In a separator or a treater there are really two levels that matter: the total height of liquid in the vessel and the boundary partway down where oil floats on top of water. A dual-float magnetostrictive transmitter reports both from a single probe by putting two floats of different buoyancy on the same waveguide, one that rides the top surface and one that settles at the oil-water interface. This page focuses on that interface job specifically, which the generic magnetostrictive device page mentions only in passing: how one waveguide handles two floats, why the floats' densities are chosen the way they are, what happens when the layer between them is an emulsion rather than a clean line, and the separate SCADA tags the probe produces for total level and interface.
Dual-float magnetostrictive interface in one line: A dual-float magnetostrictive interface transmitter uses a single magnetostrictive waveguide carrying two floats of different density, so one probe measures two positions at once: the total liquid level and the oil-water interface. The lighter float is buoyant enough to ride the top surface of the upper liquid, while the heavier float is tuned to sink through the light phase and float at the boundary between the two liquids. The transmitter times the torsional return from each float separately, so it outputs both a total-level and an interface-level reading, which in a separator or treater feed SCADA as two distinct tags for controlling the top level and the interface independently.
A magnetostrictive probe locates a float by timing a torsional wave launched where the float's magnet sits, and nothing about that mechanism restricts it to a single float. Put two floats with magnets on the same waveguide and each one launches its own return wave when the interrogation pulse passes it, so the transmitter receives two distinct return signals per cycle, one from each float, at times corresponding to their two positions. The electronics separate the two returns and report two positions, which is how one probe measures two things at once without any extra wiring or a second instrument.
The arrangement in a vessel is that both floats live on the same rigid probe, usually inside a chamber or a still-pipe so they move cleanly. The upper float rides the very top of the liquid and its position is the total level in the vessel. The lower float sits down at the interface between the two liquid phases and its position is the interface level. Because both are read off the same waveguide with the same timing principle, the two readings share the same precision and the same reference, which keeps the total level and the interface consistent with each other rather than coming from two separately calibrated devices.
The reason this matters in the field is that separators and treaters are exactly where you need both numbers and where fitting two independent instruments is awkward. One probe gives the total level for controlling the vessel's overall inventory and the interface for controlling how much water sits under the oil, from a single nozzle and a single transmitter. That economy of one device for two measurements, at custody-grade resolution, is the main reason the dual-float magnetostrictive arrangement is chosen for interface work over separate level and interface instruments.
The whole trick to interface measurement is float density, because a float only sits where its own density places it. A float floats at the level where the fluid it displaces balances its weight, so a float slightly less dense than a liquid rides on top of that liquid, and a float between the densities of two liquids sinks through the lighter one and floats on the heavier one at their boundary. The two floats on the probe are deliberately given different densities so they park at the two surfaces that matter.
For the top level, the upper float is made light enough to be buoyant in the upper liquid, so it rides the top surface wherever it is. For the interface, the lower float is the critical one: its density has to fall between the density of the light phase above and the heavy phase below, so that it is too dense to float in the oil but light enough to float on the water and therefore comes to rest exactly at the oil-water boundary. If that float's density were outside the window between the two liquids, it would either sink to the bottom or float to the top and never find the interface, so selecting it against the actual phase densities of the process is essential to getting a valid interface reading.
This density dependence is also why an emulsion layer complicates interface measurement. Real separators often have a rag or emulsion layer at the boundary rather than a sharp line, a zone where oil and water are mixed and the density changes gradually rather than stepping cleanly. The float settles somewhere within that graded density band, so what the probe reports is a position within the emulsion rather than a razor-sharp interface, and if the emulsion thickens or its density shifts the reported interface moves with it. Operators account for this by choosing the float density for where in the emulsion they want the reading to sit and by reading the interface tag as the top of the water region rather than an idealised clean line.
Because one probe produces two positions, it publishes two distinct signals, and keeping them straight is the whole point of the device in a control system. The transmitter outputs a total-level tag from the upper float and an interface-level tag from the lower float, and these drive two different jobs in a separator or treater. The total level feeds the overall liquid inventory control, typically throttling the liquid or oil outlet to keep the vessel from overfilling or emptying, while the interface level feeds the water-draw control, opening and closing the water outlet to keep the water layer under the oil at the right height so that neither oil escapes with the water nor water carries over with the oil.
A cloud SCADA platform such as Merobix takes both tags and trends them together, which is exactly how interface behaviour is best understood. Watching the interface tag alongside the total level and the inlet and outlet flows shows whether the water leg is being drawn at the right rate, and a slowly rising interface or a thickening gap between the two floats points to an emulsion building or a water-draw that cannot keep up. Alarms on the interface climbing toward the oil outlet, or on the total level running high, warn operators before the separator carries product into the wrong stream.
For separators and treaters at remote or unmanned sites, surfacing both the total-level and interface tags through cloud SCADA is what makes the dual-float probe's two-in-one measurement actually useful. On-call staff can see and control the vessel's overall level and its internal split from the same trend, distinguish a real interface excursion from noise by looking at both floats together, and catch a slowly worsening emulsion that a single-level reading would miss entirely. The device earns its place by giving two custody-grade positions from one nozzle; carrying both of those positions cleanly into remote monitoring is how that advantage reaches the operators who act on it.
The probe carries two floats on the same magnetostrictive waveguide, each with its own magnet, so when the interrogation pulse passes them each launches a separate torsional return wave. The transmitter times the two returns independently and reports two positions, the upper float giving the total liquid level and the lower float giving the oil-water interface. Because both are read off the same waveguide with the same timing principle, the two measurements share the same precision and reference from a single instrument.
The interface float must have a density that falls between the density of the light phase above and the heavy phase below, so it is too dense to float in the oil but light enough to float on the water and therefore settles at the boundary between them. If its density were outside that window it would either sink to the bottom or ride the top and never find the interface. So the float is selected against the actual phase densities of the process, and in an emulsion its density determines where within the graded rag layer it comes to rest.
A real separator often has a rag or emulsion layer at the boundary where oil and water are mixed and the density changes gradually rather than in a sharp step. The interface float settles somewhere within that graded density band, so the probe reports a position inside the emulsion rather than a clean line, and if the emulsion thickens or its density shifts the reported interface moves with it. Operators choose the float density for where in the emulsion they want the reading and treat the interface tag as the top of the water region rather than an idealised interface.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.