CMOS Op Amp - simplified models

 arctan()/(π/2) versus tanh() ::

PS! :: this model was created for the fast retrieving of the effect and the response of the particular P-MOS input stage ... the output starts behaving weird near the rails !!!

which means the model cannot be used for conventional simulation purposes !!!

---

! Update ::

---

 the (simplified-inexact MCP601) model for conventional use (compared against the Ti-s LMV831) ::

? interestingly the Ti's model has a suspicious behaviour on the input exceeding the supply rails ???

? the P-Spice model of the LMV831 accounts the common mode range ending some 1V below the upper rail so it likely suits better the simulation that needs to take into account that limitation . . . my simplified MCP601 model is brought below ::

Listing of the : MCP601SM.cir ::

* Universal Op Amp * IN NI OUT Vss Vdd .SUBCKT MCP601SM 1 2 3 4 5 PARAMS: Rt=18 gms=1 nms=1 pms=.5 * USE C=30 PF IN MAIN CIRCUIT (CA TO CB). * INPUT R1 5 51 R=Rd R2 5 52 R=Rd M1 55 55 51 51 PcMfT m=pm M2 50 55 52 52 PcMfT m=pm C1 50 55 .2p M3 43 1 50 50 PcMfT m=pm M4 44 2 50 50 PcMfT m=pm M5 43 44 41 41 NcMfT m=nm M6 44 44 42 42 NcMfT m=nm R3 41 4 R=Rs R4 42 4 R=Rs I1 55 4 35u load * DIFFERENTIAL R6 8 17 100k B1 7 17 V=V(43,44) R7 7 17 10k C2 7 17 {srf} * MEDIAN / CLIPPER / Vcp B5 10 0 V=V(5,4)/2-vfz R5 10 0 10k * CP B3 8 0 V=(V(4)+V(5))/2 R8 8 0 10k * MED B11 8 13 V=V(10) B12 12 8 V=V(10) D1 6 12 SjD D2 13 6 SjD R11 11 6 1 * GAIN / SLEW L1 10 8 300n * V.med modified ! B4 10 11 I=(V(5,4)/2+vfz)*tanh(2**bm*V(7,17))/10k R10 11 8 10k C3 11 8 {srf} * OUTP B2 9 8 V=V(11,8) R9 9 8 R=10k C4 9 8 {srf} R12 9 3 R=30 .param gm={gms} .param nm={nms*gm} .param pm={pms*gm} .param Rd={Rt*2/6} .param Rs={Rt-Rd} * .param bm=13.25 .param bm=6.223 .param vfz=0.45 .param srf={0.23n/2} .MODEL NcMfT NMOS(L=2u W=42u KP=35u LEVEL=1) .MODEL PcMfT PMOS(L=2u W=84u KP=7u LEVEL=1) .MODEL SjD D(Is=2.52n Rs=.568 N=1.752 Cjo=4p M=.4 tt=20n Iave=200m Vpk=75 mfg=OnSemi type=silicon) .ENDS

---

[EoP]