I have read about how a ground plane is not a perfect ground and i should connect output gnd to a quiet gnd point then i have a question:
Assume that the rectangular is a solid ground plane. If a current of 1A must flow from gnd load (point B) back to gnd source (point A) and assume that the resistance between A and B is 1R, then a voltage of 1V is developed between A and B. But then what is the voltage different between C (an arbitrary node in ground plane) and A or B? If we measure between A and B we will get 1V but between A and C there is resistance but no current flow so voltage difference between A and C, also B and C will be 0V different, isn't that mean between A and B there will be 0V different?
Assume that the rectangular is a solid ground plane. If a current of 1A must flow from gnd load (point B) back to gnd source (point A) and assume that the resistance between A and B is 1R, then a voltage of 1V is developed between A and B. But then what is the voltage different between C (an arbitrary node in ground plane) and A or B? If we measure between A and B we will get 1V but between A and C there is resistance but no current flow so voltage difference between A and C, also B and C will be 0V different, isn't that mean between A and B there will be 0V different?
No. Current flow in a ground plane depends on various factors. In the case of DC current the actual current path would not only be directly between points A and B. Some current would take that straight line path, but some would spread out and take a longer path. Some current would flow past point C. At much higher frequencies and in particular when there is a signal trace on top layer right above the ground plane then things are different.
Some resources:
electromagnetism - Why is the return current through a printed circuit board's ground plane concentrated below the circuit trace? - Physics Stack Exchange
PCB Return-Current Distribution in a Microstrip Line - In Compliance Magazine
How to Design Your PCB Return Current Path | NWES Blog
Reducing Ground Bounce in DC-to-DC Converters—Some Grounding Essentials | Analog Devices
Some resources:
electromagnetism - Why is the return current through a printed circuit board's ground plane concentrated below the circuit trace? - Physics Stack Exchange
PCB Return-Current Distribution in a Microstrip Line - In Compliance Magazine
How to Design Your PCB Return Current Path | NWES Blog
Reducing Ground Bounce in DC-to-DC Converters—Some Grounding Essentials | Analog Devices
Yes i know that but for dc, most of current will flow in straight line, current spread from that line is very little and can be ignore or u can imagine that point C is miles away from A and B so no current flow here.
For high frequency it's very easy to make current flow from B to A, just make a straight line for power supply to go from A to B in another layer then the question in my first post is still remain.
Think of it this way:
For lower frequencies, the current flows in the path of least resistance.
For higher frequencies, the current flows in the path of least inductance.
This means that the return current will attempt to flow back underneath
the outgoing current, so don't impede it.
How to Handle Current Return Path for Better Signal Integrity
Ground Considerations for PCB Layout of Mixed Signal Designs Part 1 - YouTube
For lower frequencies, the current flows in the path of least resistance.
For higher frequencies, the current flows in the path of least inductance.
This means that the return current will attempt to flow back underneath
the outgoing current, so don't impede it.
How to Handle Current Return Path for Better Signal Integrity
Ground Considerations for PCB Layout of Mixed Signal Designs Part 1 - YouTube
Not exactly. If C is miles away then the resistance from A to C and then from C to B still forms a voltage divider. It may be a high impedance voltage divider, so you may need to measure it with a high input impedance meter. That voltage divider is in parallel with the divider directly between A and B, and with the infinite number of other voltage dividers in parallel with that. Still there some voltage divider though point C. Also, the resistance from A to be also depends on the thickness of the copper. There will be less overall resistance between A and B if current flows through the infinite number of parallel dividers in the ground plane. As the current paths of more distant parallel dividers gets longer, then the resistance of those paths gets higher and the current that flows in each infinitesimal width path will be proportionally less.
Last edited:
If the ground plane were infinite, the voltage on a line in the center of A and B perpendicular to the line between A and B would be exactly half the voltage between A and B.
You drew C on that line, so if the plane were infinite, it would be half.
However, since you placed all three slightly to the left of the center of the rectangle, the overall half voltage line will be slightly bent.
As Rayma said, the current takes the path of least impedance. As Mark pointed out, the ground plane current will try to follow the send path as it is the path of least impedance. This of course depends on frequency, as both pointed out.
jn
You drew C on that line, so if the plane were infinite, it would be half.
However, since you placed all three slightly to the left of the center of the rectangle, the overall half voltage line will be slightly bent.
As Rayma said, the current takes the path of least impedance. As Mark pointed out, the ground plane current will try to follow the send path as it is the path of least impedance. This of course depends on frequency, as both pointed out.
jn
Some more, if you like those.
Gaps in Return Plane Demo - YouTube
Ground Considerations for PCB Layout of Mixed Signal Designs Part 2 - YouTube
What Every PCB Designer Should Know - Return Current Path (with Eric Bogatin) - YouTube
PCB layout for switched-mode power supplies (SMPS) | TI.com Training Series
Gaps in Return Plane Demo - YouTube
Ground Considerations for PCB Layout of Mixed Signal Designs Part 2 - YouTube
What Every PCB Designer Should Know - Return Current Path (with Eric Bogatin) - YouTube
PCB layout for switched-mode power supplies (SMPS) | TI.com Training Series
Another one, very good.
[LIVE] How to Achieve Proper Grounding - Rick Hartley - Expert Live Training (US) - YouTube
[LIVE] How to Achieve Proper Grounding - Rick Hartley - Expert Live Training (US) - YouTube
This is the answer i expected, thanks! So the voltage in point C is some where between 0 and 1V.
These are awesome, thanks!.
The DC current definitely isn't anything like a straight line, it spreads out rather like magnetic lines of force on a bar magnet (but constrained to 2 dimensions), spreading well beyond the points A and B, but dropping off as reciprocal square distance I think for a plane.
If the contact points at A and B are small, most of the resistance and voltage-drop is concentrated at A and B.
This is all the same as heat flow BTW, if you want some intuition - imagine a flame and an ice cube on points A and B of a metal sheet - how do you expect the temperature to distribute? Same equation.
At high frequency AC the behaviour is very different and the current will want to mirror the conductor above the ground-plane in order to minimize the inductance of the loop. (Or put another way current builds up faster in the routes of low inductance, so they dominate in AC)
Another question is if point A is power source Gnd, point B is regulator ground so what point should the load Gnd connect to for best noise performance? As in control theory, regulator Gnd should be connected to load Gnd as that's what it want to control but in that case, the Gnd is offet to 1V from the power source Gnd.