CY7B991

CY7B992

the FB and REF inputs and aligns their rising edges to ensure that all outputs have precise phase alignment.

Clock skews are advanced by ±6 time units (tU) when using an output selected for zero skew as the feedback. A wider range of delays is possible if the output connected to FB is also skewed. Since “Zero Skew”, +tU, and –tU are defined relative to output groups, and since the PLL aligns the rising edges of REF and FB, you can create wider output skews by proper selection of the xFn inputs. For example, a +10 tU between REF and 3Qx is achieved by connecting 1Q0 to FB and setting 1F0 = 1F1 = GND, 3F0 = MID, and 3F1 = High. (Since FB aligns at –4 tU and 3Qx skews to +6 tU, a total of +10 tU skew is realized.) Many other configurations are realized by skewing both the outputs used as the FB input and skewing the other outputs.

Figure 4. Inverted Output Connections

 

REF

FB

 

REF

 

FS

 

4F0

4Q0

4F1

4Q1

 

3F0

3Q0

3F1

3Q1

2F0

2Q0

2F1

2Q1

1F0

1Q0

1F1

1Q1

TEST

 

Figure 4 shows an example of the invert function of the PSCB. In this example the 4Q0 output used as the FB input is programmed for invert (4F0 = 4F1 = HIGH) while the other three pairs of outputs are programmed for zero skew. When 4F0 and 4F1 are tied high, 4Q0 and 4Q1 become inverted zero phase outputs. The PLL aligns the rising edge of the FB input with the rising edge of the REF. This causes the 1Q, 2Q, and 3Q outputs to become the “inverted” outputs with respect to the REF input. It is possible to have 2 inverted and 6 non-inverted outputs or 6 inverted and 2 non-inverted outputs by selecting the output connected to FB. The correct configuration is determined by the need for more (or fewer) inverted outputs. 1Q, 2Q, and 3Q outputs can also be skewed to compensate for varying trace delays independent of inversion on 4Q.

F

Figure 5. Frequency Multiplier with Skew Connectrions

 

 

 

REF

20 MHz

FB

 

 

REF

 

 

 

 

 

 

FS

 

40 MHz

 

4F0

4Q0

 

 

 

4F1

4Q1

20 MHz

 

3F0

3Q0

 

 

 

3F1

3Q1

80 MHz

 

2F0

2Q0

 

 

 

2F1

2Q1

 

 

1F0

1Q0

 

 

1F1

1Q1

 

 

TEST

 

 

Figure 5 shows the PSCB configured as a clock multiplier. The 3Q0 output is programmed to divide by four and is sent to FB. This causes the PLL to increase its frequency until the 3Q0 and 3Q1 outputs are locked at 20 MHz while the 1Qx and 2Qx outputs run at 80 MHz. The 4Q0 and 4Q1 outputs are programmed to divide by two, that results in a 40 MHz waveform at these outputs. Note that the 20 and 40 MHz clocks fall simul- taneously and are out of phase on their rising edge. This enables the designer to use the rising edges of the 12 frequency and 14 frequency outputs without concern for rising edge skew. The 2Q0, 2Q1, 1Q0, and 1Q1 outputs run at 80 MHz and are skewed by programming their select inputs accordingly. Note that the FS pin is wired for 80 MHz operation because that is the frequency of the fastest output.

Figure 6. Frequency Divider Connections

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REF

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FB

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20 MHz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REF

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FS

 

4Q0

 

 

 

 

 

 

 

 

 

 

 

 

 

10 MHz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4F0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4F1

 

4Q1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3F0

 

3Q0

 

 

 

 

 

 

 

 

 

 

 

 

 

5 MHz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3F1

 

3Q1

 

 

 

 

 

 

 

 

 

 

 

 

 

20 MHz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2F0

 

2Q0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2F1

 

2Q1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1F0 1Q0

1F1 1Q1 TEST

Figure 6 demonstrates the PSCB in a clock divider application. 2Q0 is fed back to the FB input and programmed for zero skew. 3Qx is programmed to divide by four. 4Qx is programmed to divide by two. Note that the falling edges of the 4Qx and 3Qx outputs are aligned. This enables the use of rising edges of the 12 frequency and 14 frequency without concern for skew mismatch. The 1Qx outputs are programmed to zero skew and are aligned with the 2Qx outputs. In this example, the FS input is grounded to configure the device in the 15 MHz to 30 MHz

Document Number: 38-07138 Rev. *B

Page 13 of 19

[+] Feedback

Page 13
Image 13
Cypress CY7B991, CY7B992 manual Inverted Output Connections

CY7B991, CY7B992 specifications

The Cypress CY7B992 and CY7B991 are advanced synchronous SRAM devices designed for high-speed applications, particularly in the field of telecommunications, networking, and high-performance computing. These SRAMs are notable for their ability to operate at high frequencies, making them suitable for systems that require rapid data access and processing.

One of the main features of the CY7B992 and CY7B991 is their support for synchronous operation, which allows for data transfers aligned with a clock signal. This capability significantly enhances performance by reducing access times and increasing data throughput compared to traditional asynchronous SRAMs. With their optimized write and read cycles, these devices can achieve low latency, enabling efficient data handling in real-time applications.

Another key technology utilized in these devices is the use of a 2-port architecture, which supports simultaneous read and write operations. This dual-port design allows for greater flexibility and efficiency in data management, making it easier to implement complex memory architectures in various applications. The architecture also supports burst mode operation, allowing for rapid sequential data access, which is crucial in environments where speed is paramount.

The CY7B992 and CY7B991 feature a wide data bus width, accommodating 32 bits to suit modern data processing needs. Their compact size and ease of integration into existing systems make them popular choices among designers and engineers. Moreover, these SRAMs offer a comprehensive range of voltage and temperature specifications, ensuring reliable performance across diverse operating conditions.

In terms of power management, the CY7B992 and CY7B991 are designed to consume low power while maintaining high performance, making them ideal for battery-operated or energy-sensitive applications. The devices include various power-saving features, such as power-down modes, enabling users to reduce overall system power consumption when the memory is not actively in use.

Overall, the Cypress CY7B992 and CY7B991 are robust, high-speed SRAM solutions that cater to the demands of sophisticated, data-intensive applications. Their synchronous operation, dual-port architecture, and efficient power management characteristics make them essential components in modern electronic systems. As technology continues to evolve, these SRAMs are poised to play a critical role in advancing the capabilities of next-generation devices.