A magnetic drive pump — the sealless centrifugal design in which an inner impeller magnet set is driven across a containment shell by an outer driven magnet assembly — is the default choice when the process fluid is toxic, flammable, or expensive enough that a mechanical seal failure is unacceptable [S2].
Selection in real engineering work is not "pick a brand"; it is gating a duty point, a fluid, a temperature, a containment material, and a torque budget against a published curve, with magnetic-drive pumps generally capped well below the head and flow envelope of equivalent seal-supported ANSI process pumps. The gate sequence matters: a pump that passes chemical compatibility can still fail on dry-run survivability, and a pump that passes both can still de-couple under upset viscosity.
Duty Point: Q, H, NPSH and the Magnetic Coupling's Torque Envelope
The first gate is a hard intersection of the system curve against the published pump curve, expressed in m³/h and metres of head, because magnetic-drive units are built on a standard centrifugal hydraulic and inherit the affinity laws of any end-suction ANSI pump [S1]. A representative small-frame chemical unit such as the Fluimac CM 30 lists flow up to 35 m³/h (1,236 ft³/h) and head to 24 m, with a fluid-temperature window of -10 °C to +90 °C and motor power 2.2–3.0 kW at 2 in / 1.5 in flanged connections [S1].
Where magnetic-drive units diverge from sealed pumps is the coupling envelope. Torque transmits from drive magnet to driven magnet across the can, and every OEM publishes a maximum continuous torque — the slip point at which the inner magnet set decouples and the impeller stalls while the motor keeps spinning. The practical rule: never size the duty point above ~80–90 % of the published decoupling torque at the operating speed, because cold-fluid viscosity spikes, vapour at low NPSH, and a partially closed discharge valve all transient-load the coupling. A duty point sitting at the Best Efficiency Point is therefore a risk-mitigation choice as much as an efficiency choice.
Containment Shell Material: the Hermetic Boundary That Replaces the Seal
The can that separates the wet end from the drive magnet is the entire reason the pump exists. Specify it from the process fluid: for chemical-industry service, PP or PVDF magnetic-drive pumps are available with fluid temperatures up to ~90 °C, while stainless-steel and alloy magnetic-drive pump constructions are also offered for chemical applications. The shell also sets the real temperature ceiling: bonded neodymium-iron-boron magnets in the inner rotor typically derate above ~120 °C and become non-viable above ~150 °C, so high-temperature service often forces a samarium-cobalt rotor or a different pump architecture entirely [S3].
Containment-shell pressure rating deserves the same attention as the flange rating. The can is a thin-walled pressure boundary welded or moulded into the casing, and its burst rating is what stands between a 10-bar reactor and the operator. Cross-check the can rating against the maximum possible system pressure under blocked-discharge or thermal-expansion transients — not just the normal duty point.
Temperature, Dry Run and the "Hidden" Failure Modes
Magnetic-drive pumps tolerate some abuse that sealed pumps do not, and they fail in some modes that sealed pumps do not. The unit's wet-end bearings — usually silicon carbide or alumina — rely on the pumped fluid for lubrication and cooling, so dry running destroys them in minutes, not hours, and a flooded suction, an open suction valve, or a vapour-locked start sequence is enough to do the damage [S2]. A bare pump datasheet rarely states dry-run survivability; treat anything described as "limited" or "with monitoring" as zero and add a flow switch, a conductivity probe, or a power monitor in the control logic.
Temperature discipline matters twice: at the wet end, where the pumped fluid sets the shell/magnet limit, and at the drive end, where the driven magnet assembly dissipates slip energy as heat into the air gap. Above ~80–90 °C fluid temperature, published horsepower derates begin, and the heat rejection from the rear housing needs to be confirmed against ambient and enclosure conditions — a hot skid in a Middle-East summer is a different thermal problem from the same pump in a temperate plant. Magnatex-class designs documented for refinery duty in Houston are built explicitly to handle the higher end of this envelope with stainless and alloy containment, ISO 9001-2008 manufacture, and both ANSI and API 685 mountings [S3].
Selection Criteria: A Side-by-Side Map of the Main Options
Process engineers in practice choose between three architectures — close-coupled plastic (PP/PVDF), close-coupled metal (316/Hastelloy/titanium), and long-coupled metal — and the decision is driven by the four gates above. The comparison below lines the typical options up against the criteria that actually fail audits and warranty claims. [S1]
On fluid compatibility, plastic-frame units win for HCl, NaOCl, and hot chromic acid; metal units win for hydrocarbons, ketones, and anything above ~90 °C. On the torque envelope, larger frame size, heavier magnets, and a tighter can wall all raise the decoupling margin. On dry-run survivability, no architecture is intrinsically good; it is a control-instrumentation problem, not a pump problem. The product specification page for magnetic materials explains the NdFeB-versus-SmCo temperature trade-off in detail, and the broader magnetic drive pump reference maps the hydraulic, material and standard families that those choices imply.
Standards, Manufacturer Map and Sourcing Reality
The standards stack is well defined: ISO 5199 / ISO 2858 covers the dimensional envelope; API 685 covers the sealless magnetic-drive centrifugal for refinery and petrochemical service, including containment-shell burst, decoupling-torque, and secondary-containment provisions; ATEX 2014/34/EU and IEC 60079-x govern the explosion-protected motor when the pump sits in a classified area; and ASME B16.5 / B16.34 govern the flange and body pressure-temperature ratings. Where the OEM is also the magnetic-coupling designer (Magnatex, Sundyne, Flowserve, Iwaki, Richter, CP&P, Hermetic), published curves can be trusted against the standard; where the coupling is sourced from a third party, the integrator's curve is the one to follow [S3].
Sourcing reality in 2026: rare-earth oxide prices climbed into July as magnet-grade supply tightened, which is pushing the bill of materials on NdFeB rotor assemblies upward and lengthening lead times for high-grade couplings (rare-earth supply update, July 2026). For spec writers, that means locking rotor material and grade at the enquiry stage and not allowing alternates at PO, because an SmCo substitute is not a drop-in for NdFeB. The same cost discipline applies to the multistage-curve work in adjacent services: the multistage centrifugal pump price 2026 breakdown shows the same raw-material and magnet-grade levers moving, just on a different frame.
Field Use, Acceptance Tests and What Spec Writers Should Demand on the Datasheet
In the field, a properly specified magnetic-drive pump will be quieter than its sealed equivalent, will not have a seal-pot or quench line, and will have a noticeably shorter stub-shaft footprint. Acceptance should be the standard hydraulic run at 1.0 × BEP flow, an extended run at 0.3 × BEP to confirm the coupling does not slip at low-flow high-head conditions, a thermal-soak run to confirm rear-housing temperature stabilises inside the magnet limit, and a helium leak test on the containment can at 1.5 × design pressure to confirm hermetic integrity before the line is energised with process fluid. The OEM certificate should be filed against the tag number, not the PO number, so the same data travels with the asset for the rest of its service life [S3].
Closing reference: a tighter spec on the front end — duty point against torque envelope, containment material against fluid chemistry, temperature against magnet grade, dry-run protection against the control system — eliminates most warranty disputes and most of the silent over-spec that drives rotor-cost up by ~15–25 % on this class of equipment. Watch the rare-earth oxide price feed into July for further rotor-cost signal, and recheck the decoupling-torque curve at every six-month service interval because the slip point drifts as the can wall thins under erosion. A useful adjacent read on the spec-gates pattern is the rotary encoder selection checklist, which applies the same "gate before brand" logic to a different component family.
For component-level specifications, see magnetic sensor.