Home / Reference Designs / REF-11KW-PFC-SIC-QD

Magnetics for the REF-11KW-PFC-SIC-QD 11 kW ANPC PFC

REF-11KW-PFC-SIC-QD is an Infineon reference design for an 11 kW bidirectional PFC and inverter that runs from a three-phase or a single-phase grid. It uses a three-level active neutral point clamped (ANPC) leg per phase at 65 kHz. Its application note describes the construction of the three line inductors and of the EMI choke in more detail than most reference designs: core, turns and wire. It does not give an inductance versus current curve, the differential mode inductance of the choke, or a loss figure for the magnetics. We build these parts from the published construction or from your modified targets.

Send your specs

What the board is

  • Three-phase input from 318 to 480 V RMS line to line (400 V typical), or single-phase from 180 to 265 V RMS (230 V typical), 50 Hz.
  • 800 V DC bus, 11 kW in three phase and 7.3 kW in single phase, in both directions.
  • Three-level ANPC legs: two CoolSiC devices switch at 65 kHz, four CoolMOS devices switch at line frequency.
  • Peak efficiency 99.15 % as a three-phase PFC and 98.95 % in single phase, at 230 V.
  • 11.3 kW/L without heatsink, 243 × 100 × 40 mm without the cooling plate.

Infineon REF-11KW-PFC-SIC-QD product page (`rel="noopener nofollow"`)

Source: Infineon REF-11KW-PFC-SIC-QD product page and application note AN123017 V1.0 (29 July 2025), including its schematics (section 7) and bill of materials (section 8), checked November 2026. See REF-11KW-PFC-SIC-QD in our reference design table.

The magnetics on the board

Magnetics on REF-11KW-PFC-SIC-QD
PartWhat the application note publishesWhat decides the design
Line inductors, L1 to L3A pair of HTEQ3626A high-flux powder EQ cores (Chang Sung), helical flat-wire winding of 7 × 0.4 mm, 50 turns, 330 µH as tested, 16 A RMS rated. BOM part MG2000010151; the schematic labels it MG1000010152Inductance held at peak current, since a powder core rolls off with bias; ripple at the worst point of the line cycle; stray field from the EQ shape near the current sensors
Common mode choke with integrated differential mode, L4 and L5Two R10K ferrite PQ35-43A cores (DMEGC), four helical flat-wire windings of 6 × 0.6 mm and 8 turns, and three sets of magnetic shunts in 0.3 mm NO 30-16 electrical steel between the windings. Stacking shunts raises the differential mode inductance. In single phase all four windings carry current. BOM: MG1000010151, "CM choke 3 phase 4 winding". No inductance valuesCommon mode inductance, differential mode inductance set by the shunts, and saturation of the shunts at full line current
Auxiliary transformers, TF1 and TF2ICE 8032.0205.021 for two quasi-resonant flybacks with the ICE2QR2280G, with ER18 and planar cores listed as B66480G0000X149 and B66453K0000X149Isolation and bias regulation, a small part of the board
Control card inductor47 µH, 0.6 A, WE-LQS 2512Catalogue part

Checks we ran on the published values

  • In a three-level leg with an 800 V bus, the inductor sees steps of 400 V. The ripple peaks where the phase voltage is a quarter of the bus, 200 V, at Vdc / (8 L f). With 330 µH at 65 kHz that is 4.7 A peak to peak. At the line peak of a 230 V phase it is only 2.8 A.
  • The application note measures 6 A peak to peak at 11 kW. Read against the same formula, that corresponds to about 256 µH at the operating current, 78 % of the 330 µH tested value, which is in line with the roll-off of a high-flux powder core under bias.
  • At 11 kW and 230 V per phase, each inductor carries 15.9 A RMS, 22.5 A peak. The 16 A RMS rating is used almost completely, and the 6 A ripple is 27 % of the peak current.
  • In single phase, the L1 and L2 terminals are joined, legs A and B run interleaved on the line, and L3 with N is the return to the DC midpoint, while leg C works as a power pulsation buffer. 7.3 kW at 230 V is 31.7 A RMS, so each of the two active inductors carries about 15.9 A RMS, the same as in three-phase operation at 11 kW. That is why the single-phase rating is two thirds of the three-phase one, and why one 16 A RMS inductor design serves both modes.
  • The protection limit for the choke temperature is 120 °C. With the chassis closed and a fan on the heatsink, the note reports a maximum of about 100 °C on the inductors across the specified input range: 20 K of margin under lab conditions.

What the public documents do not say

  • Inductance versus current for the line inductors, and the inductance at the 22.5 A peak.
  • Common mode and differential mode inductance of the EMI choke, and how many shunts the build uses for its differential mode value.
  • The values behind the loss breakdown chart. The note says that the winding losses of the inductors are among the main contributors, but gives no number.
  • Who makes the finished inductors and chokes, and their insulation class.
  • The ambient temperature behind the closed-chassis inductor measurement, and the winding hotspot temperature.

If you change the design

  • At the same inductance and frequency, the worst ripple scales with the bus voltage: a 1000 V bus raises it from 4.7 A to about 5.8 A peak to peak.
  • Halving the switching frequency doubles the ripple at the same inductance, or doubles the inductance needed for the same ripple, with a larger core.
  • A 22 kW version at the same grid doubles the phase current to about 32 A RMS. Copper loss in a winding of the same size rises fourfold, so the inductor has to be designed again, not paralleled.
  • Moving the current sensors or changing the inductor shape changes the stray field the sensors see. The application note's placement advice is valid only for its own geometry and sensor.
  • A different conducted EMI limit changes the common mode choke and its differential mode shunts. See our EMI filters.

What to send us

The reference number, your grid range, bus voltage, power and switching frequency if you changed them, the ripple target, the conducted EMI limit you test against, the case dimensions, the cooling method and the quantities you expect.

We aim to send an engineering answer with a first design proposal inside 24 hours, then 4 to 8 weeks from an agreed design to a prototype, subject to raw material stock. No minimum order quantity.

Straight answers

REF-11KW-PFC-SIC-QD, the usual questions.

Is ConverNova a partner of Infineon?
No. We are an independent designer and manufacturer of custom magnetics. We work from the public reference design and from what you tell us about your version of it.
Can you build the REF-11KW-PFC-SIC-QD line inductor from the application note?
Yes. The note gives the core, the turns, the flat wire and the tested inductance, which is enough to build it. We agree with you the inductance it must hold at peak current, because that, not the value at zero current, sets the ripple.
Why is the differential mode inductance inside the common mode choke?
The steel shunts between the windings give the leakage flux a low reluctance path, so the choke adds differential mode inductance without separate inductors. It saves volume. The note does not give the resulting value, so we design it to your EMI margin or measure it on your build.
Do the line inductors disturb the current sensors?
They can. The application note explains that the EQ shape and the powder core leak field that disturbs Hall and magnetoresistive sensors, and solves it by placing and orienting one specific sensor. If you change the inductor or the sensor, check the stray field again.