![]() |
|
If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
Datasheet File OCR Text: |
november 2013 doc id 023670 rev 1 1/24 UM1575 user manual spice model tutorial for power mosfets introduction this document describes st?s spice model versions available for power mosfets. this is a guide designed to support user choosing the best model for his goals. in fact, it explains the features of different model versions both in terms of static and dynamic characteristics and simulation performance, in order to find the right compromise between the computation time and accuracy. for example, the self-heating model (v3 version), which accurately reproduces the thermal response of all electrical parameters, requires a considerable simulation effort. finally, an example shows how the self-heating model works. spice models describe the characteristics of typical devices and don't guarantee the absolute representation of product specifications and operating characteristics; the datasheet is the only document providing product specifications. although simulation is a very important tool to evaluate the device?s performance, the exact device?s behavior in all situations is not predictable, therefore the final laboratory test is necessary. www.st.com
spice model versions UM1575 2/24 doc id 023670 rev 1 1 spice model versions st provides 6 model versions on each part number: ? partnumber_v1c ? partnumber_v1t ? partnumber_v2 ? partnumber_v3 ? partnumber_v4 ? partnumber_tn v1c version it is the basic model (level =3) enclosing c oss and c rss modeling through capacitance profile tables. it is an empirical model, and it assumes a 27 c constant temperature. v1t version it comes directly from v1c version and it also includes the package thermal modeling through a thermal equivalent network and presents two additional external thermal nodes t j and t case . this version hasn't the dynamic link between power mosfet temperature and internal parameters. v2 version it is more advanced than v1c, in fact it takes into account the temperature dependence and capacitance profiles too. it allows the static and dynamic behavior to be reproduced by user at fixed temperatures. by using this version, the simulation of self-heating effects isn't possible. v3 version it comes directly from v2 version and includes the package thermal model through a thermal equivalent network and presents two additional external thermal nodes: t j and t case . in this version, during each transient, the current power dissipation is calculated and a current proportional to this power is fed into the thermal network. in this way, the voltage at t j node contains all the information about the junction temperature, which changes internal device?s parameters. since it is a monitoring node, usually t j pin is not connected (however, to avoid warning messages on this node, the user has to add a floating wire - see figure 1 ). on contrary, t case node has to be connected, either to a constant voltage source vdc representing the ambient temperature or to a heat sink modeled by its own thermal network ( figure 1 ). v4 version it comes directly from v3 version considering the device sited in free air. it includes the package thermal modeling through a thermal equivalent network and presents three additional external thermal nodes: t j , t case and t amb . the voltage at t j node and t case node contains all the information about the junction temperature and case temperature which change internal device?s parameters. since they are monitoring nodes, usually t j and t case pins are not connected (however, to avoid warning messages on this node, the user has to add a floating wire - see figure 1 ). conversely, t amb node has to be connected: to a constant voltage source vdc, representing the ambient temperature. doc id 023670 rev 1 3/24 UM1575 spice model versions 24 tn version it includes the rc thermal network only, which represents the thermal model of the package. its symbol has two pins: t j and t case . figure 1. self-heating model (v3 version) note: t j is a monitoring node and it is not connected; t case is connected either by using a vdc, representing the ambient temperature (on the left-side), or by heat-sink thermal network (on the right-side). c1 r2 r1 tcase d s g tj zth 25 tcase d s g tj zth 25 c2 0 0 tamb tamb gipd081020130954fsr spice model symbol UM1575 4/24 doc id 023670 rev 1 2 spice model symbol for each model version, st provides the appropriate symbol as shown below: figure 2. model symbols tj tcase tcase d s g tj zth tamb tj 1 tcase 2 tcase d s g tj zth v1c version v2 version v1t version v3 version v4 version tn version gipd081020131007fsr doc id 023670 rev 1 5/24 UM1575 spice models - instructions to simulate 24 3 spice models - instructions to simulate in spice simulator, user has to upload the device symbol (.olb file) and the spice model (.lib file) to simulate transistors in the schematic. 3.1 installation in the package model, there are the following files: ? name.lib text file representing the model library written as a spice code; ? name.olb symbol file to use the model into orcad capture user interface. in capture open the menu dialog window "pspice" "edit simulation profile". go to "configuration files" tab and "library" category. select the library (*.lib) path by "browse?" button and click to "add to design" (see figure 3 ) figure 3. capture dialog window to select the library (*.lib) to include the symbol *.olb in the schematic view, open the menu dialog window "place" "part" (or simply pressing "p" key in keyboard) and click the "add library?" button (or pressing alt+"a") to select the file (see figure below). gipd081020131721fsr spice models - instructions to simulate UM1575 6/24 doc id 023670 rev 1 figure 4. capture dialog window to include the symbol (*.olb) finally, you can simulate your circuit choosing the simulation type and parameters. 3.2 typical simulation parameters / options as our models contain many non-linear elements, the standard simulation parameters are often not suitable. the following values can facilitate convergence (set them in dialog window "pspice" "edit simulation profile" "options" tab): note: if the following error message appears during the simulation of one of device models: ==> internal error -- overflow in device..... <== you have to edit the 'pspice.ini' file by inserting the following line behind the headline [pspice] as follows: [pspice] mathexceptions = off ...... do not change any other lines already present gipd081020131724fsr abstol= 1na (best accuracy of currents) chgtol= 1 pc..10 pc (best accuracy of charges) itl1= 150 (dc and bias 'blind' iteration limit) itl2= 20...150 (dc and bias 'best guess' iteration limit) itl4= 20...150 (transient time point iteration limit) reltol= 0.001...0.01 (relative accuracy of voltages and currents) doc id 023670 rev 1 7/24 UM1575 a brief description of self-heating model (v3 version) 24 4 a brief description of sel f-heating model (v3 version) power mosfet?s spice models are behavioral and achieved by fitting simulated data with static and dynamic characterization results. the behavioral model is the best approach because it reproduces the electrical and thermal behavior of the power device through a simplified physical description of the device consisting in a set of equations ruling its behavior at terminal level. the self-heating model (v3 version) includes different analog behavioral models (abm) to describe resistors, voltage and current generator, which are temperature-dependent. a curve fit optimization algorithm extracts the mathematical expression for abm, which yields a good representation of power mosfet?s static and dynamic characteristics. in figure 5 , the self-heating spice model (v3 version) schematic is shown. a brief description of self-heating model (v3 version) UM1575 8/24 doc id 023670 rev 1 figure 5. power mosfet self-heating model schematic rtj6 v_sense3 0vdc g_power in- out+ out- in+ g_rs in- out+ out- in+ rdd11 rtj14 e22 in+ in- out+ out- r_g3 1 1 2 2 g63 g_rmos in- out+ out- in+ rg6 ecap (table) in+ in- out+ out- rdd9 e_e001 in+ in- out+ out- gcdg2 in- out+ out- in+ vread2 0vdc gcdg in- out+ out- in+ 1 2 g_bvdss in- out+ out- in+ g59 g_r_did in- out+ out- in+ rtj15 e_e001 in+ in- out+ out- e2 in+ in- out+ out- rdd10 g60 g_r_did in- out+ out- in+ rtj13 1 1 2 2 g_rmos in- out+ out- in+ rtj16 ecap2 (table) in+ in- out+ out- 0 0 0 0 gate drain source 0 0 0 0 tj 0 0 tc a s e cj6 cref 40 d1x d1y 50 g 3 v2 d1z d1 dd ss r_rmos r_ds sx d d1bvdss1 r_cgs rx1 cgs cref2 402 502 v22 alfa2 rcap2 r_gdiode rd alfa rls rcap ls lg r_gbdss 2g2 ld s r_edep rld 1 d1k r_gpower edep c_cds cj5 cj4 cj3 aa r_r001 ba cj2 r_r003 c cj1 d d_dedep t1 t2 t3 t4 gipd081020131034fsr c rss modeling c rss modeling dc modelling dc diode modeling c oss modeling c oss modelling bvdss modelling recovery diode modelling thermal impedance modeling doc id 023670 rev 1 9/24 UM1575 a brief description of self-heating model (v3 version) 24 4.1 thermal network thermal impedance network represents the basic element, which is featured inside the macro-model. it is used to transform the power dissipated inside the junction into a voltage representing the temperature (t j ). figure 6. physical structure the voltage drop across the network is detected and used as emitter value inside behavioral equations used to model other parameters. thermal impedance is the experimental data required to obtain the cauer model (see figure 6 ). figure 7. thermal impedance profile and cauer model ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? * , 3 ' ) 6 5 * , 3 ' ) 6 5 5 7 m * b s r z h u , 1 2 8 7 2 8 7 , 1 5 7 m 5 7 m 5 7 m 5 7 m 7 m 7 f d v h & |