Clock generator with programmable non-overlapping-clock-edge...

Details Clock generator with programmable non-overlapping-clock-edge... Clock generator with programmable non-overlapping-clock-edge...

2002-10-23

2003-11-25

Lam, Tuan T. (Department: 2816)

Miscellaneous active electrical nonlinear devices, circuits, and

Signal converting, shaping, or generating

Phase shift by less than period of input

C327S239000, C327S295000

Reexamination Certificate

active

06653881

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to a programmable clock generator for integrated circuits, very large scale integrated circuits (VLSI), and ultra large scale integrated circuits (ULSI). More particularly, the present invention relates to a programmable clock generator for selecting optimal non-overlapping clocking signals for controlling elements on a chip.
2. Related Art
Most microprocessor chips operate as control driven, synchronous sequential systems. This means the sequence of operations in the system is synchronized by a master clock signal (usually an external clock). This clock signal is usually one of the forms shown in
FIG. 1
; which illustrates a square wave with a 50% duty cycle.
The master clock signal allows system operations to occur at regularly spaced-intervals. In particular, operations on the chip are made to take place at times when the clock signal is making a transition from low-to-high or from high-to-low; rising edge
102
or falling edge
104
, respectively.
Many microprocessor chips have their timing controlled by two or more related clock signals generated by an on-chip clock generator based on the master clock signal.
FIG. 2A
illustrates one such combination utilizing two clock signals identified by &phgr;
1
and &phgr;
2
. This clocking arrangement provides four different edges and three different states per period, compared to only two edges and two states per period provided with a single clock signal as shown in FIG.
1
.
FIG. 2B
illustrates examples of the three possible states for clock signals &phgr;
1
and &phgr;
2
. For elements on the chip to function properly, it is important that edges of clock signals &phgr;
1
and &phgr;
2
are non-overlapping. If the edges overlap there will be more restrictions on data transfer and signal hand shaking.
Additionally, it is equally important that non-overlapping clock edges be evenly distributed to all comers of a chip regardless of the distance which those signals must travel. As chip size increases, clock signals &phgr;
1
and &phgr;
2
have to travel greater distances throughout the chip. This causes clock signals &phgr;
1
and &phgr;
2
to become degraded. As distances increase, rising edges
202
,
206
and failing edges
204
,
208
may become obscured (experience phase shifts and increases in transition times) and can overlap. This phenomenon, sometimes referred to as clock skew, is caused by a number of factors, including: loading, unwanted noise, coupling, capacitance, resistance, inductance and other debilitating effects.
To account for these factors, designers must separate the rising and falling edges
202
,
204
,
206
,
208
of different clock signals (i.e., &phgr;
1
and &phgr;
2
) with a large enough margin of time to allow for clock skew. For instance, failing edge
204
and rising edge
206
must be separated by a minimum temporal distance or amount of time (7) to avoid overlapping states; especially for level-triggering operations in metal-oxide-silicon (MOS) technology. The larger T is, the less likely the chip will fail due to overlapping signals caused by skewing. The wide range of operating environments to which the chip(s) may be subject must be considered in selecting T. Therefore, to provide an adequate margin, manufacturers are forced to select T large enough to provide functionality in a worst-case environment. However, a large T is a significant cycle time constraint. Therefore chip design is not optimized for each environment.
To illustrate this, consider current chip design practices that must account for clock skew by designing a chip with a minimum safety distance T between signals against worst-case conditions. Once T is selected the chip is manufactured and tested. If the chip designer selected a clock speed that has insufficient non-overlapping time, the chip will not function due to overlapping states for some circuits located on the chip. When a chip runs properly, chip designers assume they have chosen the correct frequency, clock states; rise and fall times, and non-overlapping time T. However, chip designers do not know whether a faster clock speed or a smaller T are possible. To find out, chip manufacturers must build entirely new chips with different process parameters, which is inefficient and expensive.
Presently, no programming or tweaking can be performed after a chip is finalized. It is possible to have an on-chip clock generator running at different clock frequencies than external crystal oscillators, but the non-overlapping time of the clock edges generated by the clock generator is fixed by circuit hardware. Therefore, what is needed is a flexible system and method of programming an on-chip clock generator at the manufacturing stage to achieve adjustable as well as optimal non-overlapping times T between clock edges.
At the post-manufacturing stage, environmental conditions, such as heat and cold can also affect clock skewing. If a chip is manufactured under laboratory conditions, it may function properly. However, temperature changes may cause the chip to malfunction due to skewed clock signals. Therefore, what is needed is an on-chip clock generator that can be dynamically programmed to select non-overlapping times T to account for environmental fluctuations while the chip is in an operational environment, such as a processor chip operating in a computer.
SUMMARY OF THE INVENTION
The present invention is directed to a system and method for providing programmable non-overlapping clock generation on a chip. The present invention includes four main embodiments. The first embodiment is directed to the overall operation of an on-chip clock generator. The second embodiment is directed to a hardware programmable clock generation system and method. The third embodiment is directed to a software programmable clock generation system and method. The fourth embodiment is directed to a combination of all three embodiments.
The programmable on-chip clock generator provides two phases of a system clock with non-overlapping edges. The programmability of the clock generator provides flexibility during chip fabrication, and when a chip is functioning in a operational environment.
During the manufacturing phases of chip production, characteristics of the on-chip clock generator are altered to ensure the edges of the two generated clocks do not overlap. This allows the manufacturer to optimize the performance of the chip while the chip is undergoing initial production testing. This feature obviates the need to perform costly and time consuming trial-and-error design and redesign of on-chip clock generators.
Additionally, the present invention provides a technique for optimizing the performance of the on-chip clock generator after the chips have left the manufacturing environment. One feature of the present invention is the ability to adjust clock generation dynamically to account for climatic changes in an operational, or other post-production, environment. This allows chips to be manufactured with wider tolerances and allows operation of the chip to be optimized when the chip is in the operational environment.
Adjustments to the on-chip clock generator during the manufacturing phase are referred to as hardware programming because the manufacturer alters the physical composition of the clock generator. Adjustments to the on-chip clock generator once the chip is fabricated and in the operational environment are referred to as software programming. This terminology reflects the fact that through the use of software commands, the characteristics of the on-chip clock generator can be adjusted to compensate for changes in the operating environment. Programming capability in both cases is accomplished by adding or subtracting delay elements in feedback paths within the clock generator circuit.
Further features and advantages of the present invention, as well as the structure and operation of various embodiments of the present invention, are described in detail below with reference to the accompanying drawing

Affiliated with

Also associated with

Rating

Clock generator with programmable non-overlapping-clock-edge... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Clock generator with programmable non-overlapping-clock-edge..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Clock generator with programmable non-overlapping-clock-edge... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3132507