If you compare a modern PCB with a typical 1980s PCB, you might notice — like [lcamtuf] did — that newer boards tend to have large areas of copper known as pours instead of empty space between traces. If you’ve ever wondered why this is, [lcamtuf] explains.
The answer isn’t as simple as you might think. In some cases, it is just because the designer is either copying the style of a different board or the design software makes it easy to do. However, the reason it caught on in the first place is a combination of high-speed circuitry and FCC RF emissions standards. But why do pours help with unintentional emissions and high-speed signals?
The answer lies in the inductance the pours add to the boards. Of course, there’s no free lunch. Adding inductance in this way also increases capacitance, which can be a bad thing.
The truth is, most of the boards we deal with would be fine with or without the pours. That’s a good thing, too, because the post illustrates how some common things can significantly reduce the effectiveness of the copper pours.
When we don’t send our boards out, we are usually more interested in removing copper. You also have to be careful when you want your PCB to radiate.
I mostly use pours to conserve the etching solution.
That was my first thought, the less copper you remove the longer it will last.
And that applies to large board houses where the etchant is a “hazardous waste” that has to be disposed of.
Not to mention in the past you might have to manually draw that pour with ink by hand!
Fine if you have a final circuit. Less so if you might want to use that space for a component in a future revision.
Well, with modern production methods it is more or less free RF shielding, so if it does no harm, why not.
It also reduces the interaction between circuits that share a common ground return path be reducing their shared resistance and inductance. At low frequencies the common ground return resistance dominates while at higher frequencies it’s the shared ground inductance that can cause problems.
Generally by using pours to generate the circuit grounds you can significantly reduce jitter and noise in both analog and digital systems.
Another factor is that with the move from wave-soldering of TH parts to reflow of SMDs, there is less risk of the thermal shock warping the board. Less copper to etch off also reduces cost in high-violume manufacture. And of course back in the days before CAD, pours were just hard to create with taped layouts.
On internal layers, pours are essential because it prevents differential expansion making the board twist. Even like 2 reflow cycles is bad enough to make a significant bend. Biggest mistake in my first 8 layer board. Always balance your copper, folks.
Beat me to it with “copper balance”
sample pics for others reading this: https://www.multi-circuit-boards.eu/en/pcb-design-aid/copper-balance.html
board sizes > 20 cm and thicker copper layers drastically magnify the imbalance problem. I’ve had a hastily designed 4L power PCB that showed curvature in both directions and made suspicious crackling noises when lightly flexing.
I’m under the impression that even local reworking made it bend even more out of shape.
Read it again.
Adding copper pours reduces inductance in power connections.
Adding copper pours increases the capacitance of power connections.
Unmentioned in the article is that copper pours help provide a defined impedance when designing transmission lines for high speed signals on PCBs.
nice Amstrad CPC board on top!
I have also wondered if ’80s boards tended to have thicker copper, which would help power traces, but so far have been unable to confirm this.
I remember leaving as much copper as possible for better analog/digital ground planes
“Increased capacitance” is a good thing. Lower inductance and increased capacitance is exactly what you need to decouple power supplies.
Modern jellybean components are very fast, e.g. 74lvg1g* driving 2.5V into 50ohms with <300ps risetime. That requires excellent PSU decoupling, and Vcc+Gnd pours are very useful at removing the UHF/microwave frequencies where capacitors are inductive (i.e. Z rises with frequency).
You might, however, have to take into account the “cavity” resonance of solid Vcc+Gnd pours.
TL;DR is that power distribution networks (i.e. Vcc, Gnd) is a non-trivial topice with modern components.
I think the right answer here is “it depends”. Ground pour between high speed traces can be bad thing if they are not properly stitched. It’s better to use distance instead and no ground pour if the ground pour between high speed signals can not be properly via-stitched to ground. As previously mentioned, it can be beneficial as RF shielding, low inductance decoupling, and for balancing PCBs, just not as a blanket statement “always good”. I think both Zach Peterson and Rick Hartley have mentioned this in various videos online.