IEEE Presentation on Properties of digital switching currents (CMOS)

Submitted By: Prabhanshu Sharma

 CMOS circuit performance is generally determined by equivalent RC delays

Equivalent resistance of driver circuit (N and P devices) driving a receiver circuit plus wire

Circuit Characterization and Performance Estimation

Load capacitance is comprised of Cgate + Cwire + Cdiffusion + Cparasitic

Cgate is the total parallel gate capacitance of receiving circuit(s) Cwire is the total wiring capacitance of the interconnect line (metal or poly) Cdiffusion is the total combined PN junction capacitance of the driving circuit Cparasitic is the total equivalent capacitance of the internal integrated wire, etc.

Driver resistance consists of some equivalent combination of pull-up and pull-down devices
Rp is equivalent resistance of the PFET pull-up device Rn is the equivalent resistance of the NFET pull-down device Metal wire resistance may or may not be important depending on length of net Polysilicon gate resistance may or may not be important depending on length of poly line

 Define:

CMOS Inverter Switching Characteristics

Rise time tr = time required for a node to charge from the 10% point to 90% point Fall time tf = time required for a node to discharge from 90% to 10% point Delay time td = delay from the 50% point on the input to the 50% point on the output Falling delay tdf = delay time with output falling Rising delay tdr = delay time with output rising

Delay Time Derivation: NMOS Discharging Cload

Assume Vin switches abruptly from VOL to VOH (VOL = 0 and VOH = VDD for CMOS) We are interested in the delay time for Vout to fall from VOH to the 50% point, i.e. to the value 0.5 x (VOH + VOL), = VDD for CMOS
For Vout between VOH and VOH VTN, the NMOS is in saturation Integrate Cload dv = I dt between to and t1 IDS = kn (Vin VTN)2 t1 to = 2 Cload VTN/kn (VOH VTN)2 For Vout between VOH VTN and VOL, the NMOS is in the linear region Integrate Cload dv = I dt between t1 and t1 IDS = kn VDS (VGS VTN VDS)

CMOS Inverter Propagation Delay Summary

Summing the two delay components from the previous chart, we obtain the expression (at left) for the propagation delay (high-to low) for an NMOS transistor discharging CL For CMOS, VOH = VDD and VOL = 0, so the propagation delay (output falling) becomes the expression shown (at left) A similar expression (left) is obtained by considering the derivation of charging Cload with the PMOS transistor when the input abruptly falls from VDD to 0 and the output rises (low-to-high propagation delay)

The above expressions for propagation delay can be reduced to the following simplified form by defining n = VTN/VDD for falling output (n = |VTP|/VDD for rising output), and = N for falling output (= P for rising output) : P = k CL/VDD where k = [2n/(1-n) + ln (3 4n)]/[1-n] = 1.61 for n = 0.2

CMOS Inverter Fall Time & Rise Time Derivation

Discharge Transient: (p device OFF)

N Saturation region (0.9Vdd>vout>Vdd-Vtn)

CLdv/dt + n (Vdd Vtn)2 = 0 t1 = 2CL(Vtn 0.1Vdd)/n(Vdd Vtn)2 N Linear Region (Vdd-Vtn>vout>0.1Vdd) CLdv/dt + n v (Vdd Vtn 0.5 v) = 0 t2 = (CL/nVdd)[{ln (19-20n)}/{1-n}] where n =
Vtn/Vdd

The combined fall time tf is given by tf = k CL/nVdd where k = [2/(1-n)][(n-0.1)/(1-n) + 0.5 ln(19-20n)] k = ~3.7 for n = Vtn/Vdd = 0.2 Charging Transient: (n device OFF)
Due to the symmetry of CMOS, a similar expression is obtained for rise time where n is replaced by p = |Vtp|/Vdd Equal CMOS rise and fall times requires n = p due to the difference in e & h mobilities.

Resistance of a Wire (Rectangular Geometry)

Resistance of a uniform slab:

R = (l/A) = (/t) (L/W) where is the resistivity in ohm-cm, t is the thickness in cm, L is the length, W is the width, and A is the cross-sectional area Using the concept of sheet resistance, R = Rs (L/W) where Rs is called the sheet resistance and given in ohms per square
Rs = / t

Apply to metal wire, poly line, or even a diffused P+ or N+ area of sufficient length

Resistance of an FET transistor (linear): R = Vds/Ids = 1/[(Vgs Vt 0.5 Vds)] As Vds 0, Rds 1/[(Vgs Vt)] = k(L/W) where k = 1/[Cox(Vgs Vt)]

Sheet Resistance for Metal, Poly, Diffusion Conductors

depends on thickness of the conductor (Rs = /t) typical thickness assumed

Sheet resistance for various conductors used in S/C fabrication is given below

Aluminum and copper metal interconnect values given for 0.18 um technology

MOSFET Device Capacitance

Cgs and Cgd are lumped at gate-to-source and gate-to-drain, respectively Cgb (or Cgx) is gate-to-substrate (or gateto-well) capacitance Csb (or Csx) and Cdb (or Cdx) are the source-to-substrate and drain-tosubstrate capacitances and are due to reverse-biased PN junctions of source/drain diffusions. Regions of operation:
OFF
Cgs and Cgd are zero (or very small due to gate overlap capacitance); Cgb is Cox A in series with Cdepl

Linear Cgs = Cgd = (1/2) Cox A; Cgx = ~ 0 Saturation Cgs =~ (2/3) Cox A; Cgd =~ 0; Cgx =~ 0 where Cox = oSiO2/tox

MOS Gate Capacitance

MOS gate oxide capacitance can be divided into three regions:

(a.) Accumulation occurs when Vg is negative (for P material). Holes are induced under the oxide. Cgate = Cox A where Cox = SiO2o/tox (b.) Depletion occurs when Vg is near zero but < Vtn. Here the Cgate is given by Cox A in series with depletion layer capacitance Cdep (c.) Inversion occurs when Vg is positive and > Vtn (for P material). A model for inversion in comprised of Cox A connecting from gate-tochannel and Cdep connecting from channel-to-substrate.

(d.) shows a plot of normalized gate capacitance versus gate voltage Vgs
High freq behavior is due to the distributed resistance of channel

Normalized Experimental MOS Gate Capacitance Measurements vs Vds, Vgs

Explanation:

Shown at the left are plots of normalized gate capacitance versus Vds with Vgs Vt as the parameter for the curves
Top figure is for a long channel MOSFET Bottom figure is a short channel MOSFET Note that for Vds = 0, the total gate capacitance Cox A splits equally to the drain and source of the transistor. For Vds > 0, the gate capacitance tilts more toward the source and becomes roughly 2/3 Cox A to the source and 0 to the drain for high Vds
Higher Vgs Vt forces this tilting to occur later, since the device is linear up to Vgs Vn = Vds

For short channel devices, the fringing fields from gate to source and drain are more important and add a component to the total normalized cap (called overlap cap)

MOS Transistor Gate Capacitance Model

A model of the MOSFET gate capacitance is given at the left with representative values below for OFF, linear, and saturated regions of operation Cox (per unit area) for tox = 100A is given by = 3.5 E-7 F/cm2 = 3.5 fF/um2 For a unit-sized transistor (min L and min W with a single contact), W = 4 and L = 2, giving Cgate = 28 fF for = 1 um.

Cox= SiO2o/tox = (3.9 x 8.85E-14 F/cm)/(100 x 1E-8 cm)

Device Junction Capacitance: Area + Perimeter Terms

PN junction capacitance is given by both an area term and a perimeter term (as shown by equation at left). SPICE models allow specification of the source & drain area and perimeter
SPICE computes the total capacitance for each source and drain junction

Junction capacitance has a voltage dependency (reversed-bias junction) where m = 2 for an abrupt junction and m=1.5 for a linear-graded junction.

Cj = Cjo[1 Vj/Vb] -m

 For wide conductors with W >> H, capacitance to substrate (of any ground plane) can be determined as a parallel plate capacitor C = A/t where A is the planar area of the wire and t is the thickness of the oxide For most real conductors in todays IC technology, fringing fields contribute a major part of the line capacitance and must be included in the capacitance calculations.
For W =~ H (below), fringing fields add more than the parallel plate portion to the total line capacitance.!

Capacitance of the Interconnect Metal Wires

Metal Line Capacitance with Fringing Effect

Solution by Yuan and Trick given at right assumes the wire can be approximated by a piece of metal with thickness t and two rounded edges
parallel plate portion with width equal to W t/2 fringing term due to two hemispherical ends with exact solution to field equation

Example for wire of width W=0.30 um, thickness t = 0.30 um, and dielectric thickness h =0.35 um, gives a result C = 0.13 fF/um where the fringing part is over of the total capacitance.

Capacitance of Layered Multiple Conductors

Structure of Interconnect:

Layers 1 and 3 run along page

each can be assumed to act as a ground plane (solid plane)

Layer 2 runs out of the page

Equivalent capacitances per unit length:

Ctotal = C21 + C23 + 2 x C22 C21 is from center conductor to lower ground plane (layer 1) C23 is from center conductor to upper ground plane (layer2) C22 is from center conductor to adjacent wire on the right C22 also occurs from center conductor to adjacent wire on right assuming spacings are symmetrical

Equations at left give capacitance from center conductor to one or both ground planes

Capacitance of Layered Multiple Conductors

Equations at left give capacitance per unit length between center conductor and adjacent conductor (C22) for both cases
One ground plane only (layer 1) Two ground planes (layers 1 & 2)

Parameters:
T = wire thickness H = interlayer dielectric thickness S = wire spacing W = wire width

Interconnect Wire as a Distributed RC Network

Delay in a distributed RC ladder network is given by

n = R C n (n+1)

where R and C are the series resistance and nodal capacitance for each section, and n is the number of sections. For n large, the above expression reduces to

= r c l2
where r and c are the resistance and capacitance per unit length, and l is the total length of the wire. Note that interconnect delay is proportional to the square of wire length.

The Elmore Delay Estimation Technique

For a step input Vin, the delay at any node can be estimated with the Elmore delay equation tDi = Cj Rk For example, the Elmore delay at node 7 is give by
R1 (C1 + C2 + C3 + C4 + C5) + (R1 + R6) C6 + (R1 + R6 + R7) (C7 + C8)

Model for Distributed RC Line with Capacitive Load

A simple model for a distributed RC interconnect wire can be represented as shown at left:
driver circuit with equivalent Rdrvr Receiver circuit with capac load Cload Interconnect with total resistance Rwire and total capacitance Cwire The total delay of the wire and load can be written as t = (Rdrvr+ Rwire)(Cwire+ Cload) RwireCwire The equivalent circuit at the bottom left gives identical result to above RC model given that delay = rcl2 = RwireCwire

Thank You!!