MODELING AND BUDGETING OF TIMING JITTER AND NOISE - Advanced Signal Integrity for High-Speed Digital Design - page 569

We can then estimate the offset for the inverter as the average offset from the
typical case:

v t, inv ( max ) − v t, inv ( min )

v offset = ( volts ) (13-25)

2

With differential amplifiers the input offsets depend on matching of devices of
the same type. The transconductance parameters for devices of the same type will
be much more closely matched, leading to much smaller offsets in the differential
receiver. The input offset for a differential receiver that is based on an NMOS
source coupled pair is given by

2 (v GS − v TN ) k N
k N

W
W

L
L

v offset = v TN + volts (13-26)

Receiver sensitivity is the amount of input voltage swing that is required to get a
specified output swing [Dally and Poulton, 1998]. Reasonable sensitivity values
for designs on a 0.25- m process are 200 to 250 mV for a CMOS inverter-based

receiver and approximately 10 mV for a differential receiver.

µ

Example 13-3 Input Offsets for Single-Ended and Differential Receivers In
this example we calculate the input offsets for an inverter and a differential
receiver that are designed on a 0.25-

m process, whose parameters and variation

µ

are summarized in Table 13-3.
The worst-case device-length variation, L , does not scale with the length.
As a result, the inverter design can reduce the impact of the length variation on
the input offset by using 0.5- m gate lengths. The NMOS and PMOS design

µ

widths are 1.12 and 4 m, respectively. Table 13-4 summarizes the parameter

conditions for the minimum and maximum inverter threshold cases. Note that the
table includes variation in the supply voltage. This is because equation (13-22)
shows that the inverter threshold depends directly on the supply voltage. Since
the supply voltage may vary by up to

µ

±

10%, we choose to include the variation
in our offset calculation. To calculate the input offsets for the inverter, we use

TABLE 13-3. Example Device Parameters and Variation for a 0.25-

µ

m Process

Parameter Typical Value Variation

V TN (V) 0.43 ± 0.01

V TP (V) − 0 . 4 ± 0.01

k N (A / V 2 ) 10 − 6 10 − 6

115 × ± 12 ×

k P ( A / V 2 ) 10 − 6 10 − 6 30 × ± 3 ×

±

W (

µ

m) 0.025

±

L (

µ

m) 0.025