luke
/
ngspice
mirror of https://git.code.sf.net/p/ngspice/ngspice
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

few typos and comments

This commit is contained in:
dwarning 2020-07-24 16:03:10 +02:00
parent 14a5cd873b
commit 0efc047f5f
1 changed files with 12 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
src/spicelib/devices/hicum2/hicumL2.cpp
View File

@ -8,8 +8,8 @@ Spice3 Implementation: 2020 Dietmar Warning, Markus Müller, Mario Krattenmacher
/*
* This file defines the HICUM L2.4.0 model load function
* Comments on the Code:
* - We use dual numbers to calculate derivatives, this is readble and error proof. => you need to understand that to understand this code
* - The code is targeted to be readbale and maintainable, speed is sacrificed for this purpose.
* - We use dual numbers to calculate derivatives, this is readable and error proof. => you need to understand that to understand this code
* - The code is targeted to be readable and maintainable, speed is sacrificed for this purpose.
* - The verilog a code is available at the website of TU Dresden, Michael Schroeter's chair.
*/
@ -247,7 +247,7 @@ void HICJQ(duals::duald T, duals::duald c_0, duals::duald u_d, double z, duals::
duals::duald calc_hjei_vbe(duals::duald Vbiei, duals::duald T, HICUMinstance * here, HICUMmodel * model){
//calculates hje_vbe
//warpping in a routine allows easy calculation of derivatives with dual numbers
//wrapping in a routine allows easy calculation of derivatives with dual numbers
duals::duald vj, vj_z, vt, vdei_t, hjei0_t, ahjei_t;
if (model->HICUMahjei == 0.0){
return model->HICUMhjei;
@ -279,8 +279,6 @@ void hicum_diode(duals::duald T, dual_double IS, double UM1, double U, double *I
//wrapper for hicum diode equation that also generates derivatives
duals::duald result = 0;
// printf("executed diode");
duals::duald is_t = IS.rpart;
result = HICDIO(T.rpart(), is_t, UM1, U+1_e);
*Iz = result.rpart();
@ -409,7 +407,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
double Vbiei, Vbici, Vciei, Vbpei, Vbpbi, Vbpci, Vsici, Vbci, Vsc;
double Vbici_temp, Vaval;
// Model flags
//Model flags
int use_aval;
//helpers for ngspice implementation
@ -463,7 +461,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
double Temp;
double Tdev_Vrth; //derivative device temperature to Vrth
// below variable has a real part equal to the device temperature and a dual part equal to dTdev/dVrth
//below variable has a real part equal to the device temperature and a dual part equal to dTdev/dVrth
//this is necessary, since for some Vrth, HICUM sets Tdev constant (eg very high self heating beyond 300K)
//then, dTdev/dVrth=0. Else it is equal to 1.
duals::duald Temp_dual;
@ -538,8 +536,6 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
// Q_fC, Q_CT: actual and ICCR (weighted) hole charge
// T_fC, T_cT: actual and ICCR (weighted) transit time
// Derivative dfCT_ditf not properly implemented yet
// feenableexcept(FE_INVALID | FE_OVERFLOW); //debuger catches NANS
std::function<void (duals::duald, duals::duald, duals::duald, duals::duald, duals::duald*, duals::duald*, duals::duald*, duals::duald*)> HICQFC = [&](duals::duald T, duals::duald Ix, duals::duald I_CK, duals::duald FFT_pcS, duals::duald * Q_fC, duals::duald * Q_CT, duals::duald * T_fC, duals::duald * T_cT)
{
duals::duald FCln, FCa, FCa1, FCd_a, FCw, FCdw_daick, FCda1_dw, FCf_ci, FCdfCT_ditf, FCw2, FCz, FCdfc_dw, FFdVc_ditf, FCf_CT, FCf1, FCf2, FCrt;
@ -895,7 +891,6 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
} else{
iavl = 0;
}
// Note that iavl = 0.0 is already set in the initialization block for use_aval == 0 (Markus: not for this lambda!)
return iavl;
};
@ -1086,7 +1081,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
//final calculations afterwards, see later in load where this is called
} //if
}
return Q_pT;
};
@ -1149,7 +1144,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
use_aval = 0;
}
// end of Model_initialization
// end of Model_initialization
/* loop through all the instances of the model */
for (here = HICUMinstances(model); here != NULL ;
@ -1822,10 +1817,6 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
Vbci = model->HICUMtype*Vbci;
Vsici = model->HICUMtype*Vsici;
Vsc = model->HICUMtype*Vsc;
//Dietmar, what about this
// Vxf = model->HICUMtype*Vxf;
// Vxf1 = model->HICUMtype*Vxf1;
// Vxf2 = model->HICUMtype*Vxf2;
if (selfheat) { // Thermal_update_with_self_heating
Temp = here->HICUMtemp+Vrth;
@ -1999,10 +1990,10 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
here->HICUMtf = Tf;
//NQS effect implemented with LCR networks
//NQS effect implemented with gyrator network
//Once the delay in ITF is considered, IT_NQS is calculated afterwards
//Diffusion charges for further use (remember derivatives if this will be used somebday)
//Diffusion charges for further use (remember derivatives if this will be used someday)
Qdei = Qf;
Qdei_Vbici = Qf_Vbici;
Qdei_Vbiei = Qf_Vbiei;
@ -2191,7 +2182,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
// Excess Phase calculation -> hand implementation
if ( nqs){ // && (ckt->CKTmode & (MODETRAN | MODEAC) ) ) { //evaluate nqs network only in TRANSIENT and AC modes.
if (nqs) { // && (ckt->CKTmode & (MODETRAN | MODEAC) ) ) { //evaluate nqs network only in TRANSIENT and AC modes.
Ixf1 = (Vxf2-itf)/Tf*model->HICUMt0;
Ixf1_Vxf2 = 1.0/Tf*model->HICUMt0;
Ixf1_ditf = -Ixf1_Vxf2;
@ -2230,7 +2221,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
Qxf = model->HICUMalqf*model->HICUMt0*Vxf; //for RC nw
Qxf_Vxf = model->HICUMalqf*model->HICUMt0; //for RC nw
Qdeix = Vxf; //for RC nw
Qdeix = Vxf; //for RC nw
Qdeix_Vxf = 1.0;
Qdeix_Vbiei = 0;
Qdeix_Vbici = 0;
@ -2381,17 +2372,7 @@ HICUMload(GENmodel *inModel, CKTcircuit *ckt)
}
// Following code is an intermediate solution (if branch contribution is not supported):
// ******************************************
//if(model->HICUMflsh == 0 || model->HICUMrth_scaled < MIN_R) {
// I[br_sht] <+ Vrth/MIN_R;
//} else {
// I[br_sht] <+ Vrth/rth_t-pterm;
// I[br_sht] <+ ddt(here->HICUMcth_scaled*Vrth]);
//}
// ******************************************
// ********************************************
// For simulators having no problem with Vrth) <+ 0.0
// with external thermal node, following code may be used.
// Note that external thermal node should remain accessible