Selection between a magnetic float level switch and a magnetostrictive level transmitter is decided by one engineering question: do you need a single on/off point or a continuous, calibrated 4-20 mA reading of level and interface?
The float switch is a passive or reed-based point sensor — typically brass stem, NBR float, threaded or flanged mount, IP65, working from -10 °C to 80 °C and 0 to 10 bar in the LM1 series configuration [S6]. The magnetostrictive transmitter is a microprocessor-based field device that resolves the position of a float along a rigid waveguide with sub-millimetre repeatability, output as 4-20 mA with HART or digital protocols, and used for both single-liquid level and oil/water interface measurement in tanks up to several metres tall [S1][S2][S3].
Operating principle and what each device actually outputs
A float switch closes (or opens) a reed contact when the buoyant magnet in the float passes a fixed point on the stem; it has no HART layer, no calibration curve, and no analog output — only a discrete state change [S6]. A magnetostrictive level transmitter works by sending a current pulse down a waveguide, which generates a torsion wave at the float's permanent-magnet position; the return time, measured by the on-board microprocessor, gives absolute position of both the top liquid surface and, when two floats are fitted, the lower interface surface [S1][S3]. The Yokogawa-sourced ISE-Magtech LTM-300FF extends the same family to hygienic and sanitary service classes [S4].
Forced choice: if the I/O card is a digital input and the loop is a single SPDT contact, the level switch is the only option. If the I/O is analog and the DCS expects an engineering-unit PV, the level transmitter — magnetostrictive being one of several technologies — is the only option that also gives you interface measurement on the same probe.
Decision criteria, head-to-head
Float switches and magnetostrictive transmitters overlap on mounting, but diverge on every other axis. The Elettrotec LM1 magnetic-float switch ships with brass stem, NBR float, threaded mount, IP65, -10 to 80 °C process temperature, 0 to 10 bar process pressure, and multi-point switching as the only "configuration" knob [S6]. The ABB LMT-series magnetostrictive transmitters are field-mounted, modular, microprocessor-based, with 4-20 mA + HART, two-float interface capability, and a working envelope that comfortably covers refinery and chemical tank temperatures and pressures well beyond the LM1 limits [S1][S2][S3].
Cost axis: a brass float switch is an order of magnitude cheaper than a magnetostrictive probe plus head transmitter, so any specification that drops a magnetostrictive where a switch would do is a wasted CAPEX line. Accuracy axis: the switch is a point detector with no linearity spec at all — only a switching hysteresis band in millimetres, typically a few mm; the magnetostrictive transmitter is specified with sub-mm repeatability and full 4-20 mA linearity over metres of measurement length [S1][S3]. Functional axis: one switch equals one trip point; one magnetostrictive probe can carry a top-level float and an interface float simultaneously, replacing what would otherwise be two switches and a sight glass [S1][S3].
Where the float switch is the correct specification
Pick a level switch for high-high and low-low alarms, pump cut-out, and overfill prevention where the trip point is fixed at one elevation. The brass-stem LM1 pattern with NBR float, threaded mount, and IP65 housing fits utility water, light oils, and chemical day tanks operating up to 80 °C and 10 bar [S6]. Multi-point stems give you several trip elevations on one probe, which is the natural way to add a high-high alarm above a control-level transmitter on the same tank without a second penetration. The switch is also the right call on any loop that is intrinsically safe in a way that the magnetostrictive head transmitter is not, or where no HART communicator, no DCS analog input, and no maintenance budget for calibration exists — it is a fit-and-forget device.
Hard rule that is not negotiable: the switch is not a continuous level gauge. If anyone in the review meeting asks for "tank level trending", "daily inventory reconciliation" or "interface detection between two liquids", the switch will not satisfy that scope, no matter how many points you stack on the stem.
Where the magnetostrictive transmitter is the correct specification
Pick a magnetostrictive level transmitter whenever the scope calls for a continuous 4-20 mA or HART PV of liquid level, total volume, or interface position. Refinery separator drums, chemical reactor charge tanks, and produced-water / oil skimmers are the canonical applications, because the same probe resolves the top hydrocarbon surface and the lower water interface through two independently tracked floats on the waveguide [S1][S2][S3]. The modular LMT100 / LMT200 family covers short and long tank geometries, and the AT600 extends the offering to hygienic and higher-temperature classes [S1][S2][S3].
For comparison of transmitter technologies on a like-for-like basis — guided-wave radar, capacitance, RF admittance, and ultrasonic — the level transmitter page lays the architecture options side by side. For interface measurement specifically, the capacitance level transmitter and the RF admittance level switch entries cover the competing probe technologies used in the same separator service. When the question is "automatic tank gauging" or "strapping-table inventory" with HART data back to the DCS, the automatic level page covers the system-level integration.
Failure modes and known limitations
Float switches fail in three predictable ways: the float floods and sinks (loss of buoyancy on a damaged NBR float in hydrocarbon service); the reed contact welds closed under inductive kick from a directly-driven relay coil (no interposing relay or arc suppression fitted); and the stem corrodes at the process connection when the specified material — brass in the LM1 example — is ordered into a chemical service it was not rated for [S6]. None of these failure modes shows up on the data sheet as a single MTBF number; they are service-condition failures, so the mitigation is specification discipline, not redundancy.
Magnetostrictive transmitters fail differently: the waveguide can be bent by mechanical impact during installation, which destroys linearity over a section of the probe; the float's permanent magnet can lose field strength at sustained high temperature, shifting the indicated position; and the head transmitter's electronic module is sensitive to surge on the 4-20 mA loop, so loop-powered installations in outdoor tanks need proper surge protection [S1][S2]. The ultrasonic level meter selection gate article makes the same point about non-contact technologies — the failure mode is rarely the sensor element itself, it is the field-side loop and the installation geometry.
Side-by-side comparison for procurement and review
Use this matrix in the spec review when both devices are on the requisition line and the wrong one is being justified by price. Float switch (Elettrotec LM1 reference): magnetic-float point detector; brass stem, NBR float; threaded or flanged mount; IP65; -10 to 80 °C process temperature; 0 to 10 bar process pressure; output = SPDT reed contact; multi-point available; no analog output; no HART; no interface measurement [S6]. Magnetostrictive transmitter (ABB LMT100/LMT200, AT600, Yokogawa ISE-Magtech LTM-300FF reference): continuous 4-20 mA + HART field transmitter; two-float interface capability; microprocessor-based; modular probe lengths; broad process temperature and pressure envelope; output = engineering-unit PV plus dual alarm relays on head transmitter variants [S1][S2][S3][S4].
Translated into the four criteria a project engineer is actually graded on — accuracy, interface capability, installed cost, and commissioning time — the switch wins only on cost and commissioning speed; the transmitter wins on accuracy, interface capability, and the ability to be re-ranged in the field through HART without pulling the probe. Everything else, including the ultrasonic level meter versus level switch trade-off case laid out separately, follows the same logic: binary output versus continuous output is the dividing line, and neither technology is a substitute for the other on the wrong side of that line.
Trackable next nodes: confirm whether the tank in question is single-liquid or two-liquid — that single question alone moves the spec from LM1-class to LMT200-class; and confirm whether the DCS has a free analog input, because that gate decides whether the magnetostrictive level transmitter can even be wired in without a new I/O card. If both answers come back "yes", specify the magnetostrictive; if either answer is "no", specify the level switch and stop.