Rare Earth Permanent Magnet Motors: Reading the Magnet Grade
Quick Answer
A rare earth permanent magnet motor is a specification set built around a magnet grade, and the grade does more to define it than the product name does. Two families matter: neodymium-iron-boron, the strongest, and samarium-cobalt, the heat-tolerant one. Inside NdFeB the suffix is the real decision, and what it buys is coercivity from dysprosium or terbium, paid for with a little remanence. Two motors carrying the same 42 number can differ by 120 degrees of working temperature. At the hot end that difference costs about 13 percent of flux, 15 percent more current and 33 percent more copper loss at the same torque, before the winding resistance rise is counted.
- Only four rare earth elements reach the magnet
- NdFeB and SmCo: strength traded against heat
- A grade carries two pieces of information
- The arithmetic of a hot magnet
- The worst case is a short circuit, not an overload
- How the magnet changes the rotor
- Auditing an efficiency claim
- What to write on the specification sheet
- Where it wins, and where it does not
- Six mistakes that reach production
- Troubleshooting: symptom, cause, action
- How Greensky Power specifies magnets
- References
- FAQ
Only Four Rare Earth Elements Reach the Magnet
The rare earth family contains seventeen elements: the fifteen lanthanides plus scandium and yttrium. Almost none of them end up in a motor. Commercial permanent magnets use neodymium and praseodymium as the main rare earth constituent, samarium for the cobalt family, and dysprosium and terbium in small amounts to raise coercivity. Everything else in the group is either a by-product of separation or belongs to a different industry.
This matters commercially, because “rare earth motor” gets used as though it described a single material. It covers at least two families with different physics, and within each family a dozen grades with different thermal limits. A buyer who specifies “rare earth magnet” and nothing else has specified almost nothing.
| Element | Role in the magnet | Practical note |
|---|---|---|
| Neodymium (Nd) | Main rare earth in NdFeB | Sets the remanence; usually supplied as didymium with praseodymium |
| Praseodymium (Pr) | Substitutes for part of the neodymium | Slightly raises coercivity and corrosion resistance |
| Dysprosium (Dy) | Raises intrinsic coercivity at high temperature | Used sparingly; each addition lowers remanence a little |
| Terbium (Tb) | Same role as dysprosium, stronger per unit mass | Concentrated at grain boundaries in modern processes |
| Samarium (Sm) | Main rare earth in SmCo | Buys temperature capability instead of peak strength |
Why the 4f shell matters
The 4f electron shell sits deep enough that its magnetic moment is largely shielded from the crystal field, and that produces very large magnetocrystalline anisotropy. The material therefore resists being magnetised along any axis but its preferred one, and resists being pushed out of that state by an opposing field. That resistance is coercivity, and it is what lets a thin magnet in a small air gap hold a large flux.
Ferrite has the same job at roughly a third of the flux density. Sintered ferrite reaches about 0.38 to 0.44 tesla of remanence, sintered NdFeB about 1.1 to 1.45 tesla. For a fixed air gap and stator, that ratio is the difference between a motor that needs a large frame and one that fits in a hand. The comparison in more depth, including where the induction machine still wins, is in permanent magnet motor versus induction motor.
NdFeB and SmCo: Strength Traded Against Heat
The two families are not better and worse versions of each other. They are two answers to the same question. What settles the choice is the temperature the magnet has to survive; the torque requirement barely enters into it.
| Property | Sintered NdFeB | Sintered SmCo (2:17) |
|---|---|---|
| Remanence Br | 1.1 to 1.45 T | 0.85 to 1.15 T |
| Maximum energy product | 200 to 420 kJ/m3 | 130 to 260 kJ/m3 |
| Temperature coefficient of Br | about -0.11 to -0.13 %/K | about -0.03 to -0.04 %/K |
| Curie temperature | about 310 to 340 °C | above 700 °C |
| Corrosion behaviour | Poor; coating is mandatory | Good; coating usually unnecessary |
| Mechanical strength | Brittle, crack-sensitive | Brittle but tougher in tension |
| Relative material cost | Lower | Higher |
| Typical use | Traction, servo, pumps, appliances | Aerospace, high ambient, vacuum, safety-critical |
When samarium-cobalt is the right answer
A 120 K rise settles the comparison. Neodymium loses about 13 percent of its remanence across that span, samarium-cobalt under 4 percent. Add a Curie temperature more than twice as far above the working range, no corrosion coating to specify, and a rotor that tolerates a hot soak without recalculation, and the higher material cost stops being the deciding number.
In practice the environment forces the decision more often than the designer’s preference does. A pump in a hot process line, an actuator inside a vacuum chamber where convection is absent, a motor that has to keep working after a fire event, a downhole tool: all of these push the decision to SmCo regardless of what the torque requirement says. For everything else the answer is NdFeB, and the real work moves to choosing the grade.
A Grade Carries Two Pieces of Information
The standard designation packs an energy product and a temperature class into one short string. In N42SH the 42 is the maximum energy product in megagauss-oersted, roughly 334 kJ per cubic metre, and the letters SH are the temperature class, which carries a minimum intrinsic coercivity and a maximum operating temperature with it.
| Suffix | Typical maximum operating temperature | Minimum intrinsic coercivity | Intrinsic coercivity in SI |
|---|---|---|---|
| N (none) | 80 °C | 12 kOe | 955 kA/m |
| M | 100 °C | 14 kOe | 1 114 kA/m |
| H | 120 °C | 17 kOe | 1 353 kA/m |
| SH | 150 °C | 20 kOe | 1 592 kA/m |
| UH | 180 °C | 25 kOe | 1 989 kA/m |
| EH | 200 °C | 30 kOe | 2 387 kA/m |
| AH | 230 °C | 35 kOe | 2 785 kA/m |
The suffix is not a strength rating
The most expensive misunderstanding in this field is that a higher suffix means a better magnet. It means a more heat-resistant one, and it is bought at a price in both directions. Dysprosium and terbium raise coercivity but couple antiferromagnetically with the iron sublattice and pull remanence down, so 35UH has a slightly lower Br than plain N35 even though both are nominally 35-grade.
The cost is real as well: dysprosium and terbium are the two least abundant of the four elements in use. A design that reaches for UH because the application feels important, when a smaller temperature rise or a better cooling path would have kept the magnet at 110 °C, has paid twice, once in heavy rare earth and once in the flux it gave up. The minimum values quoted for each material class are standardised in IEC 60404-8-1, which is the document a grade claim should be checked against [1].
The same substitution logic runs in the opposite direction for cost-sensitive grades, where part of the neodymium is replaced by the far more abundant lanthanum or cerium. Remanence falls and the grade number falls with it, but the price falls further. A buyer comparing two quotations for the same nominal duty should ask which direction the supplier has moved, because the answer is not visible on the rotor.
Maximum operating temperature is not a material constant
The figures in the table assume a typical magnetic circuit. Resistance to demagnetisation depends on the permeance of the circuit the magnet sits in, not only on the material, so the same grade tolerates more heat in a thick, well-closed circuit and less in a thin, open one. A supplier quoting 180 °C for UH without stating the permeance coefficient has quoted a rule of thumb, not a guarantee.
That is why the demagnetisation analysis has to be run on the actual rotor geometry at the actual worst-case temperature, as set out further down. Two rotors using identical magnets can carry different safe operating limits, because one has a larger leakage path or a thinner magnet.
The Arithmetic of a Hot Magnet
Temperature enters a permanent magnet motor twice. The magnet loses flux and the copper gains resistance. Both move the same way, and the combined penalty is much larger than either effect on its own.
For a three-phase machine with p pole pairs and a peak flux linkage per phase of lambda, the torque per amp is 1.5 times p times lambda. In a four-pole motor p is 2, and a flux linkage of 0.159 weber at 20 °C gives a torque constant of 0.477 newton-metre per amp, so 10 A produces 4.77 newton-metres. The table below carries that machine up to 140 °C and beyond. The derivation of that torque expression is set out in how to calculate motor torque.
| Magnet temperature | Relative flux | Torque constant (N·m/A) | Current for 4.77 N·m | Copper loss, relative |
|---|---|---|---|---|
| 20 °C | 100.0 % | 0.477 | 10.00 A | 1.000 |
| 80 °C | 93.4 % | 0.446 | 10.71 A | 1.146 |
| 120 °C | 89.0 % | 0.425 | 11.24 A | 1.262 |
| 140 °C | 86.8 % | 0.414 | 11.52 A | 1.327 |
| 180 °C | 82.4 % | 0.393 | 12.14 A | 1.473 |
| 200 °C | 80.2 % | 0.383 | 12.47 A | 1.555 |
Two temperature effects that compound
The table stops at the magnet, but the winding runs at roughly the same temperature, and copper resistivity rises by 0.393 percent per kelvin, so a winding at 140 °C has about 47 percent more resistance than the same winding at 20 °C. Combining the two, 1.327 for the extra current and 1.472 for the resistance, gives 1.95. The same torque at 140 °C costs nearly twice the copper loss it cost when cold.
That factor is why the magnet temperature belongs in the design loop. It also explains a failure that looks like a sizing error and is not: a motor that performs correctly on a cold bench and overheats after an hour in a hot enclosure was probably specified at 20 °C and is running at 130. The loop closes on itself: the extra copper loss heats the winding, the winding heats the rotor, and the rotor heats the magnet.
Always state the temperature at which a torque constant is quoted, because a bare figure in newton-metres per amp cannot be checked. Where the application is hot, the cheap fix is usually on the thermal side: more cooling surface, a lower-loss stator or a shorter duty cycle will often let a cheaper grade survive where a higher suffix was about to be specified.
The Worst Case Is a Short Circuit, Not an Overload
Reversible loss recovers when the magnet cools. Irreversible loss does not, because the operating point has fallen below the knee of the second-quadrant demagnetisation curve and part of the material has flipped. The design question is where that knee sits at the highest temperature the magnet will see, and how much field it takes to reach it.
Engineers tend to test this by imagining an overload. In most machines the governing case is a terminal short circuit at high temperature. The stator current is limited only by the machine impedance, the field it produces directly opposes the magnets, and because the short removes the terminal voltage there is no counter-flux from the supply to help. Field weakening beyond the inverter’s capability produces the same result from the other direction.
Do not assume the winding thermistor protects the magnet. A PTC or PT100 embedded in the winding measures the winding. The magnet is a separate thermal mass with its own time constant, sitting in a rotor that may be cooled only by radiation and a thin air gap. In a short-circuit event the magnet can pass its knee temperature while the winding sensor is still reporting a comfortable number, and the damage is done well before any protection trips.
Why the thermistor does not protect the magnet
A defensible analysis couples the two: compute the peak current the inverter can deliver into a terminal short at the highest coolant temperature, convert it into the demagnetising field in the magnet layer, compare that with the knee of the B-H curve at the maximum magnet temperature, and leave margin. A supplier who can produce that calculation for the actual rotor geometry is worth more than one who quotes a grade and stops.
Segmentation of the magnets reduces eddy-current heating at high speed, which keeps the magnet further from its limit. Rotor surface speed raises magnet temperature through harmonic losses, so a high-speed design with the same grade has less margin than a low-speed one at the same winding temperature.
How the Magnet Changes the Rotor
Once the grade is fixed, the question becomes where to put the magnets, and that choice changes the machine more than the grade does. The two standard topologies behave differently enough that they serve different applications.
| Property | Surface-mounted (SPM) | Interior (IPM) |
|---|---|---|
| Magnet position | On the rotor surface, retained by a sleeve or band | Buried in slots inside the rotor lamination |
| Torque components | Magnet torque only | Magnet torque plus reluctance torque |
| Saliency | None, so inductance is uniform | Strong, so the d and q axes differ |
| Field weakening | Requires continuous negative d-axis current | Reluctance torque helps; wider constant-power range |
| Demagnetisation risk | Higher, because weakening current flows against the magnet | Lower, because the magnets sit behind iron |
| Mechanical strength at speed | Retention sleeve is the limit | Iron bridges carry the load |
| Cogging | Low to moderate | Higher unless the pole shape is optimised |
Grade and topology interact directly. An interior design with UH magnets can hold a four-to-one constant-power range without risking demagnetisation, because the magnets sit behind iron and the reluctance torque does part of the work. A surface design attempting the same range needs a continuous weakening current that pushes the operating point down, which means either a higher coercivity grade or a lower top speed. The two decisions therefore have to be settled at the same time.
For servo duty the comparison is usually against other permanent magnet machines rather than against induction, and the trade-offs are set out in BLDC motor versus PMSM motor. The product range that follows from these choices is described on the PMSM motor page.
Auditing an Efficiency Claim
Permanent magnet motors are usually sold on efficiency, and the efficiency claim is the part of the datasheet most often stated in a way nobody can check. The four documents below do different jobs, and most of the confusion comes from quoting a class from one against a test method from another.
| Instrument | What it covers | What it does not cover |
|---|---|---|
| IEC 60034-30-1:2025 | Line-operated single-speed motors, 0.12 to 1 000 kW, 2 to 8 poles, including line-start permanent magnet machines | Converter-fed machines, multi-speed machines, motors with mechanical commutators |
| IEC TS 60034-30-2:2016 | Efficiency classes for variable-speed machines not covered by 30-1 | Line-operated machines |
| IEC 60034-2-1:2024 | How losses and efficiency are determined by test for line-operated machines | Converter-fed losses |
| IEC 60034-2-3:2024 | Loss and efficiency determination for converter-fed machines, at seven standardised load points | Line-operated machines |
Three corrections to the usual claim
The class is technology-neutral. IEC 60034-30-1 states explicitly that no distinction is made between motor technologies, which is what allows an induction motor and a permanent magnet motor to be compared at all. A well-designed induction machine can reach IE4, and a synchronous reluctance machine can reach IE5. The magnet does not confer the class [2].
The gap between classes is also smaller than the marketing implies. For a 7.5 kW four-pole machine at 50 Hz the standard sets IE4 at 92.1 percent and IE5 at 93.0 percent, and those percentages hide the useful number: 643 W of loss at IE4 against 565 W at IE5, a difference of about 79 W, or 12 percent of the losses. Real, worth having, and much smaller than the impression a label creates.
A converter-fed machine falls outside that standard altogether. Efficiency classes for variable-speed machines sit in IEC TS 60034-30-2, and the test method in IEC 60034-2-3 evaluates seven standardised operating points instead of one [3][5]. An IE5 badge on a drive-fed servo therefore rests on a different document, and the buyer should ask which one. The measurement chain for line-operated machines is in IEC 60034-2-1 [4]. The practical difference between the two supply cases is covered in VFD motor versus normal motor.
What the magnet genuinely removes is the rotor winding loss: an induction rotor carries induced current and therefore real copper loss, and in small machines that term is one of the two largest in the loss budget. Deleting it is why a permanent magnet machine can reach the top classes without growing a frame size. The wider comparison of efficiency classes is covered in what a motor efficiency class means and IE2 versus IE3.
What to Write on the Specification Sheet
A purchase order that says “rare earth permanent magnet motor, 5 kW” gives the supplier almost unlimited room. These are the lines that close it.
| Line item | Why it matters |
|---|---|
| Grade and suffix, for example 42SH | Fixes both the energy product and the temperature class |
| Minimum intrinsic coercivity in kA/m | The suffix sets a floor; state it so the floor is contractual |
| Minimum remanence or energy product | Prevents a high-coercivity, low-flux substitution |
| Maximum magnet temperature and the assumed permeance coefficient | Makes the temperature claim checkable rather than rhetorical |
| Coating type and thickness | Nickel-copper-nickel or epoxy; NdFeB corrodes without it |
| Demagnetisation calculation at short circuit and maximum temperature | The only evidence that the grade is sufficient rather than merely fashionable |
| Torque constant and the temperature at which it is quoted | A Kt without a temperature is not a specification |
| Basis of the efficiency claim, with the standard number | Separates a 60034-30-1 class from a 60034-30-2 class |
| Vibration grade to IEC 60034-14 | Magnets add rotor mass and stiffness that change the balance case |
| Enclosure to IEC 60034-5 | The magnet is inside, but the winding still has to survive the environment |
The torque constant without a temperature reference is the most frequent omission in permanent magnet datasheets, and the demagnetisation calculation is the one document that turns a grade choice from an opinion into an engineering decision. Both lines are cheap to add. Rating and thermal behaviour for the machine as a whole remain governed by IEC 60034-1 [6], and enclosure protection by IEC 60034-5 [7]. What actually reaches the nameplate, and what a replacement decision can rely on, is set out in how to read a motor nameplate.
Where It Wins, and Where It Does Not
The permanent magnet machine wins wherever torque density, part-load efficiency or a wide constant-power range matters more than first cost. It loses wherever the duty is intermittent, the ambient is uncontrolled and the price is the deciding variable.
| Application | Why the magnet helps | The constraint that decides |
|---|---|---|
| Traction and e-mobility | Torque density and a wide constant-power range | Magnet temperature under repeated hill climbs |
| Industrial pumps and fans on drives | Part-load efficiency, which is where most operating hours sit | Whether the IE claim is under 30-1 or 30-2 |
| Servo axes and robotics | Torque per frame size, low inertia, precise control | Cogging and the demagnetisation margin at peak torque |
| Compressors and HVAC | High speed in a small frame | Rotor retention and magnet heating at speed |
| Aerospace and vacuum actuators | No field winding, high temperature capability with SmCo | Qualification cost, and SmCo supply lead time |
| Fixed-speed, low-duty machinery | Little; an induction motor is cheaper | First cost, not efficiency |
The last row is worth stating plainly, because it is the commercially useful one. A machine that runs two hundred hours a year at fixed speed will never repay the price difference, and a magnet motor specified into that duty is a cost with no return. The energy argument works where annual hours are high and the load varies. It does not work as a general principle.
Motor selection for a driven machine follows the same logic as for any other prime mover; the sequence for a fluid load is worked through in choosing the right motor for your pump, and the family relationships are mapped in AC synchronous motors and synchronous versus induction motor.
Six Mistakes That Reach Production
- Treating “rare earth motor” as one specification. NdFeB and SmCo differ by a factor of three in temperature coefficient. A quotation that names neither is a quotation for either.
- Buying the highest suffix for safety. It costs remanence, money and sometimes a redesign of the rotor to fit a weaker magnet. The cheaper fix is usually thermal.
- Quoting a torque constant without a temperature. The same motor has a torque constant 13 percent lower at 140 °C than at 20 °C. A Kt without a temperature reference cannot be verified.
- Assuming the winding sensor protects the magnet. The magnet is a separate thermal mass and can pass its knee while the winding reads normal.
- Testing demagnetisation with an overload. A terminal short at high temperature is the governing case in most machines, and it is the one that is rarely analysed.
- Using an IE5 claim outside its standard. A drive-fed machine is classified under IEC TS 60034-30-2, not IEC 60034-30-1, and the two are not interchangeable on a datasheet.
Troubleshooting: Symptom, Cause, Action
| Symptom | Most likely cause | Action |
|---|---|---|
| Torque falls after an hour of running, recovers when cold | Reversible flux loss with magnet temperature | Log magnet or rotor temperature; check the cooling path before changing the grade |
| Torque never recovers after an event | Irreversible demagnetisation past the knee | Measure back-EMF at fixed speed against the original value; replace magnets and review the short-circuit case |
| Back-EMF lower than drawing on a new batch | Wrong grade, or a high-coercivity low-flux substitution | Check remanence and coercivity certificates; add both to incoming inspection |
| Rust or blistering on the magnet surface | Coating damage or an uncoated NdFeB part | Inspect handling and fixturing; specify nickel-copper-nickel or epoxy explicitly |
| Vibration rising with speed | Rotor balance after magnet assembly, or magnet retention wear | Rebalance the assembled rotor; check sleeves and retaining bands against IEC 60034-14 |
| Overcurrent trip at start on a drive | Damaged magnets producing uneven back-EMF, or a rotor position offset | Check the three line-to-line back-EMF waveforms for symmetry, then re-identify the encoder offset |
| Hot spot on one pole of the rotor | Localised eddy-current heating or a shorted lamination | Inspect lamination insulation and magnet segmentation; consider a segmented magnet design |
The first two rows are the pair that separates a thermal problem from a materials problem, and the back-EMF measurement is what settles it. A machine that lost flux permanently will show a permanently low back-EMF at a fixed test speed, which is a fifteen-minute check and does not require dismantling the rotor.
How Greensky Power Specifies Magnets
We treat the magnet grade as part of the motor specification, not as a purchased component. That means the grade, the coercivity floor and the assumed permeance coefficient are agreed with the customer before the rotor is drawn, and the demagnetisation case is analysed at the maximum magnet temperature rather than at the winding temperature.
In practice a quotation from us carries four things a magnet-agnostic one cannot: the grade and suffix with minimum coercivity, the torque constant together with the temperature at which it applies, the maximum magnet temperature supported by the cooling arrangement we are proposing, and the basis of any efficiency claim with the standard number attached. Where the duty is hot, we say so before the drawing is issued. A field failure is an expensive place to find out.
If you already have a rotor design and want the grade checked against a real duty cycle, send the torque profile, the ambient and the inverter’s peak current into a short circuit. Those three numbers decide the answer. Send the duty cycle to our engineering team, or start from the motor application library and the custom motor programme. The efficiency argument in more depth is in what makes a high-efficiency motor more efficient.
Send us the torque profile, the ambient temperature and the inverter’s short-circuit current, and we will return the grade, the coercivity floor and the demagnetisation margin in writing. Contact Greensky Power.
References
- IEC. (2023). IEC 60404-8-1:2023, Magnetic materials – Part 8-1: Specifications for individual materials – Permanent magnet (magnetically hard) materials, Edition 4.0. IEC Webstore. https://webstore.iec.ch/en/publication/68440/
- IEC. (2025). IEC 60034-30-1:2025, Rotating electrical machines – Part 30-1: Efficiency classes of line operated AC motors (IE code), Edition 2.0. IEC Webstore. https://webstore.iec.ch/en/publication/91195
- IEC. (2016). IEC TS 60034-30-2:2016, Rotating electrical machines – Part 30-2: Efficiency classes of variable speed AC motors (IE-code), Edition 1.0. IEC Webstore. https://webstore.iec.ch/en/publication/30830
- IEC. (2024). IEC 60034-2-1:2024, Rotating electrical machines – Part 2-1: Standard methods for determining losses and efficiency from tests, Edition 3.0. IEC Webstore. https://webstore.iec.ch/en/publication/67756
- IEC. (2024). IEC 60034-2-3:2024, Rotating electrical machines – Part 2-3: Specific test methods for determining losses and efficiency of converter-fed AC motors, Edition 2.0. IEC Webstore. https://webstore.iec.ch/en/publication/67758
- IEC. (2026). IEC 60034-1:2026, Rotating electrical machines – Part 1: Rating and performance, Edition 15.0. IEC Webstore. https://webstore.iec.ch/en/publication/89961
- IEC. (2020). IEC 60034-5:2020, Rotating electrical machines – Part 5: Degrees of protection provided by the integral design of rotating electrical machines (IP code) – Classification, Edition 5.0. IEC Webstore. https://webstore.iec.ch/en/publication/64224
- IEC. (2018). IEC 60034-14:2018, Rotating electrical machines – Part 14: Mechanical vibration of certain machines with shaft heights 56 mm and higher – Measurement, evaluation and limits of vibration severity, Edition 4.0. IEC Webstore. https://webstore.iec.ch/en/publication/59889
- European Parliament and Council. (2024). Regulation (EU) 2024/1252 establishing a framework to ensure a secure and sustainable supply of critical raw materials. Official Journal of the European Union. https://eur-lex.europa.eu/eli/reg/2024/1252/oj
- U.S. Geological Survey. (2025). Mineral Commodity Summaries 2025: Rare Earths. U.S. Department of the Interior. https://pubs.usgs.gov/periodicals/mcs2025/mcs2025-rare-earths.pdf
FAQ
Q: What is a rare earth permanent magnet motor? It is a motor whose excitation comes from permanent magnets made of rare earth elements instead of from a field winding or an induced rotor current. In practice that means neodymium-iron-boron or samarium-cobalt magnets inside the rotor. The motor produces torque from the interaction between the magnet flux and the stator current, so it needs no rotor copper loss and no external excitation.
Q: What does the suffix in a magnet grade such as N42SH mean? The number is the maximum energy product in megagauss-oersted, so 42 means roughly 42 MGOe or 334 kJ per cubic metre. The letter code is the temperature class and it sets a minimum intrinsic coercivity: N for about 80 °C, M 100, H 120, SH 150, UH 180, EH 200 and AH 230. Two magnets with the same number and different letters have almost the same room-temperature strength but very different resistance to demagnetisation when hot.
Q: Why does a hot magnet make the motor less efficient? Remanence falls with temperature at about 0.11 percent per kelvin for standard neodymium grades, so a magnet at 140 °C produces about 13 percent less flux than at 20 °C. The torque constant falls in proportion, which means about 15 percent more current for the same torque. Because copper loss follows the square of current, that is about 33 percent more copper loss, and the winding resistance is itself higher at that temperature, so the two effects compound.
Q: What causes permanent demagnetisation? It is a demagnetising field strong enough to push the magnet operating point below the knee of its second-quadrant curve at that temperature. The worst case in a motor is usually a terminal short circuit at high magnet temperature, not a mechanical overload, because the current is large and there is no air gap flux to help. Once the knee is crossed the loss is not recovered by cooling.
Q: Is a rare earth motor always IE4 or IE5? No. IEC 60034-30-1 makes no distinction between motor technologies, so an induction motor can also reach IE4, and a reluctance machine can reach IE5. What the magnet does is remove the rotor winding loss, which is one of the largest single loss terms in a small induction motor. A converter-fed permanent magnet machine is outside the scope of that standard and is classified under IEC TS 60034-30-2 instead.
Q: Neodymium or samarium-cobalt? Neodymium-iron-boron gives the highest remanence, about 1.1 to 1.45 tesla, and is the default choice for traction, servo and appliance motors. Samarium-cobalt gives less, about 0.85 to 1.15 tesla, but holds its flux far better when hot and has a Curie temperature above 700 °C, so it is chosen for high ambient, vacuum, aerospace and safety-critical duties where a magnet failure cannot be serviced.
