Servo drives marketed with a 300% peak current rating are built to deliver three times their continuous RMS current for a short, well-defined window, typically 1-5 seconds depending on the manufacturer and the cooling method, after which the drive folds back to its continuous rating to protect the IGBT power stage [S2][S3][S5].
Engineers reading a datasheet should not treat that 300% figure as a free torque multiplier for the full move profile; it is a thermal envelope, sized for acceleration transients, punch-and-return cycles on presses, and short overload events like a packaging-machine index [S2][S5]. The two numbers that always travel with the 300% rating are the peak duration (T_pk) and the recovery time, and both are spec items, not footnotes [S3][S4].
What the 300% rating actually means on a datasheet
Industry convention groups servo drive current ratings into a continuous (nominal) value and a peak (transient) value, with peak typically 2x or 3x continuous on most modern industrial units [S2]. The peak is the value the drive can deliver "for a defined peak time" while staying within the thermal limits stated in the manual, ambient temperature and heatsink conditions assumed [S3]. Celera Motion's Ingenia documentation defines this explicitly: peak current is the transient value the drive can withstand for a stated peak time, expressed either in DC amperes or RMS amperes depending on the product line [S3]. The same paper warns that "peak" is ambiguous unless qualified, and a True-RMS reading on a multimeter will not capture it; an oscilloscope and current probe, or the drive's own firmware register, are the honest tools [S3].
RMS vs crest is a recurring trap. For sinusoidal commutation, crest = RMS x sqrt(2), so a 17 A RMS phase current implies roughly 24 A crest at the rated operating point [S3]. When a vendor lists 300% peak against a continuous RMS number, the underlying crest multiplier is larger; the 300% is conventionally stated against the continuous rating in the same unit system the datasheet uses, not a mixed unit [S3]. For an application engineer the practical rule is: convert everything to one unit (crest or RMS) before comparing two drives side by side, and size the upstream transformer, bus capacitor, and regen resistor against the peak in the same unit [S3].
How long the 300% actually lasts: typical duration numbers
Peak current duration is a separate, named parameter on most datasheets, and it varies widely by product family. Rockwell's Kinetix 5500 design guide defines T_PKmax (peak current duration) as 1.0 second for that drive family, with the further caveat that peak current must also be bounded by the motor's peak stall-current values [S4]. Kollmorgen's S300 (Servostar) datasheet lists peak output current as "max. 5 s" at rated DC link voltages of 145-360 V and 560-675 V, with tolerance specified at plus or minus 3% [S6]. SIGMATEK's DIAS Drive 2000 series pushes the overload factor to 300% with a 3x rated-to-peak relationship (5 A rated to 15 A peak in size 1, 10 A to 30 A in size 2), at controller cycle times of 62.5 microseconds [S5].
These are not the same product, and the durations reflect different thermal strategies. A 1-second peak is typical of drives sized to ride through a single acceleration ramp; a 5-second peak suits presses and indexing tables that need a sustained punch force. Beyond the published peak window, the drive drops to its continuous limit, and any further demand above continuous must wait for the thermal mass to recover [S3][S4]. The servo drive product page covers how the internal DC bus, IGBT stage, and firmware current loop cooperate to enforce that envelope.
Why 300% specifically, and not 200% or 400%
The 2x to 3x peak-to-continuous range is a thermal and semiconductor trade-off, not a marketing accident. Raising peak current forces larger IGBT dies, bigger bus capacitors, and more copper in the motor windings to absorb the I-squared-R heating during the burst [S2][S3]. A drive that advertises 300% peak is generally a compact multi-axis unit, with modular axis modules and shared DC-link capacitance sized for the aggregated peak demand of several axes accelerating at once [S5]. Going past 3x typically requires a physically larger module, active liquid cooling, or a derated continuous current, which erodes the size advantage that drives the multi-axis trend in packaging and robotics [S5][S7].
Pulse-duty servo systems are the design point where the 300% number is most useful: a lower-inertia rotor, a high-overload drive, and a duty cycle where the peak demand is short and well-spaced. Motion Control Tips notes that pulse-duty drives have a high current overload rating by design and are paired with low-inertia motors, exactly the configuration that benefits from a 3x ceiling [S7]. For an axis that runs at 50% continuous load for 90% of the cycle and then needs a 3x burst for the remaining 10%, the servo motor page describes the matching thermal-class considerations on the rotor side.
Sizing the DC bus, supply, and regen path for the 300% case
Three subsystems must be sized to the peak, not the continuous number, or the drive will brown out, trip DC bus undervoltage, or cook the supply during a burst. First, the DC power supply must handle the peak inrush, and Advanced Motion Controls' 2026 sizing guide recommends determining the continuous current, peak current handling, output voltage, and regenerative capacity of the supply as the first step, not an afterthought [S8]. Second, the DC-link capacitance inside the drive must hold bus voltage during the peak; undersized capacitance causes the bus to sag and the drive to drop out of the current loop, which shows up as torque ripple or a fault [S1]. Third, regeneration during deceleration must be absorbed by an external brake resistor sized for the peak's energy, not the continuous rating, or the drive will trip DC bus overvoltage on every stop [S8].
Voltage-range selection is the upstream lever that lets a designer cut the current requirement before the peak even happens. Elmo's voltage-budget guidance is concrete: calculate the worst-case minimum and maximum bus voltage from supply tolerance, cable drop, load sag, and regenerative rise, then add 10-15% headroom on both ends [S1]. For a given power, raising bus voltage proportionally lowers current, which cuts I-squared-R losses, permits thinner cables, and often allows a smaller, less costly drive, but it raises the back-EMF and changes the speed ceiling on the motor [S1]. Matching the servo drive voltage range to the mechanical demand is the single biggest lever for keeping the 300% rating from being a thermal liability instead of an asset.
Reading duration, duty cycle, and recovery from the datasheet
Three numbers must come off the datasheet before the drive is approved for a 300% application, and each is sourced. First, peak duration T_pk: 1.0 s for the Kinetix 5500 [S4], 5 s for the S300 [S6], and a "very short" controller cycle of 62.5 microseconds for the DIAS Drive 2000 (cycle time, not peak duration, but it sets the dynamic response) [S5]. Second, the recovery or cool-down time the drive enforces after a peak before another peak is allowed, which is rarely in the headline spec and is usually a firmware register or an I-squared-t integrator output [S3]. Third, the peak-to-peak frequency the drive will accept, because the IGBT junction temperature accumulates over time and a 300% burst every 200 ms is a different thermal problem from a single 3-second burst every 30 s [S3][S10].
Parker's OEM300 sizing manual frames this as a duty-cycle calculation: measure the duration of the peak demand, find how frequently it occurs, and calculate its duty cycle, then check that against the drive's I-squared-t thermal model [S10]. The Kollmorgen selection guide reinforces the point that 2-3x peak-to-continuous is the market norm, and that the peak is for short bursts, not sustained operation [S2]. A reader who treats 300% as a free torque multiplier without computing the duty cycle will see a drive that runs hot, derates itself in the field, or trips on IGBT over-temperature within weeks of commissioning.
What the 300% rating is not, and where it fails
The 300% peak rating does not extend to the motor unless the motor is also rated for it, and that is the most common engineering miss. Motor peak current is a user-configured parameter bounded by the motor's thermal mass, the encoder feedback resolution, and the demagnetization limit of the magnets, and it is typically lower than the drive's peak output, so the system peak is the smaller of the two [S3][S4]. A drive that can deliver 30 A peak into a motor rated for 15 A peak will either trip a motor overcurrent fault or, worse, slowly demagnetize the rotor [S3]. The Kinetix 5500 design guide explicitly requires that the drive's peak output be checked against the motor's peak stall-current values, not just the drive's own rating [S4].
The 300% rating also does not extend to the upstream AC mains or the DC supply. A multi-axis system where three axes all demand 300% simultaneously pulls nine times the continuous current from the supply, and that aggregated peak must be budgeted into the feed, transformer, and fuse selection [S5][S8]. Fails to do this and the supply sags, the drives fault on DC undervoltage, and the line shows up as a 100 ms brownout on the plant power monitor. For drives packaged with the motor as a matched drive motor set, the vendor's combined datasheet usually covers this, but third-party pairings require the engineer to redo the aggregation by hand.
Practical selection checklist for a 300% application
Four checks decide whether a 300%-rated drive is the right pick for a given axis. First, the peak demand on the cycle: compute the I-squared-t of the worst-case move and verify it fits inside the drive's T_pk plus the recovery window [S3][S10]. Second, the duty cycle: a 300% drive makes sense when the peak is a small fraction of the cycle, typically under 20% on-time; for higher duty cycles, oversize the continuous rating instead [S2][S7]. Third, the motor pairing: confirm the motor's peak current, peak torque, and thermal time constant all support the same peak, and that the encoder bandwidth can close the current loop at the required rate [S3][S4]. Fourth, the supply and regen: verify the DC supply, bus capacitance, and brake resistor are sized for the peak in the same unit system the datasheet uses, with a 10-15% voltage headroom on top [S1][S8].
Engineers working on servo press applications should weigh the 300% peak against the 500% or higher peaks some press-duty drives advertise, because the mechanical demand on a press is closer to a single short-duration impulse than a continuous overload. Slewing drive and eddy-current-test eddy current tester applications, by contrast, rarely need more than 2x because the load is constant and slow. Matching the peak envelope to the application is the whole game; the 300% number is a tool, not a target.
The next signal to track is the revision of IEC 61800-1 and IEC 61800-2, which govern adjustable-speed power drive systems and which several working groups are aligning with the newer wide-bandgap semiconductor (SiC, GaN) datasheets, where 300% peak is becoming the entry-level rather than the premium. A second watch item is the migration of pulse-duty drive/motor pairs into collaborative-robot joints, where the 300% envelope on a low-inertia rotor is replacing the older 2x continuous on a high-inertia servo as the default design point for cobot gearboxes.
This topic is covered further in Air impact wrench price: drive size and torque class dominate the cost stack.