Arif:

It would be helpful if you provided some sort of schematic or table that details how all of these devices are interconnected to the Port 0 of your microcontroller. The reason I say this is because the AT24C01 is an I2C device. The AT25010A and the AT93C46 are SPI type devices which would need separate chip select lines. Lastly the LCD module requires yet another interface with either 4-bit or 8-bit interface if the module follows the old standard Hitachi character mode LCD interface scheme.

It is possible to envision an implementation specific interface between all these devices that could share some of the port lines in common. Of course this is taking into consideration that all the devices are powered up at the same time, that the port pins assignments are made in a carefully considered way and that the firmware in the MCU only accesses one device at a time. It is also necessary to very carefully consider any output pins of the attached devices to ensure that there is not any active contention between parts.

Now let us consider the issue if one or more devices are powered down. The powered down device is highly likely to cause problems for the signal lines that it is connected in common with other devices. The powered down device may have input protection diodes between the input pin and the chip VCC pin and between the input and the chip GND pin. These diodes are designed to clamp input pin voltages to between the part voltage rails. When such device is powered down the clamp pin to the device VCC pin gets forward biased when the signal pin from the MCU is high. This ends up causing two issues; the signal pin ends up getting clamped to something like about 1.2 volts and attempts to try to power the "off" chip from inside. For some types of MCU pins that can source a large amount of current in the high level this clamping and powering of the "off" chip can be overcome and the signal line made to appear to work normally but this is not recommended as it can stress the MCU pin and could lead to undue stress of the "off" chip.

There are some types of devices that contain a different type of input protection diode. Instead of a pair of diodes as described before the input is connected instead to a zener diode type component designed to clamp inputs to a voltage that is typically in the range of 5.6 to 7 volts. This type of device can generally be safely connected in common to other devices on the same signal pins and then be in the "off" state without causing clamping action on the signal line.

There is sometimes a work around for "off" devices that implement the first type of input protection diodes as described above. It may be possible to insert a series resistor in the signal path right in front of the "off" device which isolates it from the rest of the devices on the common signal line. If there are multiple "off" devices on the same line each would likely require its own series resistor. This resistor scheme would work OK with CMOS or NMOS type "off" devices but will not work well at all for older type bipolar type chips like TTL, LSTTL logic chips. The series resistor idea usually does not work if the signal connection at the "off" device is to an output pin on the device because the part may not be able to drive back properly to the MCU through the resistor when it is not an "off" device. For cases where the series resistor is just going to an input on the "off" device there may be input logic threshold problems or signal integrity loss due to the stray capacitance on the node after the series resistor.

In conclusion it can be recommended that making devices in common on MCU pins takes careful consideration under cases where all devices are powered. It takes some analysis and knowledge of how all the parts work together in actual operation with the MCU firmware. In the case of shared signal lines going to some "off" devices it takes intimate knowledge of how those devices behave in their powered down state. The types of input protection diodes in use on a particular part are often not shown in device data sheets and could even vary between functionally similar devices from different vendors. Sometimes it may take experimentation to learn how certain devices behave and whether use of series resistors will provide suitable performance when the "off" device is called into duty.

Michael Karas

Edit: Let me add - If an "off" device's high side input protection diode tries to clamp the signal line voltage to the ~1.2 volt level that I mentioned it is possible that the clamp level voltage may rise higher as the "off" device tries to power itself up internally. Due to this behavior it is possible to see the clamping voltage level which may rise to even over 2 volts and becomes dependent on source impedance of the signal line coming from the MCU. Karas