Proper Handling of Reset Pin (When you have no intention of resetting)

Other Hardware 3rd Party Boards
1

Looking at other examples of how the reset pin is handled on AVRs I see the following methods:

  1. Leave it unconnected
  2. Connect to 5V with a 10K resistor
  3. Connect to 5V with a 10K resistor and decouple with .1uF cap

In the particular project I'm working on I have no intention of resetting my chip while it's in circuit; I keep it in a socket and remove it to program. This is also the first time I've built a circuit destined to be used in a small vehicle and I'm trying to make sure I won't get messed up by engine noise.

The datasheet for the ATTiny85 states that the reset pin has an internal pullup when not configured as an input. In that respect it seems like option 1 would be the way to go (noise can't interfere with it?) but I'm open to suggestions on what would be the best practice.

2

Chagrin:
In that respect it seems like option 1 would be the way to go (noise can't interfere with it?) but I'm open to suggestions on what would be the best practice.

The internal pull-up on RESET is weak; 30 K to 60 K. Noise can interfere with it; cause reset to trigger. I know from experience that a nearby (one to six feet away) static discharge can trigger a reset.

Were I in your shoes, I would use a strong pull-up resistor; something like 1 K.

3

There is an internal pull-up but experience has shown that (particularly when programming) it is more reliable with the 10K external pull-up. I've never seen it decoupled myself.

Here's a comparison ... the TV-B-Gone schematic:

877×884 12.2 KB

Now that does not (appear to have) a pull-up resistor.

4

Chagrin:
3) Connect to 5V with a 10K resistor and decouple with .1uF cap

I wonder if the example you're referring to there has a cap connected not to GND for decoupling, but to the RTS pin on an FTDI connection for programming?

Geoff

5

That might depend a lot on the wires/PCB tracks attached to the pin.

If the pin isn't connected to anything then the internal pullup might be enough (snip the pin off? :slight_smile: )

6

in tests i was hired to perform connecting ANYTHING to that pin always made things worse. caps, resistors, short to vcc, all caused more trouble in noise tests. even unconnected ic pad on the board picked up pulses from his welding equipment. in fact for ultimate protection we ended up cutting off the reset pin completely but that was extreme.

ops #1 option is #1 in my book.

7

john1993:
for ultimate protection we ended up cutting off the reset pin completely

Hah! I was right! :slight_smile:

8

where were you when we needed you? haha

9

strykeroz:
I wonder if the example you're referring to there has a cap connected not to GND for decoupling, but to the RTS pin on an FTDI connection for programming?

Just a few random examples:

http://plischka.at/Akkuwarner.html
http://elbastl.sweb.cz/clock.htm
http://hobby-electrons.sourceforge.net/projects/t13-dev-board/index.html

10

Atmel's guidelines...

...section 3 Connection of RESET pin on AVRs. Resistor, capacitor, and diode are on the list.

If you won't be programming in-circuit, I assume you can just tie RESET to VCC.

11

john1993:
...short to vcc, all caused more trouble in noise tests...

Doesn't that imply a problem with the power supply? Or an isolation problem?