# 0 ohm resistor selection for single-point grounding

1. Jan 1, 2009

### rogerdi

Hi all, I want to use a 0 ohm resistor for single-point grounding to connect analog/digital ground together. Since the current could be pretty large (several hundred mAs), I'm considering how to choose the resistor for its power consumption. Yes it's "0 ohm", which makes it seem that any size would be ok (DIP/0805/0603/0402/...). But is this ture in the real world?

Happy new year!

2. Jan 1, 2009

### Staff: Mentor

Welcome to the PF, and excellent question. There should be a spec for the resistance of the 0-Ohm resistors that you are considering, because you are right, they are just very small resistance values, but they are non-zero. And many of us rely on bounding that resistance in our 0-Ohm jumper applications. I've gotten so many questions back from our Manufacturing Engineering folks on "How can there be a tolerance on a 0-Ohm jumper?!"

Check the datasheets, and hopefully they will give you what you want for specs. If not, call the tech support folks at the manufacturing companies, and ask them for clarification. If that still doesn't help, post some links to the datasheets here, and we can follow up with some suggestions.

However, I have to take issue with what you say about the currents that you are anticipating flowing between your analog and digital domains. Not good. What is the source of these currents, and how can you use better partitioning or guarding to reduce them?

The goal of Star Grounding is to reduce shared impedance to zero for return currents in the different power domains. That can mean more than just separating the grounds sometimes....

3. Jan 2, 2009

### fidel

Thank you Berkeman for the information. Yes I will turn to the datasheets for details. Is it enough that I find the max possible resistance, with the calculated P=I^2*R to select the appropriate size? I used to use inductances for the separation, but in a tiny system, 0 ohm resistors should be the best. Am I right?

I might have made some mistakes in my expression. Actually it should be Star Grounding, using 0 ohm resistors to connect DNGD and AGND to power ground entry.

4. Jan 2, 2009

### Staff: Mentor

The max resistance is usually all you are concerned with for "0-Ohm Jumpers", although I suppose that power dissipation could be a consideration in some unusual applications.

It sounds like you know what "Star Grounding" means, and what it is used for. In sensitive analog+digital systems (like RF transceiver applications), the separation, star configuration, no shared return impedances, and capacitive guarding, can make the difference between a product that works, and one that cannot meet the performance specs that are needed to be a real product.

5. Jan 6, 2009

### yungman

You have to be very careful on cutting the ground plane. I did a lot of very sensitive mixed signal designs and PCB layout. In the first pass, I do not separate ground plane.

I don't know anything about your application so I cannot even start to say anything. If you mix digital and analog circuit on a PCB. Make sure you use a ground plane. Return current always follow the path of the trace back to the source. This is called image current. Your placement of the signal traces is as important as anything. This is a big topic on signal integrity which I have done a lot.

If you really have to have large current and your application is not what I described. I would use a large wire winding on a toroid with a few turns and connect the two grounds together. If you are only dealing with low frequency, you might even want to use a choke transformer. So you can supply large current due to the low resistance of the fat wire but an open circuit in AC point of view. In this way you jumper don't have to carry the current. Hope that help. As I have said, personally, I never separate grounds, just solid ground planes. Noise generate mainly due to the return image current get interupted.

Anything more, you have to give example. I don't think we can really give you useful answer until that point.

Last edited: Jan 6, 2009