Dynamic IGBT model.
For high power systems and industrial drives, the IGBT replaces the Mosfet as the workhorse for power electronics. The model of the IGBT is build from a Mosfet and BJT. The input of the IGBT is modeled by the Mosfet, while the output is modeled by a BJT. Therefore the parameters are equal to these from the Mosfet and additionally there are parameters describing the on-state voltage drop of the BJT. The IGBT exhibits a tail current while turning off. This tail current is modeled using two extra parameters.
Depending on the application where the IGBT is used, also the model should be chosen carefully. In many cases the ideal IGBT modeled by a on-state voltage drop in series with an on-off resistance controlled from a block diagram is sufficient.
The dynamics can be included by adding a RC or time delay to model the delay caused by the gate charging and some wire inductance to model the overvoltage during switching. The tail current is minimized in modern IGBT's
If more detailed wave forms are required, the complete non-linear IGBT model would be more appropriate.
The non-linear capacitances of the mosfet finally determine the dynamic transient response during switching, while the tail current is typically modeled bya decaying current flowing throught the IGBT after turn off of the IGBT
What model is required
In the first place the static characteristic determines the static (mostly on-state) losses. If the IGBT is switching fast, the model can be approximated by a simple on-off switch with an on-state voltage in series with a resistance.
More detailed modeling would also include the dependency on the gate voltage, where the on resistance is modeled as a non-linear function of VCE and VGE.
This is required if also the driver is modeled in detail and a more accurate simulation result is expected regarding the parasitic components surrounding the IGBT.
Static transfer function
The static transfer function for the IGBT is equal to the static transfer function of hte mosfet, but now the parameter KP is defined as the gain for the IGBT:
| IGBT | VC-VE=Vce+RCE·ICE |
| where | |
| ICE is modulated by the drain current from the mosfet and approximates (KP/2) · (VGE-VTO)2 | |
| RCE is modeled by the parameter RD | |
| VCE is given as a constant parameter VCEon |
Dynamics
The gate capacitance CGS is constant in this model. The other capacitances are non-linear and are modeled by the parameters CGD, CDS, VJ, M and FC.
The tail current is modeled by ITail_over_IForward and TailCurrentDecayTime. It is not modeled by a controlled current source but rather depending on the voltage over the IGBT. In this way the voltage over the IGBT influences the tail current.
| ITail_over_IForward | The parameter ITail_over_IForward defines the initial tail current as a fraction of the forward current. Set this value between 0 and 1. Default it is set to 0.1, meaning that the tail current wil start when the current ICE dropped to 0.1· IForward. |
| TailCurrentDecayTime | The tail current will exponentially decay to zero with a time constant modeled by the parameter TailCurrentDecayTime |
Integrated reverse conducting diode
The reverse conducting diode of the IGBT is always included in the model of the IGBT. The parameters can be set such to make it a very worse diode and also the reverse recovery of the diode can be included. The dynamic model for the diode is based on partly based on spice or on measurement parameters. A great advantage compared to the original Spice model is that the dynamic diode model can also simulate reverse recovery.
The reverse diode is modeled as an ideal diode, but including reverse recovery.
The reverse recovery in the diode model is based on stored charge during the conduction interval. Dependent on the way the diode is forced to turn-off, the reverse recovery current is provided by the stored charge in the diode.
This can be modeled and parameterized in various ways:
- Spice parameter based
- Physical parameter based
- Measured data based
-
Spice parameter based
The original spice parameters are not that bad. The only problem is the model, that was originally designed for small signal diodes. However the parameter TT, modeling the transit time, can be used to model the reverse recovery behavior.
The parameter TT can approximately be chosen as TT=40ns for a diode with a blocking voltage of 100V, to TT=4us for a diode capable of blocking 1000Volt. If TT is set to zero the transit time is approximated from the reverse breakdown voltage BV. - Physical parameter based
The forward storage time in the diode is modeled by the parameter TT, modeling the transit time. The parameter is the same as in the original spice specification, so it can be used to model the reverse recovery behavior.
The parameter TT can approximately be chosen as TT=40ns for a diode with a blocking voltage of 100V, to TT=4us for a diode capable of blocking 1000Volt. If TT is set to zero the transit time is approximated from the reverse breakdown voltage BV.
The time constant with which the reverse recovery is ending is specified by the parameter τrr. If τrr (tau_rr)=0, the diode snaps off very fast. A value greater than zero defines the time constant by which the reverse recovery current decays from IRR towards zero. - Measured data based
If measured data is available, the parameters IF, dIF/dt, QRR and TRR can be specified.
In the datasheet the parameters QRR and TRR are specified for a typical measurement, where IF and dIF/dt are the test circuit conditions.IF The maximum forward current during the conduction of the diode. During conduction the total amount of charge is depending on this value. dIF/dt The gradient of the diode current during turn of, measured at the zero crossing of the diode current. This value is depending on the load circuit connected to the diode and the parasitic inductance in series with the diode. QRR The reverse recovered charge is taken from the specification in the data sheet and is specified for a typical forward current IF and turn off gradient dIF/dt of the forward current TRR The reverse recovery time is taken from the specification in the data sheet and is specified for a typical forward current IF and turn off gradient dIF/dt of the forward current
Losses and Thermal simulation
The IGBT model has a thermal connection that has to be connected to a heat sink model.
The temperature rise due to the conduction, switching and reverse recovery losses is modeled on this connection.
A heat sink is build from the components found in components/library/Heatsink The parameter Rth and Cth model the thermal model from junction to case. The initial temperature of the junction is modeled by the parameter Tth0.
If a more detailed thermal model for the junction to case thermal path has to be build, Rth and Cth simply model the first chip-layer and the following layers are modeled by subsequent thermal models.
Overview of the parameters
The parameters for the IGBT are summarized in the following table. For the parameters that are compatible with the spice diode model the column Spice shows the spice parameter name.
Parameters that do not exist in the spice model are indicated with N.A. Default values for the parameters are given.
| Parameter | Default | Spice | Function |
| VTO | 3 | VTO | Gate threshold voltage |
| KP | 6.4 | KP | Gain for both the Mosfet and BJT |
| VCEon | 2 | N.A. | On-state Collector Emittor voltage |
| RG | 1 | RG | Internal Gate resistance |
| RD | 1u | RD | Internal Collector (Drain) resistance |
IGBT Dynamic Parameters
| Parameter | Default | Spice | Function |
| CDS | 10pF | CDS | Collector Emittor Capacitance maximum value |
| CGD | 1000pF | CGD | Gate Collector Capacitance maximum value |
| CGS | 640pF | CGS | Gate Emittor Capacitance constant value |
| FC | 0.5 | N.A. | Forward bias junction fit parameter |
| M | 0.5 | N.A. | Grading coefficient |
| VJ | 1 | N.A. | Junction Potential |
IGBT Tail Current Parameters
| Parameter | Default | Spice | Function |
| ITail_over_IForward | 0.1 | N.A. | Initial Tail current as fraction ofthe forward current. Set this value between 0 and 1. |
| TailCurrentDecayTime | 50n | N.A. | Time constant for hte exponential decay |
IGBT Wire Parameters
| Parameter | Default | Spice | Function |
| LD | 10nH | N.A. | Collector wire inductance |
| LG | 5nH | N.A. | Gate wire inductance |
| LS | 12nH | N.A. | Emittor wire inductance |
IGBT Reverse Diode Parameters
| Parameter | Default | Spice | Function |
| IS | 10-14 | IS | Saturation current. |
| BV | 106 | BV | Reverse breakdown "knee" voltage. |
| N | 1.5 | N | Emission coefficient. |
| TT | 0 | TT | Forward Storage Time (Transit Time). |
| RS | 1mOhm | RS | On resistance. |
| τrr (tau_rr) | 0 | N.A. | Decay time constant of the reverse recovery current after IRR. If this value is set to zero, the diode has a snappy recovery. |
| IF | 0 | N.A. | Measured maximum forward current. |
| dIF/dt | 0 | N.A. | Measured current gradient during turn off. (Measure at the zero-crossing) |
| QRR | 0 | N.A. | Measured reverse recovery charge. |
| TRR | 0 | N.A. | Measured reverse recovery time. |
Thermal Parameters
| Parameter | Default | Spice | Function |
| Rth | 1 | N.A. | Thermal junction-case resistance. |
| Cth | 0.5m | N.A. | Thermal junction-case capacitance. |
| Tth0 | 25 | N.A. | Initial junction temperature. |