Is an external pulldown resistor needed for an IC with an internal pulldown resistor?

Thread Starter

Seric

Joined Oct 20, 2013
18
I am working with the Atmel SAM4E16E microcontroller. I have been reading the datasheet trying to familiarize myself with it. Certain pins are described in the datasheet as being connected to an internal pull-down resistor. For example, the NRST pin is connected to an internal pull-down resistor, and will reset the chip if it is pulled high. So if I connect the reset pin to a pushbutton (normally opened) which goes 3V, then pushing the button will reset the chip.

In this situation, I would normally add a pull-down resistor to the circuit. But since the microcontroller already has an internal pull-down resistor on that pin, do I still want to?

From my understanding, doing so would result in having the two resistors being in parallel, therefore reducing their resistance. But I just would like to confirm this.
 

ScottWang

Joined Aug 23, 2012
7,397
The NRST pin was used a 100K connected to VDDIO, that is a pull high, so it should be actived by pull low. (check 13.6.2 NRST Pin on page 305)

For a cmos IC to using 100K to pull high, that is enough to hold the voltage level, it's no need to adding any resistor.
 

Thread Starter

Seric

Joined Oct 20, 2013
18
You're right, my mistake! I was going by memory when I wrote the post, I should have double checked the datasheet.
Thanks for the help!
 

MikeML

Joined Oct 2, 2009
5,444
The time you would add a parallel, external pull-up or pull-down is for better noise immunity.

If the Reset button was to be say 5m from the processor, then 5m of open wire at the button could act as an "antenna", especially sensitive to stay capacitive coupling to other wires. 100K might not act as a "terminator" to keep the input pin at the correct voltage. Paralleling 1K or 5K at the button end might keep the pin from being perturbed.

Another consideration: Most switches do better at keeping an oxide film off their contacts while switching a few mA compared to switching the tiny current sourced by 100K...
 

Thread Starter

Seric

Joined Oct 20, 2013
18
The time you would add a parallel, external pull-up or pull-down is for better noise immunity.

If the Reset button was to be say 5m from the processor, then 5m of open wire at the button could act as an "antenna", especially sensitive to stay capacitive coupling to other wires. 100K might not act as a "terminator" to keep the input pin at the correct voltage. Paralleling 1K or 5K at the button end might keep the pin from being perturbed.

Another consideration: Most switches do better at keeping an oxide film off their contacts while switching a few mA compared to switching the tiny current sourced by 100K...
I'm not sure that I understand. Paralleling 1K gives a fairly low combined resistance, so wouldn't that defeat the purpose of a pull-down resistor?
In this particular application, the reset button will be more like 5 cm from the processor. Do I still need to worry about the antenna in that case?
 

MikeML

Joined Oct 2, 2009
5,444
I'm not sure that I understand. Paralleling 1K gives a fairly low combined resistance, so wouldn't that defeat the purpose of a pull-down resistor?
In this particular application, the reset button will be more like 5 cm from the processor. Do I still need to worry about the antenna in that case?
Look at this:

Notice that the noise coupling is a tiny 1pF. The Noise source is any unrelated signal with a fast dv/dt.
V(porta) clearly crosses the logic threshold, while V(portb) does not...
 

Attachments

Last edited:
Top