters away) turns on an LED on a 3V power supply.

I also have a software defined radio (SDR) that can pick up a wide range of frequencies (1kHz to 2 GHz, continuous). I would like to use this in tandem with my coherer, to get a measure of the signal strength simultaneously (in an effort to answer the question, what is the minimum signal strength required to get a response from the coherer).

I understand the signal from a spark is broad and basically interference across a range of frequencies, but I was curious if there is one frequency that might (for some reason) be better for transmitting a spark than another.

I have done a lot of experiments with coherers. They do not work by making the metal filings cling together due to RF. Rather, they are a threshold device which works on DC as well as RF.

There seem to be three modes of operation. If we apply a voltage of say 1 to 3 volts, DC or RF, the filings will suddenly conduct. This is thought to be due to the breakdown of an oxide layer. The effect occurs even at 40GHz. If the supply has a high resistance, the contact tends to break again when the signal ceases, but if it has a low resistance then micro welding occurs and the contact is permanent until it is tapped.

Once a contact has occurred, we can, by tapping the device, place it is a very sensitive condition where it is slightly conducting. The coherer can then act as a square-law device and demodulate radio signals. It appears then the oxide layer allows electron tunnelling, so that voltages below the threshold can produce a small current - rather like a diode. But the difference from a diode is that the device is bi-directional, the current increasing with a square law in both directions.

To obtain demodulation we need to bias the device to one side, so that the curvature produces an unbalanced response. Used in this way it is possible, using a long antenna, to hear off-air MF and HF signals and the response is, from memory, about 6dB below that of a Germanium diode.

To make a coherer you need metals which from an oxide layer, such as copper, iron, nickel or mercury. Carbon will not work. I have used a circuit where the bias potential is applied from a centre tapped potentiometer circuit, so I can increase the voltage from zero in either direction.

Regarding frequency of operation, it is best to use a resonant device such as a dipole so that the energy of ther spark is contained in a small bandwidth. Otherwise much of the energy could lie outside your measurement range. We do not want to be trying to measure the energy in a nanosecond (or shorter) pulse.

 

[mishima]

Thanks for that info, I've had the best result with copper shavings (ordered from a chemical supply) with galvanized 1/4 bolts on either side for contacts. I tried many other things...copper powder, iron fillings, silver powder, aluminum contacts...

 

[tech99]

That is interesting. I have also used Mercury between Copper electrodes. That is what Marconi used for the trans-Atlantic experiment, together with sensitive earphones. I think the operating condition is when one electrode makes solid contact and the other makes gentle contact. A good way to adjust the coherer is to apply a signal from either a mechanical buzzer or a modulated signal generator.

 

[Baluncore]

Thanks for that info, I've had the best result with copper shavings (ordered from a chemical supply) with galvanized 1/4 bolts on either side for contacts.

Note that both copper oxide and zinc oxide are semiconductors.

Copper oxide was used to make metal rectifiers.
https://en.wikipedia.org/wiki/Metal_rectifier#Description

Zinc oxide is used in varistors.
https://en.wikipedia.org/wiki/Varistor#History

 

[tech99]

Although some metal oxides are semi conductors, it appears that the coherer is a bidirectional device. I have taken characteristic curves on several occasions and it is always of the double type, like back-to-back diodes. It seems to be electron tunneling through the thin oxide insulating layer. In order to obtain detector action using earphones (rather than the "threshold mode" using relay) we need to apply a bias of either polarity of about the same voltage at the peak signal. It might be supposed that two diodes are formed in the coherer, one at each electrode, but this seems not to be the case, because we rarely find unidirectional action where just one electrode might be rectifying.
By the way, I made a mistake with the Mercury coherer, which is Iron - Mercury - Iron.

 

[Baluncore]

Although some metal oxides are semi conductors, it appears that the coherer is a bidirectional device.

That is because the coherer is made from losely packed particles with random ± junction orientation. It can be modeled as a back-to-back diode pair.

The coherer ±3 volt bandgap needs to be biassed to gain sensitivity. While one polarity of particle junction is selected to maximise resistance slope change, the other polarity is reverse biassed. The low reverse voltage breakdown for metal particle junctions is protected by the forward conduction voltage of the other half of the particles.

 

[tech99]

I agree it is simulated approximately by a diode pair. However, the action appears to be either open circuit, short circuit or an S-shaped curve with good bi-directional symmetry. This makes me think it is not the chance creation of diode pairs. Also we do not see a stand-off voltage like a semiconductor diode - conduction commences immediately at the smallest voltage. We do not see series diode strings occurring as far as I have noticed. The optimum bias is very small, as I mentioned, equal to the peak signal voltage, and can be adjusted for each signal to obtain best results.
For the case of the Iron-Mercury-Iron coherer, we have a simplified structure with planar electrodes, so there is not the random orientation. We do not see single diode action happening at one electrode only, which I would expect to see if we were getting semiconductor action.

 

[Baluncore]

However, the action appears to be either open circuit, short circuit or an S-shaped curve with good bi-directional symmetry.

You make it sound like a switch with metalic contacts.

Also we do not see a stand-off voltage like a semiconductor diode - conduction commences immediately at the smallest voltage.

Which I interpret as parallel resistive pathways.

We do not see series diode strings occurring as far as I have noticed.

The multiple diode strings are short circuited by the single diode strings.

 

[mishima]

I think its amazing this ostensibly simple little device was only around 20 or 30 years (made obsolete by the popularity of AM radio) and remains to be experimentally interesting. From a wireless educator's standpoint, the simplicity of construction is on part with an electromagnet and totally approachable for kids (more so than say, a crystal radio kit). They love the little finger tap required to decohere and be usable again.

Is it safe to say the theory of how coherers work is still a bit debatable, or only how they are modeled for circuit analysis?

Are any science suppliers offering a commercial coherer? (my usual go-tos for high school do not)

 

[Tom.G]

One on Ebay:
https://www.ebay.com/itm/194514061470?hash=item2d49f1089e:g:jfoAAOSw2Apd-TMt

 

[tech99]

There was a commercial device a few years back for detecting very fast pulses and it looked like a glass fuse. Ideal for lightning strike detection. One of the early investigators, Chandra Bose, used a coherer device at 40 GHz; as far as I know, there is still no other device which can detect large picosecond pulses without risk of burn out.