Electronic digital logic circuitry – Interface – Current driving
Reexamination Certificate
1999-03-09
2001-06-19
Tokar, Michael (Department: 2819)
Electronic digital logic circuitry
Interface
Current driving
C326S028000, C326S083000
Reexamination Certificate
active
06249147
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to the field of data transmission, and more particularly to an apparatus and method for increasing the propagation speed of a signal along signal paths in an integrated circuit.
BACKGROUND OF THE INVENTION
The signal propagation delay time increases as the signal path length through a network increases. The propagation delay time is also relatively higher for a signal transmitting across a “heavily-loaded” network (i.e., a network or net with a large capacitive load), since the large capacitive load increases the RC delay time of the propagating signal. One example of a heavily-loaded net is an SRAM word line.
As feature sizes decrease, the metal layers in integrated circuits increase in resistance value. The higher resistance values increase the RC delay for signals transmitting across the nets formed in the metal layers.
The microprocessor cycle time increases if the propagation delay time increases for signals processed by the microprocessor. Additionally, timing requirements in “critical nets” (“critical paths”) may also not be met if signal propagation delay time increases along a critical net.
According to conventional approaches, repeaters, normally in the form of inverters, may be inserted in a long net to increase the signal propagation speed. The repeaters divide the long net into multiple shorter-length nets wherein each repeater drives one of the shorter length nets. In many instances, the desired signal timing (or optimized timing) is attained by insertion of an odd number of inverters. However, the odd number of inverters reverses the polarity (voltage swing) of the propagating signal. To obtain the original polarity of the propagating signal, an additional inverter is inserted in the net so that an even number of inverters is implemented. However, the additional inverter adds delay and, as a result, the desired signal timing constraint (or optimized timing value) may not be satisfied for the net.
Conventional approaches also have the “neighbor effect” problem. The neighbor effect occurs when signals propagating along neighboring nets switch in opposite directions. The neighbor effect leads to a higher effective switching capacitance that also increases the signal propagation delay time.
Accordingly, it is desirable to provide a method and apparatus that can increase the propagation speed of a signal across a net in an integrated circuit and that can overcome the above mentioned deficiencies of conventional approaches.
An important case of the RC delay problem is a net driven by one of the multiple drivers attached to the net. The net cannot use repeaters because repeaters, being unidirectional, disallow the drivers to reach all portions of the net. An example of such net is the result bus of a multiple functional unit. It is desirable to provide a method and apparatus that improve the propagation delay in this case. The conventional approach to solve this case is the use of a bi-directional repeater. However the bi-directional repeater is slow and requires a control circuit to direct the direction of the signal flow. The control circuit is likely to not only be costly in terms of area, but also can introduce speed problems by itself.
SUMMARY OF THE INVENTION
The present invention provides an apparatus for achieving high speed signal propagation across a net driven by one of the multiple drivers in an integrated circuit. The apparatus includes a first driver for driving a signal across the net. A first transition assist driver (TAD) can pull the voltage level at a first node in the net in response to the voltage level at the first node reaching a threshold value as the signal approaches the first node. If the first node is precharged to a voltage level of logic level one, the first TAD can pull the voltage level of the first node to logic level zero. If the first node is precharged to a voltage level of logic level zero, the first TAD can pull the voltage level of the first node to logic level one. When the first TAD pulls the voltage level of the first node, the propagation speed of the signal across the net increases.
In another aspect of the present invention, a second TAD is coupled to the net at a second node and is capable of pulling the voltage level of the second node as the signal approaches the second node. By pulling the voltage level of the second node, the propagation speed of the signal is increased further. Additional TADs may be coupled to other nodes in the net to further increase the propagation speed of the signal across the net.
In another aspect of the present invention, the above threshold value can be adjusted by programming a TAD in accordance with the invention so that the switching speed or, alternatively, the noise immunity of the TAD increases.
In conventional approaches, repeaters (inverters) are used to increase the signal propagation speed across the net. However, an odd number of repeaters reverses the polarity (voltage swing) of the propagating signal. The present invention advantageously avoids the use of repeaters for increasing the signal propagation speed across a net. In addition, a TAD in accordance with the present invention does not invert the polarity of the propagating signal.
In another aspect of the present invention, a precharge scheme is used wherein neighboring nets are precharged to a particular voltage level. This precharge scheme avoids the “neighbor effect” problem of conventional approaches, wherein the neighbor effect occurs when signals propagating along neighboring nets are switching in opposite directions. The neighbor effect problem leads to a higher effective switching capacitance that also increases the signal propagation delay time. In the precharge scheme of the present invention, signals along neighboring nets will not switch in opposite directions. Thus, when a signal in one net is switching in one direction, another signal in a neighboring net is either switching in the same direction or remains at its current polarity.
A second driver may be coupled to the net for driving signals that propagate in a direction opposite to the direction of the signals driven by the first driver. As a result, bi-directional signal transmission can occur in the signal propagation system in accordance with the present invention. Additional drivers can be inserted at various points in the net to permit bi-directional signal transmission across the net.
A TAD in accordance with the invention can pull the voltage level of an associated node in the net, irrespective of the propagation direction of a signal across the net. The TAD can automatically detect the change in the voltage level of an associated node due to a signal propagating across the net and can pull the voltage level of the associated node to increase the propagation speed of the signal.
REFERENCES:
patent: 3819877 (1974-06-01), Blahut et al.
patent: 4012593 (1977-03-01), Yamaguchi
patent: 4091386 (1978-05-01), Luedtke et al.
patent: 4096566 (1978-06-01), Borie et al.
patent: 4154978 (1979-05-01), Tu
patent: 4387466 (1983-06-01), Sire
patent: 4512016 (1985-04-01), Fulcomer, Jr. et al.
patent: 4584441 (1986-04-01), Chance et al.
patent: 4591669 (1986-05-01), Duttweiler et al.
patent: 4598410 (1986-07-01), Smith et al.
patent: 4604576 (1986-08-01), Barrault
patent: 4628157 (1986-12-01), Chance et al.
patent: 4635260 (1987-01-01), Sestan
patent: 4658396 (1987-04-01), Barden
patent: 4779013 (1988-10-01), Tanaka
patent: 4847831 (1989-07-01), Spiesman et al.
patent: 5414312 (1995-05-01), Wong
patent: 5475605 (1995-12-01), Lin
patent: 5497108 (1996-03-01), Menon et al.
patent: 5546083 (1996-08-01), Elam
patent: 5568064 (1996-10-01), Beers et al.
patent: 5572441 (1996-11-01), Boie
patent: 5578939 (1996-11-01), Beers et al.
patent: 5604450 (1997-02-01), Borkar et al.
patent: 5629860 (1997-05-01), Jones et al.
patent: 5649170 (1997-07-01), Chappell et al.
patent: 5655107 (1997-08-01), Bull
patent: 5659575 (1997-08-01), Vielhaber et al.
patent: 5668482 (1997-09-01), Roskell
patent: 5715405 (1998-02-01), McC
Patwa Nital P.
Vinh James
Fenwick & West LLP
Fujitsu Ltd.
Le Don Phu
Tokar Michael
LandOfFree
Method and apparatus for high speed on-chip signal propagation does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for high speed on-chip signal propagation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for high speed on-chip signal propagation will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2543938