Why aren't ground planes more common? (pardon the pun...)

[Circlotron]

Whenever I do a board layout I pour a board-wide polygon that is connected to the GND net. This greatly simplifies the layout. And it guarantees very low resistance between points connected to GND. Also it means that less etchant is consumed by the board manufacturer. Maybe board yield numbers taking a slight hit because of the greater likelihood of etch faults courtesy of all the tracks being in close proximity to the polygon. Except for issues with capacitance with HF signals, why don't people more often use polygons like this?

[ataradov]

I see polygons everywhere. Do you have examples of well known boards that don't have ground pours?

[brabus]

Single sided boards (the real deal in mass electronics) may warp if one side has too much copper.

[james_s]

Is it also that copper is expensive stuff? Sure the board starts out completely covered and it gets etched away, but don't they recover copper from the spent etchant in large scale commercial operations? No reason to use a ground plane if the circuit does not require it. I do typically use them where I can on my own boards though.

[Feynman]

Are ground planes really uncommon? I don't think so

[S. Petrukhin]

I think solid poligon have a termal problems and capasite problem. I'm use grid poligons.

[filssavi]

Is it also that copper is expensive stuff? Sure the board starts out completely covered and it gets etched away, but don't they recover copper from the spent etchant in large scale commercial operations? No reason to use a ground plane if the circuit does not require it. I do typically use them where I can on my own boards though.

While they do recover copper from etchant it is actually to save money on the etchant, and it’s disposal, rather than the copper itself.

While copper might be expensive (mostly due to speculation actually) you also don’t use that much in a PCB, also the PCB fab houses will buy ready made cores that are already copper plated, thus the low purity copper scrap they get from the solution is of no use to them, they will just sell it for dimes on the dollar to a recycling company.

The truth is that ground planes are ubiquitous where it makes sense to have them, that is multilayer boards and simple digital two layers.

On complex 2 layer boards it is often impossible to have a ground plane as it would be too chopped up by traces  and components to be effective. On regular power supplies (that s low-ish frequency) it is often useless to have a ground plane, as it is much easier to control current paths with discrete traces, and you don’t have high speed signals anyway

[T3sl4co1l]

Low purity what, recovered copper is electrolytic grade.

Also, are you just assuming, or do you have data from operating fabs?

In either case, it's down in the noise for the cost of fabbed boards.  Unless you're doing millions, what's it really matter?

I regularly do 2-layer boards with dense layouts, this one for example: https://www.seventransistorlabs.com/Images/LimitingFusePcb.jpg (I don't have a 2D view handy; there are a few traces on the back, routed to carefully avoid cutting off the ground plane between top and bottom.)  Though I wouldn't say these sorts of layouts are "typical" in commercial use.

Back in the days of 1-layer boards, the next best thing was ground generally poured around everything, using jumpers to close loops only where absolutely needed.  Also, they made a lot of boards suspiciously like they were either saving etchant or maximizing conductivity (which may well be important with the typical 0.5oz foil?).  An aside, I never did understand how they made those; they were made in the CAD era, not freehand taped, but I've never seen an EDA that draws in negative space exclusively.  Possibly some Japanese exclusive.

(Not that 1-layer boards are gone, they're still quite common in power supplies.  These days, most products seem to be complicated enough to need multiple layers, or whole ass planes for signal quality alone.  Which is understandable, given the ever-higher density and flexibility of MCUs and whatnot.)

Tim

[filssavi]

I am assuming, and most definitely not operating my own fabs (well that seemed a bit obvious )

However I still don’t believe the  high purity argument, as you said as a cohost/saving is going to be down in the noise, thus I don’t expect them to invest in equipment capable of separating copper from all other crap that will get etched away from the acids.

Again  this is still pure speculation but my guess is they even large scale (consumer market level) are just reprocessing the process fluid to recover the acid, and just selling the residue (which will be mainly copper) to company in the business of refining metals, they will clean this raw copper waste from all other junk that goes with it resist, PCB material (while the etch rate for that is much  much smaller than copper it will still probably not be null) and given a large enough volume of liquid you will get some contamination in the output copper.
I am sure the final separation stage to actually get pure copper will not be rocket science, however still why would a PCB manufacturer, especially at large scale set up equipment and processes to do that when the regular metal smelters/recyclers are already set up to do that at a much larger scale (and thus much more quickly), when the amount of recovered value is so low

[rs20]

Whenever I do a board layout I pour a board-wide polygon that is connected to the GND net. This greatly simplifies the layout. And it guarantees very low resistance between points connected to GND.

EDIT: I may have been a bit strong with my retort here as pointed out by Tim. I'll keep the content below, but please interpret them as just 'points to consider' rather than an objection to the OP's post.

-- snip --

• If you properly design and lay out your ground traces (and then pour a big GND plane afterwards), then you're not simplifying the layout at all.
• If you just pour a big GND plane and assume that everything will get connected together, then there's every risk that your ground loops will be very circuitous, which is nothing like guaranteeing very low resistance between points connected to ground (or at least, the loop inductances will end up very bad, which is often just as problematic as resistance.)

Exceptions to the above objection may include: if you're working with something huge like DIP packages, in which case the GND can flow between pins, making the second claim slightly more feasible. Or if you're talking about having dedicated ground planes as in, e.g., a 4-layer board -- which is something I totally do for all the reasons you mentioned. But apart from those exceptions, treating GND as a lowest class citizen by just assuming the GND plane will stitch it together is IMHO bad practice.

[T3sl4co1l]

Well yes, given the brevity of that comment, it's no guarantee that a spray-and-pray strategy wasn't meant.  But it's also no guarantee against it.  An inner plane is perfectly consistent with the phrasing, as is properly stitched pour on signal layers.  Rather than jumping against such statements, I'd rather air on the optimistic side, and provide clarification if follow-up is requested.

Tim

[Circlotron]

Well yeah, I was initially thinking of mostly through hole boards. At work I often do some relatively simple and uncritical stuff that I have to build myself, and double sided through hole is plenty good enough. But even back in the day when this sort of stuff was mainstream, flooding the spare space with copper didn’t seem all that common to my recollection. Generally speaking, if there are no high currents or high voltage or other critical factors involved I just shoot for minimum total track length and that tips the odds in favour of having the most uninterrupted area for a copper pour. Do the same with the bottom layer and put some vias through to the other side where you have a dead end or an area connected only by a narrow isthmus. I’ve had great success doing that for many years.

[T3sl4co1l]

Also, if you go back far enough, you probably find it was outright harder to do (ain't no one got time for all that Rubylith!), or actively worse (tube circuits are quite high impedance, the added capacitance of a plane often being more detrimental than the coupling between traces, never mind magnetic induction).

Tim

[Psi]

For it to be classified as a power or ground plane then power or ground needs to be the only copper on that layer.

However, you always end up needing to route at least some signal tracks on that layer just to make things work or to stop the layout getting too messy. This turns the layer from a power/gnd plane into a signal layer (still mostly power/gnd, but no longer a plane).

If you absolutely must have a solid power/gnd plane you usually end up needing to pay more for extra PCB layers than you wouldn't have needed if you made it a signal layer with a few tracks on it.

And cost usually wins in the end

[coppice]

Ain't no one got time for all that Rubylith!

Who ever used Rubylith for a PCB?

[Psi]

Also it means that less etchant is consumed by the board manufacturer.

They don't care, I get front/rear case panels made from PCBs and the rear panel for one of my projects has zero copper on it. I'm sure they just etch it all off.

[james_s]

Also it means that less etchant is consumed by the board manufacturer.

They don't care, I get front/rear case panels made from PCBs and the rear panel for one of my projects has zero copper on it. I'm sure they just etch it all off.

Sure, they'll make whatever you ask for, for low volume stuff it doesn't matter. If you wanted 1,500,000 of them for a volume consumer device though it becomes important to optimize the pennies.

[PlainName]

Ain't no one got time for all that Rubylith!

Who ever used Rubylith for a PCB?

"Before Dot and Tape (see below) there was Rubylith. This consisted of a thin red film stuck to a mylar film. Generating a PCB layout was done in the negative, removing the red film where there was to be track. The technique required a completely different mindset to today’s “positive” approach. It was really good for earth planes, but I am not sure if double sided layout was possible and certainly a track between IC pins was impossible. I was only introduced to this at the start of my career in 1976 and never actually worked with it. I don't really recall how one created IC and transistor pads. I can only assume that they used some template placed on the Rubylith and exposed the area where the metal would be. The finished piece of Rubylith was the negative of the layout (and in 1:1 scale) and was placed over the photo resist and then exposed to light. There was no photographic step needed."

https://www.eetimes.com/how-it-was-pcb-layout-from-rubylith-to-dot-and-tape-to-cad/

[coppice]

Ain't no one got time for all that Rubylith!

Who ever used Rubylith for a PCB?

"Before Dot and Tape (see below) there was Rubylith. This consisted of a thin red film stuck to a mylar film. Generating a PCB layout was done in the negative, removing the red film where there was to be track. The technique required a completely different mindset to today’s “positive” approach. It was really good for earth planes, but I am not sure if double sided layout was possible and certainly a track between IC pins was impossible. I was only introduced to this at the start of my career in 1976 and never actually worked with it. I don't really recall how one created IC and transistor pads. I can only assume that they used some template placed on the Rubylith and exposed the area where the metal would be. The finished piece of Rubylith was the negative of the layout (and in 1:1 scale) and was placed over the photo resist and then exposed to light. There was no photographic step needed."

https://www.eetimes.com/how-it-was-pcb-layout-from-rubylith-to-dot-and-tape-to-cad/

I know what rubylith is. I've used it. I've never seen it used for PCBs, though. Only for laying out ICs. Dot and tape on mylar film for PCBs goes back to the 1960s, so rubylith for PCBs must have disappeared very early on.

[james_s]

Maybe different techniques were common in different countries or regions? Most of that is before my time.