EE 420L Engineering Electronics II Lab - Lab 8
Mario Valles

April, 17 2015

vallesm@unlv.nevada.edu

---

In this lab you will characterize the transistors in the CD4007 and generate SPICE Level=1 models. Assume that the MOSFETs will be used in the design of circuits powered by a single +5 V power supply. In other words, don't characterize the devices at higher than +5 V voltages or lower than ground potential.

• Experimentally generate, for the NMOS device, plots of: 

• ID v. VGS (0 < VGS < 3 V) with VDS = 3 V 

NMOS	PMOS

• ID v. VDS (0 < VDS < 5 V) for VGS varying from 1 to 5 V in 1 V steps

• NMOS	PMOS

  
  

• ID v. VGS (0 < VGS < 5 V) with VDS = 5 V for VSB varying from 0 to 3 V in 1 V steps.
  
• 
  

• Assuming that the length of the NMOS is 5 um and its width is 500 um calculate the oxide thickness if Cox (= C'ox*W*L) = 5 pF.

• NMOS	PMOS
  C'ox=Cox/(L*W)=5pF/(500um*5um)=2 mF/m   TOX=Eox/C'ox=8.85e-12*3.97/(2e-3)=18nm	C'ox=Cox/(L*W)=5pF/(500um*5um)=2 mF/m   TOX=Eox/C'ox=8.85e-12*3.97/(2e-3)=18nm

  
  
• From this measured data create a Level = 1 MOSFET model with (only) parameters: VTO, GAMMA, KP, and TOX.

• Transistor	NMOS	PMOS
  VTO	1	1.3
  GAMMA	0.8	0.5
  KP	3.2u	3.6u
  TOX	18nm	18nm

  NG = dVTO/dVSB = 0.8   PG = DdVTO/dVBS = 0.5
  
• KPN = 2ID/(W/L(VGS-VTO)^2) = 2*1.6e-4/(100(2-1)^2) =3.2
• KPP = 2ID/(W/L(VGS-VTO)^2) = 2*1.6e-4/(100(2-1.3)^2) =3.2
• 
  
• 
  
• Compare the experimentally measured data above (the 3 plots) to LTspice-generated data (again, 3 plots) and adjust your model accordingly to get better matching.

• TRANSISTOR	NMOS	PMOS
  DVG
  DVD
  DVB

  
  
• Experimentally, similar to what is seen on the datasheet (AC test circuits seen on page 3 of the datasheet), measure the delay of an inverter using these devices (remember the loading of the scope probe is around 15 pF and there is other stray capacitance, say another 10 pF).
• We got 36ns, however the picture did not showed, we erased the other pictured.
  
• 
  
• Using your model simulate the delay of the inverter and compare to measured results. Adjust your SPICE model to get better matching between the experimental data and the measured data.
• The delay in the simulation is not appreciated as much is in the experiment having only a delay of about 10ps
  

CKTs used