[steveh2112]

hi everyone, i'm trying build a simple horizontal level sensor using a circular bubble type spirit level and a couple of electrodes inserted into the fluid cavity, to detect the bubble versus the fluid, hence level or not level. (i don't think anyone makes a simple low cost sensor for this, at lease i haven't found one)

the problem is the fluid in there is non conductive so i need to replace it with something that it. i also want a fluid a little more viscous than plane water so it dampens out vibration a bit.

so far washing up liquid seems to do the trick but i don't want it foaming up under intense shaking.

any suggestion for commonly available liquids i might try?

thanks

[chenbeier]

You add salt and glycerine or other higher viscose liquid like poly ethylen glycol.

[Enthalpy]

Accelerometers are level sensors too. Built by semiconductor technology, they are rugged, small, cheap, sensitive, low-noise, and the electronics is often integrated on the chip or in the package. Just check if their zero is stable enough.

Bubble sensor: I saw one with electrodes three decades ago. Its development was already old. That company developed it for its own needs and didn't sell components.

The difficulties facing such a sensor are tightness and corrosion, which appear after years hence are lengthy to observe. My suggestion would be: don't start that development if your company affords only few months work.

As liquid-tight contacts through an insulator, I know metal-ceramic sealing, and nothing else: epoxies are corrosive, and O-rings don't give the electrodes a stable position. But you can find such feedthroughs already made, with a thread to fit at your container: car spark plugs are an example, feedthrough capacitors an other one. One difficulty is that this leaves little choice for the metal, often molybdenum, and such metals don't resist corrosion well. Cover with gold maybe?

I'd largely prefer a capacitive scheme, with the electrodes evaporated on the glass container. No corrosion, no leaks, stable positions. You can reuse an existing level and experiment with aluminium foil, with metallic adhesive tape (exists at 3M), or with conductive glue used as a paint.

In any case, the reading must be AC. DC voltages drift too much with time and corrosion. AC too needs enough electrodes so some inject the current and others measure voltages.

A machine washing liquid could be a start material. They have anti-foaming and anti-corrosion properties. Water-based hydraulic fluids, especially with polyglycol, are an other start. They have wetting, defoaming, anticorrosion agents. Both present a signifiant conductivity. If adding ions for conductivity, not NOT take a chloride not bromide, as these corrode metals. A phosphate would be better, or just a plain base. However, most anticorrosion agents work by creating a stable layer at the metal's surface, and this goes against a stable metal-liquid voltage. One more reason for AC operation.

Reading the bubble's position optically is horribly more complicated, but it's cheap since bubble levels and PC cameras exist already. Add software.

[Enthalpy]

If using the better capacitive scheme, a liquid metal improves the performance (except for the viscosity). Ga-In-Sn alloys are liquid at -19°C
https://en.wikipedia.org/wiki/Galinstan

[steveh2112]

thanks for the suggestions. i just got my MPU6050 so testing to see if it can maintain calibration. if not, i like the idea of a capacative sensor, better than electrical connection.

also thought about an optical sensor using a bubble level, i'll try that too, easy to interface to arduino

[Enthalpy]

Here's a diagram with capacitive coupling to avoid corrosion worries. The electrodes are an the bulb's outer face, for instance evaporated, painted, deposited chemically. 4mm*8mm area on 1mm glass gives up to 1.3pF depending on the bubble's position and also the glass composition. At example 40.68MHz the impedance is -j3.1kohm.

The liquid can be an electrolyte, better than alloys at cold. 10S/m provide roughly 10ohm between sensing and excitation electrodes, so the liquid's position and the resulting capacitance decides the current. Wetting of the glass over the bubble must be checked.

The LC primary resonant circuit gets a strong amplitude because it is tuned to compensate the total 1.3pF electrode capacitance. Depending on detailed values and component losses, it gets the full voltage of the generator when tilted, or even more. If the capacitor is adjusted at 20pF, the 0.72µH primary for 40.68MHz can have D=10mm, short length, and ~8 turns on air. The secondary can be the same, with an additional 50ohm output, unless the signal conditioning is local. The signal is very strong. Damping should result first from the load hence from design choices, then from the electrolyte, finally from the resonant circuits.

The differential transformer rejects interferences and provides zero output at zero tilt. Synchronous detection can reduce the noise further if any necessary and reject interferences. Shielding is necessary to reduce stray emissions and is mandatory at ISM frequencies too.

The primary coil could also be wound directly on the bulb and the excitation injected at the ends. The other design is more flexible.

Marc Schaefer, aka Enthalpy