resistor in series to micro inputs 1

[perts]

i have seen lot of people put like 100k resistors in series on digital inputs to a micro.

why is that? how does affect the digital signal.

 

[Skogsgurra]

The resistor is usually put there to limit possible input current during transients. Without the resistor, the current could zap the substrate diodes (internal diodes connected between input and Vee and Vcc).

The resistors are normally not needed on "inter-IC" signals, but are used on signals that come from outside the board. Or sometimes as level shifters when you go from a higher level system to a TTL or other lower level system.

100 k is a lot of resistance, though. It should not be used on high speed signals since the resulting filter (pin and trace capacitance) will have a cut off frequency in the 100 - 1000 kHz region. A 100 k resistor will reduce (yes, reduce, not enhance) internal EMI immunity since the input is no longer tied to Vee or Vcc as effectively as without the resistor. A paradox is that it enhances external EMI immunity and that the immunity can be further enhanced by adding capacitance between input pin and Vee.

A more normal resistor value would be 10 kohms. See also "10 kohm disease".

 

[MacGyverS2000]

skogs,

Meant to ask this one and never got around to it... you mentioned no need for "inter-IC" signal resistors. At what point do you consider them necessary?

For example, if I wanted to connect multiple latch chips to a micro's data bus, at what point would you suggest inserting resistors to limit the initial current rush?

 

[logbook]

These protection resistors are not used to protect against current transients due to the logic switching. Skogs mentioned external world signals. A 15kV ESD transient is not something you want to stick into an ordinary IC. Modern interface chips are ruggardised to handle this sort of thing but ordinary logic is not. This is why you see clamp diodes, capacitors and all sorts of other analog stuff around digital input lines which go to connectors to the "outside world".

 

[MacGyverS2000]

Egg on my face for not reading his response more closely.

Be that as it may, I've seen arguments for and against using resistors between chips. It seems the only prudent thing to do, but I was wondering others thoughts on the matter. Maybe they're making chips stronger these days and the resistors are no longer required...

 

[BrianR]

You do NOT want resistors between chips except perhaps in odd cases where you are changing from one logic type to another and level translation is needed eg CMOS driving TTL.

Best operation is achieved with a fast risetime input signal. Adding a resistor will, as skoggsgurra said, create a low pass filter, delaying the point of switching and adding a delay of uncertain length.

Low value resistors (100 ohm - 4700 ohm) in series with inputs from the external world are a good idea as stated in earlier posts.

 

[felixc]

Resistors are part of your defense against the outside world, as the operating conditions of ICs are very limited. For Inter IC signals, they can be used for impedence matching purposes. But as MacGyverS2000 states, many ICs, particularly programmable parts, now offer output characteristics with adjustable output impedance and rise time control, removing the need for series resistors in inter IC connections.

 

[bradhuntley]

Another thought.
Some automated production board testers (ICT) simply measure the I/O diodes inside the chips to determine if the chip is ok. They cannot tell the difference between an output diode and the next stages input diode. Adding inter IC resistors in these lines helps the tester detect the presence of these protection diodes and therefore the assumed condition of the chip.