Utility of copper pours on four-layer boards

and

the

although

and

Ok, in the field I was thinking of they don't use FR4 and cost isn't really an issue. Much of it is financed through the generous donations we all put in the offering plate by April-15.

Ah, OK.

Absolutely, but appears that you employ a design process where, by the time you get to what's supposed to be the "real" board, the chance is pretty small. Granted, with analog design usually you don't have a lot of choice in the matter (I can't imagine there are many analog designers out there who would capture a schematic of a tricky analog circuit they've never done before and expect that without crunching some numbers, simulation or prototyping it would have a near-100% chance of working on the first go), but I've seen plenty of digital guys who are more than happy to draw up schematics even when there are somewhat tricky portions of, e.g., an interface where they haven't yet fully thought through how it's all going to work.

I remember one board that had a BGA DSP on it where we were in "must got PCB made ASAP" mode, and I had mentioned that since we hadn't fully flehsed out the DSPmicrocontroller interfaces it would behoove us to take every single BGA pad and dog-bone via it in case we needed to get to some of the (currently unused) pads. This suggestion was dismissed in the name of saving a bit of time and of course it wasn't two weeks after getting the PCB back that it turned out we did need access to some additional pads and the board had to be re-spun. On the second go-around they were gun shy enough that they switched to a TQFP part for the DSP, though (and then finally switched back to BGA for the production PCBs).

Yeah, although he really didn't seem to mind all that much. :-)

---Joel

Thanks for the advice, Terry. I'll avoid a detailed reply as you wrote quite well and I'm generally very much in agreement with it.

---Joel

Do you typically remove the soldermask as well (assuming there isn't a safety issue with the voltages present)?

---Joel

I dunno ... see Episode 2 at

As explained to me 10 years ago, the outer layers are etched, then plated up to desired copper weight. Thus, having copper pour on the outer layers requires more plating.

I wouldn't say pouring the outer layers with ground is the norm. I'll do it only if there's a good reason to. If you're running high-speed signals around an FPGA or other HS parts, then a ground pour on the FPGA mount side may help.

Mark

Actually I often do have to go to layout without sims. I breadboard the most critical stuff but the remainder goes straight into CAD.

Ouch! The switch to TQFP was a smart move, I'd have stayed with that. Especially if this has to go into RoHS countries.

Oh, he did mind.

That increases the chance of attracting dirt, grease, finger prints and all that. Can become ugly, to the point where a service guy doesn't want to touch it anymore. One guy told me that some patient interfaces can look like evidence pieces from a crime scene.

Hi Joerg,

OK; I still expect you're noticeably above average when it comes to not making mistakes.

The final product didn't have the space for it, otherwise we (probably) would have.

Hmm... yeah, I suppose he did; it would seem to hinge on how much he knew of what was going to happen next, perhaps.

---Joel

Hmm, doesn't that only apply to thicker copper such as 2oz?

You'll probably see somewhat of a renaissance of copper pour coming, on account of all those chips that require a heat sinking pad.

Older, mostly :-)

But yes, with age comes experience from the numerous bloody noses one got in the past. Such as to avoid LDOs like the plague. Just one of those can cause a lot of grief on a design that was only simulated and not breadboarded.

I try to avoid BGA wherever I can.

He knew all of that and had told the others. Yet many of them still didn't get it. Thomas was by far not the only one in doubt.

I have had 4- and 10-ounce PCBs made. with these the mfg started with

I have also done a lot of 1-4-4-1 Oz 4-layer PCBs, again starting with a

I guess it all depends on who makes your board, what foils they stock etc.

Cheers Terry

No, I dont.

not for this reason (although its a good one).

solder mask stops evil bridging. Having been not so much bitten as having had several limbs torn off by this once in the past, I leave the solder mask alone unless absolutely necessary - eg if using zebra strip LCD connectors dont have solder mask anywhere near them, its thick enough to interfere with the contacts.

Cheers Terry

Understood. What I meant was on large, contiguous copper pours that are acting as heat sinks (e.g., the 1+ sq. in... er, 645+ mm^2 that you have poured around the source terminal of a D-PAK to help it get the heat out).

---Joel

Wow! Some day archeologists will wonder why New Zealanders copper-plated everything. I can't remember any designs I did where we used over 2oz.

Yep. Same for cheap coin cell holders where a gold-plated pad on the board is one of the contacts. The spec sheets often don't talk about solder mask clearances. Can be a nasty surprise for those who didn't know.

Don't overdo it with heat sinking into FR4. If the board gets noticeably hot with soldermask chances are that you'll see some delamination in the field. It won't become much better without the solder mask.

And something has to protect the copper or it'll look like the roof of a Gothic cathedral some day ;-)

Hi Joerg,

Quantifying "how much better" is what I'm trying to get at here; apparently the "how much" is "not very." I was thinking it was potentially worthwhile based on the thermal resistance of solder resist being something like 6x that of air, although I realize the thickness of the solder mask is pretty small.

I've got plenty of milled boards around that are already heading that way. (On the other hand, some decades back my mother cut a large "Loch Ness monster" out of some rather thick copper and mounted it on a similarly cut-out sheet of plywood to hang on the wall as decoration. She claimed that eventually it'd turn nice and green -- sea monster green! --, but now ~25 years later it's just a very tarnished brown. :-) Back at University there was one building with a copper roof that had turned a nice green though, so I guess it is just a matter of waiting.)

Hmm... I have a suspicion we won't be seeing many new buildings with copper roofs any time soon!

---Joel

thanks Joerg, I'll add that to my list of "dont do this"

Cheers Terry