Pin pitch is ultimately the spacing between traces. The traces are not as big of an issue as the actual spaces between the traces. This clearance is where things get tricky with making printed circuit boards. The process of masking off some circuit is not that hard. The way the stuff you want to keep is isolated from the copper you want to remove is the hard part. One of the issues is that you need an acid to take away the copper, but not the mask, but copper has a thickness. As the copper is etched away the acid moves sideways into the thickness too. Copper never etches completely uniformly either. The larger areas of open copper that need to be removed will etch much faster than a bunch of thinly spaced gaps. One of the tricks to design is finding ways to etch consistently with the process you build.
If you want to make super tiny traces that still have the right amount of copper and have all the gaps etched away consistently, the process of the etching toolchain becomes more expensive. You will need a stronger acid with a very good way of removing the etchant that is close to the copper and loaded with copper already. This is usually done with a stream of small bubbles, but it is risky because it could impact the adhesion of the masking material over the traces you want to keep. The stronger, hotter, and now agitated acid requires that the copper clad board is extremely clean and the photoresist used to mask the stuff you want to keep must be a very high quality. Also the resolution of this photoresist requites a much more precise form of UV exposure and development (about like developing old film photos).
So you need a better mask development toolchain, better quality photoresist. You might get away with not using photoresist at all in some other cheaper low end processes. You need the highest quality copper clad that etches more evenly, and you need a stronger acid to etch quicker straight down because a slower acid will move further sideways and ruin the thin traces to keep.
The pic has old school dip chips in a static resistant foam. Those are the classic standard 1/8th inch (2.54mm) pin pitch. The easiest types of boards to make yourself are like the island soldering style board with the blue candy soldered on. That is a simple coalpits oscillator for testing crystals. Then there are protoboards like the homemade Arduino Uno pictured. Then you get into the etched boards. Some of these were done with a laser printer toner transfer method. That is like the least accurate DIY and somewhat analogous with the cheapest boards from a board house. Others were made using photoresist. This method is more accurate but involved and time consuming. One of the boards pictured is a little CH340 USB to serial board with a USB micro connector. That is getting close to my limits for etching easily. Another board has a little LCD and text. There is a small surface mounted chip pictured on the foam and that is a typical example of what kinds of pin pitches are common for the cheapest level of board production. Now there are two USB-C female connectors pictured. One has a larger pin pitch and is made for USB 2.0 connections and power. However, that other one with all those tiny tiny connections at the back – that is a full USB-C connector. That thing is a nightmare for tiny pin pitch. There is also a USB-C male connector with a little PCB attached. These are the types of solutions people have tried to come up with where only some small board is actually of a much higher resolution. It is not the best example but I’m not digging further through stuff to find better.
The actual pins on the little full USB-C connector are inverted to be able to flip the connector. There is a scheme present to make this a bit easier to match up the connections but it is still a pain in the ass to juggle everything around. All of the data trace pairs are differential too, which basically means they must be the same length between the source and destination. So any time they are not equal, the shorter trace must zigzag around in magic space you need to find just to make them even.
Pin pitch is ultimately the spacing between traces. The traces are not as big of an issue as the actual spaces between the traces. This clearance is where things get tricky with making printed circuit boards. The process of masking off some circuit is not that hard. The way the stuff you want to keep is isolated from the copper you want to remove is the hard part. One of the issues is that you need an acid to take away the copper, but not the mask, but copper has a thickness. As the copper is etched away the acid moves sideways into the thickness too. Copper never etches completely uniformly either. The larger areas of open copper that need to be removed will etch much faster than a bunch of thinly spaced gaps. One of the tricks to design is finding ways to etch consistently with the process you build.
If you want to make super tiny traces that still have the right amount of copper and have all the gaps etched away consistently, the process of the etching toolchain becomes more expensive. You will need a stronger acid with a very good way of removing the etchant that is close to the copper and loaded with copper already. This is usually done with a stream of small bubbles, but it is risky because it could impact the adhesion of the masking material over the traces you want to keep. The stronger, hotter, and now agitated acid requires that the copper clad board is extremely clean and the photoresist used to mask the stuff you want to keep must be a very high quality. Also the resolution of this photoresist requites a much more precise form of UV exposure and development (about like developing old film photos).
So you need a better mask development toolchain, better quality photoresist. You might get away with not using photoresist at all in some other cheaper low end processes. You need the highest quality copper clad that etches more evenly, and you need a stronger acid to etch quicker straight down because a slower acid will move further sideways and ruin the thin traces to keep.
The pic has old school dip chips in a static resistant foam. Those are the classic standard 1/8th inch (2.54mm) pin pitch. The easiest types of boards to make yourself are like the island soldering style board with the blue candy soldered on. That is a simple coalpits oscillator for testing crystals. Then there are protoboards like the homemade Arduino Uno pictured. Then you get into the etched boards. Some of these were done with a laser printer toner transfer method. That is like the least accurate DIY and somewhat analogous with the cheapest boards from a board house. Others were made using photoresist. This method is more accurate but involved and time consuming. One of the boards pictured is a little CH340 USB to serial board with a USB micro connector. That is getting close to my limits for etching easily. Another board has a little LCD and text. There is a small surface mounted chip pictured on the foam and that is a typical example of what kinds of pin pitches are common for the cheapest level of board production. Now there are two USB-C female connectors pictured. One has a larger pin pitch and is made for USB 2.0 connections and power. However, that other one with all those tiny tiny connections at the back – that is a full USB-C connector. That thing is a nightmare for tiny pin pitch. There is also a USB-C male connector with a little PCB attached. These are the types of solutions people have tried to come up with where only some small board is actually of a much higher resolution. It is not the best example but I’m not digging further through stuff to find better.
The actual pins on the little full USB-C connector are inverted to be able to flip the connector. There is a scheme present to make this a bit easier to match up the connections but it is still a pain in the ass to juggle everything around. All of the data trace pairs are differential too, which basically means they must be the same length between the source and destination. So any time they are not equal, the shorter trace must zigzag around in magic space you need to find just to make them even.