How it works
A dual active bridge (DAB) converter connects two full bridges through a high-frequency transformer and a series inductance. Power flows from the leading bridge to the lagging bridge, and its amount is set by the phase shift between the two bridge voltages. Because both bridges are active, power can flow in either direction, which makes DAB a common choice for battery chargers, energy storage and solid-state transformers. This tool covers forward power, from V1 to V2.
Conventions used everywhere on this page: the transformer is 1:n with n = N2/N1, V2' = V2/n is the secondary voltage referred to the primary, M = V2/(n·V1), and Th = 1/(2·f_s) is the half period. L is on the V1 side and currents are V1-side inductor currents. Every phase ratio is a fraction of Th.
Single phase shift (SPS) keeps both bridges at full square waves and uses one ratio D between them. Dual phase shift (DPS) adds an inner shift D1 in both bridges, so each bridge voltage is a pulse of width (1 − D1)·Th, and D2 is the outer shift between the bridges. Triple phase shift (TPS) sets the two pulse widths independently: here D1 is the primary pulse width ratio, D2 is the secondary pulse width ratio and φ is the shift between the pulse centers. Note that D1 means different things in DPS and TPS; a DPS setting with inner shift D1 equals TPS with both widths at 1 − D1, and SPS equals TPS with both widths at 1.
SPS and DPS have closed forms, shown below and printed next to the exact value as a self-check. TPS has no simple closed form across all its modes, so this tool computes P, I_rms, I_peak and the switching currents exactly from the piecewise-linear inductor current. Between voltage edges di/dt = (v1 − v2')/L is constant, so no numerical integration is involved. The model assumes ideal switches, a lossless L and no magnetizing current. For fixed pulse widths P rises monotonically with the outer shift from 0 to 0.5, so the outer shift is found by bisection. The DPS and TPS optimum is found numerically on a grid; this is not the closed-form solution of the cited TPS papers, which introduce the modulation concept. The closed-form DPS expressions shown here were checked against the exact time-domain evaluator (max difference 1e-15). Bai and Mi (2008) print different expressions, twice these values where D2 ≤ D1; see Zhao, Song and Liu (2012) for a later treatment.
ZVS is checked with one rule for all modes. At each switching instant the inductor current must have the polarity that discharges the capacitance of the incoming switch, and it must carry enough energy: ½·L·i² ≥ ½·(2·C_eq)·V². This takes the node capacitance of one leg as twice the switch capacitance C_eq, and it gives I_min = V·√(2·C_eq/L). A status of marginal means the polarity is right but the current is below I_min. It is a per-leg energy estimate: it ignores dead-time limits and the two legs that swing together in SPS, so leave margin.
Design procedure
- 1Fix the operating range: V1, the minimum and maximum of V2, the power range and the switching frequency.
- 2Choose n so that M = V2 / (n · V1) is close to 1 over the most common operating range. Leaving n empty uses V2_max / V1.
- 3Choose D_max for the corner (V2_min, P_max). A smaller D_max means a smaller inductance and lower RMS current but less control headroom; values around 0.25 to 0.35 are common.
- 4Read the four corner rows: D, peak and RMS current, and the ZVS status of each bridge. Enter C_eq to apply the energy criterion.
- 5If ZVS is lost or RMS current is high at light load or when M is far from 1, compare DPS and TPS at the same L and n, and check whether their optimum keeps ZVS.
Formulas
- Definitions
- M = V2 / (n · V1), V2' = V2 / n, Th = 1 / (2 · f_s)
- Power base
- k = V1 · V2 / (2 · n · f_s · L)
- SPS power
- P = k · D · (1 − D)
- Series inductance (SPS design)
- L = D_max · (1 − D_max) · V1 · V2_min / (2 · n · f_s · P_max)
- DPS power, D1 inner, D2 outer (use 1 − D2 for D2 > 0.5)
- D2 ≥ D1: P = k · [D2 · (1 − D2) − D1² / 2]; D2 ≤ D1 and D2 ≤ 1 − D1: P = k · [D2 · (1 − D1) − D2² / 2]; 1 − D1 ≤ D2 ≤ D1: P = k · (1 − D1)² / 2 (0 ≤ D2 ≤ 0.5)
- TPS: exact time-domain evaluation
- v1, v2' piecewise → di/dt = (v1 − v2') / L → P = (1/T) ∮ v1 · i dt
- ZVS minimum current
- I_min = V · √(2 · C_eq / L)
Worked example
With V1 = 48 V, V2 = 350 to 400 V, 100 W to 1 kW, 100 kHz and D_max = 0.35, leaving n and L empty gives n = 8.33 and L = 2.29 µH on the V1 side. In SPS, (350 V, 1 kW) needs D = 0.350 with I_peak = 38.59 A and I_rms = 30.24 A; (400 V, 1 kW) needs D = 0.274 with 28.71 A and 25.95 A. At (350 V, 100 W), D = 0.023 with 8.67 A and 4.40 A, and the V2 bridge hard-switches because its commutation current is −4.10 A. DPS with minimum peak current at that corner gives D1 = 0.41 and D2 = 0.040, with 7.51 A and 4.22 A, but the V2 bridge still hard-switches. TPS with minimum RMS current gives D1 = 0.590, D2 = 0.675 and φ = 0.039, with 7.39 A and 3.43 A. Entering C_eq_i = 300 pF and C_eq_o = 50 pF shows both bridges as marginal there: 0.33 A against an I_min of 0.78 A and 0.03 A against 2.31 A.
Common mistakes
- Mixing conventions for n, D and the side to which L is referred. This page uses n = N2/N1 and L on the V1 side. Many papers use n = N1/N2 or refer L elsewhere, so convert before comparing.
- Using D close to 0.5 at full power. Power is maximum at D = 0.5, so there is no control margin left for transients.
- Ignoring the transformer leakage inductance, which adds to the external series inductance.
- Checking ZVS only at full power. SPS usually loses ZVS first at light load when M ≠ 1.
- Taking the current optimum as a finished design. The TPS and DPS optimizers ignore ZVS, and low RMS operation often leaves very little switching current.
Frequently asked questions
When should I use DPS or TPS instead of SPS?
When M moves far from 1 or the load is light, SPS loses ZVS and circulates current. In the example above, TPS cuts the RMS current at (350 V, 100 W) from 4.40 A to 3.43 A. At full power near M = 1 both optimizers return full-width pulses, which is SPS, so there is nothing to gain there.
What do D, D1, D2 and φ mean?
All are fractions of the half period. D is the SPS phase shift. In DPS, D1 is the inner shift in each bridge and D2 is the outer shift between the bridges. In TPS, D1 and D2 are the pulse widths of the two bridges and φ is the shift between pulse centers. For SPS, maximum power is at D = 0.5.
Why is L designed at V2_min rather than V2_max?
P = k·D(1 − D) and k is proportional to V2, so at P_max the lowest V2 needs the largest phase shift. Designing L there sets D = D_max at the hardest corner. Designing at V2_max would leave (V2_min, P_max) needing more than D_max, and possibly more power than the converter can reach.
Why does the ZVS check include parasitic capacitance?
The inductor current at the switching instant must charge and discharge the switch output capacitances before the next switch turns on. Right polarity alone is not enough: when the stored energy ½·L·i² is below ½·(2·C_eq)·V², the transition does not finish and the status is marginal.
References
- R. W. De Doncker, D. M. Divan, M. H. Kheraluwala, "A three-phase soft-switched high-power-density DC/DC converter for high-power applications," IEEE Trans. Ind. Appl., vol. 27, no. 1, pp. 63–73, 1991.
- H. Bai, C. Mi, "Eliminate reactive power and increase system efficiency of isolated bidirectional dual-active-bridge DC–DC converters using novel dual-phase-shift control," IEEE Trans. Power Electron., vol. 23, no. 6, pp. 2905–2914, 2008.
- B. Zhao, Q. Song, W. Liu, "Power characterization of isolated bidirectional dual-active-bridge DC–DC converter with dual-phase-shift control," IEEE Trans. Power Electron., vol. 27, no. 9, pp. 4172–4176, 2012.
- B. Zhao, Q. Song, W. Liu, Y. Sun, "Overview of dual-active-bridge isolated bidirectional DC–DC converter for high-frequency-link power-conversion system," IEEE Trans. Power Electron., vol. 29, no. 8, pp. 4091–4106, 2014.
- J. Huang, Y. Wang, Z. Li, W. Lei, "Unified triple-phase-shift control to minimize current stress and achieve full soft-switching of isolated bidirectional DC–DC converter," IEEE Trans. Ind. Electron., vol. 63, no. 7, pp. 4169–4179, 2016.
- F. Krismer, J. W. Kolar, "Closed form solution for minimum conduction loss modulation of DAB converters," IEEE Trans. Power Electron., vol. 27, no. 1, pp. 174–188, 2012.
Results are engineering estimates. Verify your design with datasheets, simulation and measurement.