Inductance of a Loop in Air and Water?

[KerimF]

The inductance of a loop (say, 1 m diameter) of enameled wire is L_air at 1 Mhz.
Its inductance is L_water when the loop is immersed in fresh water, also at 1 Mhz.
Which is, in your opinion, the true case?
L_water < L_air
L_water = L_air
L_water > L_air

Thank you.
Kerim

[TimFox]

What is the conductivity of the "water"?
Absolutely pure water has very low conductivity, regular water is mediocre, and saline solutions have high conductivity.

[KerimF]

What is the conductivity of the "water"?
Absolutely pure water has very low conductivity, regular water is mediocre, and saline solutions have high conductivity.

You are right about water conductivity.

I am wondering about inductance (in Henry).
Naturally, if no one knows for sure the answer, I have to do an experiment using a small practical loop (not necessarily just one turn) to know it.

[Vovk_Z]

It depends (pure, non-pure, frequency, etc).
Article

[ejeffrey]

Water is not magnetic so the loop inductance won't change.  Capacitance and dissipation will increase, so this will have an effect on the overall impedance.  This may affect your measurement of the "inductance" especially if you are anywhere near the self resonant frequency.

The way to think about how conductivity and loss affect the circuit, consider a transformer. If you measure with the secondary open circuit you measure the primary inductance.  If you add a load resistor, the impedance seen at the primary drops but the primary inductance itself remains the same.

[TimFox]

Conductivity surrounding a coil can be considered a "shorted turn" with loss coupled to the inductor.
Water is also slightly diamagnetic (q.v.), with a relative permeability < 1.

[KerimF]

Conductivity surrounding a coil can be considered a "shorted turn" with loss coupled to the inductor.
Water is also slightly diamagnetic (q.v.), with a relative permeability < 1.

I understand that water introduces sort of RF losses when the loop is energized by a 1 Mhz signal.
Do you think an enameled wire (used in transformers), as mentioned in OP, will also let the loop be shorted by the water conductivity? Thank you.
 

[MrAl]

The inductance of a loop (say, 1 m diameter) of enameled wire is L_air at 1 Mhz.
Its inductance is L_water when the loop is immersed in fresh water, also at 1 Mhz.
Which is, in your opinion, the true case?
L_water < L_air
L_water = L_air
L_water > L_air

Thank you.
Kerim

Hello,

Do you have a specific application for this?
I ask because that often matters.

[TimFox]

Conductivity surrounding a coil can be considered a "shorted turn" with loss coupled to the inductor.
Water is also slightly diamagnetic (q.v.), with a relative permeability < 1.

I understand that water introduces sort of RF losses when the loop is energized by a 1 Mhz signal.
Do you think an enameled wire (used in transformers), as mentioned in OP, will also let the loop be shorted by the water conductivity? Thank you.

Yes, just as using enameled wire in a transformer primary still works with a shorted external turn.

[KerimF]

Hello,

Do you have a specific application for this?
I ask because that often matters.

Very good question.
Lately, I had somehow a silly idea to test. I liked to monitor the water height linearly (not just at certain levels) in my reservoir/tank. But where I live, I can't get any special component which suits this job (pressure sensors, ultrasonic transceivers... etc.).

First, I had the thought of immersing two long electrode to sense the gradual variation of their capacitance while the water height varies. I couldn't find a practical way to do it.

The alternative, I thought of, since it will not be a NASA project, was to use a narrow long closed loop of enameled wire instead. It will play the L of an LC oscillator (around 400 kHz, I am not sure). Its construction is simpler than of the two previous electrodes. I supposed that L, hence the frequency, will change gradually while the water height rises. This may be wrong. But I have the impression that the frequency will change though not necessarily because of L inductance. The LC resonant tank will likely be disturbed by the change of the electromagnetic wave speed (water/air) in its two parts (below and above the water surface).

Even if the test will let me realize that the alternative solution is also not practical, I am curious to know what I can get from the measured data at different frequencies.
 
In case it will be practical, the function, height versus frequency, will be adjusted in software with the knowledge it is sensitive to temperature and the water characteristic (I guess even the drinkable water of the city network is likely not the same always). And I am afraid that the function's algorithm has to be somehow intelligent to adjust itself by learning (mainly the frequencies of the upper and lower limits). After all, the accuracy is not too important. The most important is to know if water is filling the tank (when a pump is activated) or the tank loses water because of a fault, before it is too late.

[TimFox]

Will the detailed composition of the water solution be absolutely constant during your measurements?
If the result depends only on diamagnetism (i.e., ultra-pure water), this might work.
If affected by conductivity, that varies all over the map with the solutes and concentration.

[KerimF]

Will the detailed composition of the water solution be absolutely constant during your measurements?
If the result depends only on diamagnetism (i.e., ultra-pure water), this might work.
If affected by conductivity, that varies all over the map with the solutes and concentration.

You have a good point to test. Thank you.
I guess, my test will have to be done for many days and at different temperature (in these days, we have warm daylights and cold nights).

[ejeffrey]

The relative magnetic permeability of water is 0.99999 (where vacuum/air is 1.0).  The dielectric constant is 80.  Even assuming ultra-pure deionized water with no conductive losses, you are trying to measure and effect 10 million times smaller than the capacitance change that "didn't work."  If you build the circuit you describe it might work, but it will be entirely due to the dielectic constant or conductivie losses.  Measuring the RF permeability of water is an extremely difficult measurement because of this.

[IanB]

First, I had the thought of immersing two long electrode to sense the gradual variation of their capacitance while the water height varies. I couldn't find a practical way to do it.

The alternative, I thought of, since it will not be a NASA project, was to use a narrow long closed loop of enameled wire instead. It will play the L of an LC oscillator (around 400 kHz, I am not sure). Its construction is simpler than of the two previous electrodes.

The only functional difference between a long narrow loop of enameled wire and two parallel lengths of enameled wire is that the former (loop) is closed at the bottom, and the latter (parallel wires) is open at the bottom.

I cannot conceive of a reason why one is more practical to achieve than the other.

Given that the capacitive approach is much more sensitive to water level, it would be a much more likely candidate to use.

[KerimF]

The only functional difference between a long narrow loop of enameled wire and two parallel lengths of enameled wire is that the former (loop) is closed at the bottom, and the latter (parallel wires) is open at the bottom.

To get a practical capacitance, the spacing between the two wires of the latter needs to be small.
It is relatively wide in case of the former

Given that the capacitive approach is much more sensitive to water level, it would be a much more likely candidate to use.

This is what I like to find out from the test of the inductance approach.

[KerimF]

If you build the circuit, you describe, it might work, but it will be entirely due to the dielectric constant or conductivity losses.

Very true.
And I can't be sure now if it will be a practical approach to monitor the water height.

[IanB]

To get a practical capacitance, the spacing between the two wires of the latter needs to be small.

Given that commercial capacitive level sensors exist, and are small and practical, I would suppose that it is probable such a device could be made by the home experimenter.

[TimFox]

Capacitive liquid sensors are common for petroleum products and other non-conducting media.

[gcewing]

Insulated metal strips rather than wires would give you more capacitance to work with.

[KerimF]

Given that commercial capacitive level sensors exist, and are small and practical, I would suppose that it is probable such a device could be made by the home experimenter.

Capacitive liquid sensors are common for petroleum products and other non-conducting media.

After reading your comments, I will likely use the capacity approach.

I started to think of a simple 'temperature to frequency' circuit (to compensate the variation of the water dielectric coefficient with temperature). The 'capacitance to frequency circuit' is relatively easy to build.

And I will have to find out how to send the two frequencies down to the MCU of the in-door meter via a pair of wire (which will also power the sensor circuit at the water tank).

[golden_labels]

A capacitive sensor as suggested above is the response to your question.

But the original post was more general and perhaps somebody else may come across it later. So here’s a neat experiment from Alpha Phoenix. He compared the signal propagation speed in a wire submerged in air and in water. The difference is probably too small for your purposes and the entire thing is the sub-microsecond realm. Nonetheless a good demonstration fitting the topic.

[MrAl]

Hello,

Do you have a specific application for this?
I ask because that often matters.

Very good question.
Lately, I had somehow a silly idea to test. I liked to monitor the water height linearly (not just at certain levels) in my reservoir/tank. But where I live, I can't get any special component which suits this job (pressure sensors, ultrasonic transceivers... etc.).

First, I had the thought of immersing two long electrode to sense the gradual variation of their capacitance while the water height varies. I couldn't find a practical way to do it.

The alternative, I thought of, since it will not be a NASA project, was to use a narrow long closed loop of enameled wire instead. It will play the L of an LC oscillator (around 400 kHz, I am not sure). Its construction is simpler than of the two previous electrodes. I supposed that L, hence the frequency, will change gradually while the water height rises. This may be wrong. But I have the impression that the frequency will change though not necessarily because of L inductance. The LC resonant tank will likely be disturbed by the change of the electromagnetic wave speed (water/air) in its two parts (below and above the water surface).

Even if the test will let me realize that the alternative solution is also not practical, I am curious to know what I can get from the measured data at different frequencies.
 
In case it will be practical, the function, height versus frequency, will be adjusted in software with the knowledge it is sensitive to temperature and the water characteristic (I guess even the drinkable water of the city network is likely not the same always). And I am afraid that the function's algorithm has to be somehow intelligent to adjust itself by learning (mainly the frequencies of the upper and lower limits). After all, the accuracy is not too important. The most important is to know if water is filling the tank (when a pump is activated) or the tank loses water because of a fault, before it is too late.

Ok, so how deep is the water normally and how deep can it get as a maximum?

The inductance idea probably isn't a good one because it's too hard to measure these quantities.  A better idea is to force the situation.
Maybe you can work this next idea out. One way is to measure the temperature of the water and use that as a reference.  Then have the other temperature sensors be heated to something above the reference temperature.  When the water touches one of the heated sensors, it's temperature will be reduced, and in that way you know the water is up to that level so far.  If it touches the next sensor up, it reduces the temperature of that sensor also so you know it's even higher now.
The reference temperature is needed so you know how much to heat the sensors.  Once a sensor is under water, it will stay cool until the water level falls again and then it will heat up again.

This idea assumes the at least the following:
1.  There is enough water so that the needed power delivered to the sensors does not heat the water too much, but if it does heat a little then it's not enough to cause an extreme temperature rise.
2.  The water is allowed to vary in temperature naturally, but the sensor logic has to allow enough time between readings so none of them are falsely reported as submerged.
3.  The needed water level resolution does not have to be too small so that you don't need 100 sensors, unless you care to do that.
4.  Resistors can be used along with some self heating, but you'd have to work out the details for how much to heat and a decent way to measure the resistance changes (like with a microcontroller).  If not, resistors can be used as the heat elements along with good temperature sensors.  Thermistors may work also, with a little forced self heating.

Something like this has to be fully tested of course in the same environment that it will actually be used in or at least a reasonable facsimile.  Be aware of all possible changes that can occur naturally in the environment.