𝗠𝘂𝗹𝘁𝗶𝗹𝗮𝘆𝗲𝗿 𝗯𝗼𝗮𝗿𝗱𝘀: 𝗣𝗿𝗼𝗼𝗳 𝗼𝗳 𝗲𝘅𝗽𝗲𝗿𝘁𝗶𝘀e, 𝗼𝗿 𝘄𝗮𝘀𝘁𝗲 𝗼𝗳 𝗺𝗼𝗻𝗲𝘆?
Multi-layer boards are VERY cool. But there is a reason we try to stick to 2 layers when possible: COST.

A high layer count comes about for one of 4 reasons:
• Component density neccesitates it.
• You need to get it done in a hurry and to hell with cost
• You lack the experience to get it done with less
• You think more layers are a badge of honour

The moment you pay for it yourself, your perspective changes. A given board design may look like it needs 4 layers, but after staring at it for a while, you start to wonder: Is it perhaps possible to make this work on only 2 layers? What if all it took was careful and focussed effort? Will the ground net be a bit fragmented, yes definitely. But can I make it work well enough despite this? Perhaps...

𝗕𝘂𝘁 𝘁𝗵𝗲 𝗿𝗲𝗮𝗹 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻, 𝗶𝘀 𝘄𝗵𝘆?
Because saving money here is basically free money. Let's say I put in an additional 24 hours of effort to drop two layers and save $2 per unit in the process. If I build 5000 of those boards over the course of a year or two, that is $10k saved. Am I happy to be paid $10k for 24 hours of effort? I think so yes!!

𝗛𝗼𝘄 𝗱𝗼 𝗜 𝗱𝗼 𝗶𝘁? Experience and focussed intent.

𝗠𝘆 𝗳𝗶𝗿𝘀𝘁 𝘄𝗲𝗮𝗽𝗼𝗻 is smaller components. It is amazing how much space you can save by going to 0402 instead of 0603 sized passives. But this is sometimes counter productive. Sometimes you REALLY need to pass a track in between the pads of a passive, and the 0402 makes that difficult, so here and there you stick to 0603 (where it is doable), because an 0603 is smaller than an 0402 and 2 vias.

𝗠𝘆 𝘀𝗲𝗰𝗼𝗻𝗱 𝘄𝗲𝗮𝗽𝗼𝗻 is to lean into the PCB manufacturer's capabilities. I don't like to be on the bleeding edge of what they can do, but if I can reduce track width and spacing, and reduce via diameter safely, I do so.

𝗠𝘆 𝘁𝗵𝗶𝗿𝗱 𝘄𝗲𝗮𝗽𝗼𝗻 is to be fanatical about GND net connectivity. This is the part that makes a complex design originally intended for 4 layers, actually work well on 2 layers. GND must reach everywhere and do so in a solid way. Copper pours and sometimes thousands of stiching vias serve as my makeshift GNDplane.

There are other tricks as well, but you get the point. It is important to note that this has to be validated. Just because it is working on the bench is no guarantee of production-ready stability, especially if there are high speed requirements.

Sometimes the only thing that stands between you and a significant cost saving is a bit of focussed intent.