Taking note of asterisked margins shown on the comparison sheet: tAD, tRD, tRR and tDW, it is seen that they are all taken care of by introducing a wait state. The double aster- isked margins deal with the tRV spec on the 8255A-5, 8253-5 and 8279-5 peripherals. tRV is the time from the rising edge of WR or RD to the next falling edge. To allow sufficient time for this spec it is necessary to delay the commands sent to these three peripherals. Enough dead time must occur to make up for the entire negative portion of the margin (for example: 790ns in the 8253-5 medium system). Since in the 8085A-2 every machine cycle is at least 200ns long, 4 ma- chine cycles are sufficient time to allow peripheral control signal recovery (tRV).

One may notice that all of the 8085A instructions take at least 4 T-states (providing a minimum of 800ns) giving ample time to meet this requirement, just by programming one instruc- tion in between every command sent to the peripheral. I/O mapped I/O, which results in using the Input, Output instruc- tions has this delay time built in when moving the data to be transferred into the accumulator. With memory mapped I/O, any instruction that accesses memory for data will provide the time necessary to not violate tRV as a second fetch is performed.

Bus· Loading Considerations· Oecoupllng

For the cost conscious designer it is always helpful to know when buffering is needed and when it is not. How much can I load the 8085A output pins down? To answer this it is helpful to first list the DC requirements of the common types of logic loading and compare this to the capabilities of the 8085A.

 

Maximum

Maximum

 

High-Level

Low-Level

 

Input Current

Input Current

TIL (single load)

40p,A

1.6mA

Schottky or HTIL

40p,A

2.0mA

MOS

10p,A

10p,A

LSTIL (single load)

20p,A

400p,A

The 8085A is capable of an 10L of 2mA (low) and 10H of - 400p,A. With this spec it 'iseasy to come up with the pos- sible combinations of D.C. loading that the designer can use without bufferi ng:

8085A,

LOADSA-2 limiting factor (level)

1 TIL + 1 LSTIL

LOW

1 TIL + 36 MOS*

HIGH

1 SCHOTIKY or 1 HTIL

LOW

40 MOS (various combinations possible)'

HIGH

5 LSTIL

LOW

• Exceeds capacitive loading limit, to be discussed

If a user exceeds these DC loading limitations he must buffer that particular signal. Another factor that the designer must consider is the capacitive load that is seen by the 8085A outputs, which may very well be excessive even if DC loading is not. One may note that even though the 8085A can handle a DC load of 40 MOS devices or 36 MOS + 1 TIL, their collective input capacitances exceed the 150 pF max spec.

The timing specs of the 8085A are guaranteed as long as the 150 pF maximum loading is not exceeded, which includes the wires, components and parasitics. If the user exceeds this value and wants to guarantee his system timing he must either derate the system timings or use buffering.

What if you choose to ignore this limit and say you can live with the performance degradation? First the timing perfor- mance is not all that would degrade, a user must be willing to give up some reliability of his components (All MOS devices have this restraint). This is caused by the excessive switching currents that are needed for this extra loading capacitance. If reliability is not an important consideration, the user can load up to 300 pF on the 8085A bus, but the following correction factors must be used to adjust the timings:

for 150 pF < 300 pF add .13 ns/pF

conversely if less than 150 pF:

for 25 < CL < 150 pF you can subtract .1/ns/pF.

What happens after 300 pF? If the user exceeds this, the noise levels become excessive and problems will result. How much is to much noise? 350 mvolts zero to peak. Prudent designers will always buffer when noise approaches this level, especially in the case of going from orie board to another.

The above takes into consideration the actual specification considerations of when to buffer, but there are also transmis- sion line and noise effects that must be considered. When working with dynamiC RAMs small (20-30 ohm) resistors are commonly put in series in the address lines to help match impedance levels and reduce reflections. Note that this re- sistor should be chosen such that it does not severely degrade the voltage levels of the signal. Long parallel board traces with signals that could adversely affect each other should also be avoided to prevent cross talk problems.

By-passing is very important to prevent intermittent problems which often plague the board designer. Large bulk capacitors should be used at strategic locations on the board to prevent power supply droop. This becomes a major factor when there are many devices that can turn on at once and produce a considerable drain from the power supply (such as burst re- fresh in dynamic RAM).

To help smooth out the current spikes that naturally occur when devices turn on and off, it is recommended to liberally use small capacitors such as the monolithic and other ceramic capacitors which have low inherent inductance. Attached in the 2117 data sheet is a suggested layout of capacitors to effectively bypass the supply lines to ensure proper system operation. Cutting corners here will often times turn around and bite you.

Proper layout is an important consideration. Power supply lines should be well gridded to supply sufficient current to all areas of the board. A strong ground layout is advised to offset noise problems. Remember if the ground plane moves up in voltage because of. excessive charge dumping in a particular area, the supply will drift up correspondingly. Sensing low levels often becomes an intermittent problem when proper ground is not provided.

A1-25

Page 166
Image 166
Intel MCS-80/85 manual Input Current TIL single load 40p,A 6mA Schottky or Htil, 10p,A, TIL + 36 MOS, Schotiky or 1 Htil

MCS-80/85 specifications

The Intel MCS-80/85 family, introduced in the late 1970s, is a seminal collection of microprocessors that played a pivotal role in the early days of computing. The MCS-80 series, initially targeting embedded systems and control applications, gained remarkable attention due to its innovative architecture and flexible programming capabilities.

The MCS-80 family is anchored by the 8080 microprocessor, which was one of the first fully integrated 8-bit microprocessors. Released in 1974, the 8080 operated at clock speeds ranging from 2 MHz to 3 MHz and featured a 16-bit address bus capable of addressing up to 64KB of memory. The processor’s instruction set included around 78 instructions, providing extensive capabilities for data manipulation, logic operations, and branching.

Complementing the 8080 was a suite of support chips, forming the MCS-80 platform. The most notable among them was the 8155, which integrated a static RAM, I/O ports, and a timer, tailored for ease of designing systems around the 8080. Other support chips included the 8085, which provided improvements with an integrated clock generator, making it compatible with more modern designs and applications.

The MCS-85 series, on the other hand, revolves around the 8085 microprocessor, which provided a more advanced architecture. The 8085 operated at clock speeds of up to 6 MHz and came with a 16-bit address bus, similar to its predecessor. However, it introduced more sophisticated features, including an enhanced instruction set and support for interrupt-driven programming. These enhancements made the 8085 especially appealing to developers working in real-time processing environments.

The MCS-80/85 family utilized NMOS technology, known for its lower power consumption and higher performance compared to previous technologies like TTL. The family’s architecture allowed for easy interfacing with a variety of peripherals, making it a favorite for educational institutions and hobbyists embarking on computer engineering projects.

With its robustness, versatility, and affordability, the Intel MCS-80/85 microprocessors laid the groundwork for many subsequent microcomputer systems and applications. The legacy of this powerful family continues to influence modern microprocessor design, emphasizing the importance of reliable architecture in a rapidly evolving technology landscape.