Connector selection and footprint design require careful DfM consideration, as tolerances, spacing and assembly constraints directly impact manufacturability and performance.

When it comes to interconnects for printed circuit boards, I found one vendor with 25,217 options. A specialty vendor offered only about 9,000 different connector SKUs. Luckily, connector selection can typically be narrowed by pin count, pitch and other key parameters.

To be honest, the parametric part numbers were what bothered me about capturing connector footprints. A few letters designated the family, followed by a series of numbers that fleshed out the characteristics of each connector. Imagine the old days, thumbing through a catalog to find the connector style, then parsing through other pages to decode the numerous details of all the options. There was one drawing, and it covered about 300 different connector configurations for the whole family of connectors.

Figure 1. Narrowing our search to PCB connectors returns 25,217 results from a single vendor. These 10 options fan out to almost limitless configurations. (Source: TE Connectivity)

The good news is that there will likely be a table that provides the overall size values of the connector body based on the number of pins by rows and columns. The length of any through-hole pins is another option, depending on the PCB thickness. Larger connectors feature alignment pins or holes for mounting hardware. Retention clips are common with higher pin-count connectors.

The plating type is usually immaterial to the footprint geometry but is often incorporated somewhere in the vendor part number. Alignment pins are one area where a connector vendor will push the limits of what can be constructed on the PCB. Positional and size tolerances can be too tight.

Other miscellaneous issues could also crop up. The gap between a non-plated hole and an SMD pin can be too small for the PCB fabricator. Beware of a board thickness callout that is non-manufacturable. Component engineers earn their money when they must qualify a connector scheme. The board designer is a backstop for the DfM requirements. The board is only as good as its design. How it interfaces with the world depends on sound engineering judgment.

Figure 2. Filtering the search to find board-to-board receptacles with 60 positions using two rows on 0.1" pitch, plated holes with vertical launch narrows the field from 13,903 possibilities down to twelve. There are 24 further fields to specify just about any requirement missed by the first few attributes. If you can’t find a connector among these, outfits will create connectors to order. (Source: TE Connectivity)

Just like any component vendor, the connector industry wants to give its product the best odds of success. That innate desire not to be a spectacular failure is a strong incentive to over-constrain the design. This is where tolerances are flagged during the DfM cycle at the fabricator. This is not the stage to discover required compromises. A call ahead of time can save the endgame.

Figure 3. The title page of a typical connector drawing includes a decoder for the part number and the standard views. The other three pages have cross-section cutaway views and various details, but no actual PCB footprint recommendations. We do have a pin-one indicator, so that’s a good start. (Source: Samtec)

Now for the fine print: The upper-left corner of the title block is reserved to indicate the tolerance allowance based on the number of digits to the right of the decimal point for each dimension. The dimension with the tightest tolerance would be any three-digit dimension where the maximum deviation from the print is +/-0.050mm (2 mils). Most dimensions use two digits past the decimal point, giving a more reasonable tolerance of +/-0.13mm (5 mils).

These connector dimensions sometimes creep over to PCB outline dimensions and, of course, board thickness. A check of the component footprint should include considering the DfX guidelines the vendor chain would require regarding the dimensions and tolerance stackups of the single board.

Given that connectors are often on the edge, the spacing requirements for engagement and disengagement apply. Compared to some connectors, our fingers look like they are on the hand of The Incredible Hulk. For some of those ham-fisted Army sergeants, that may be true, so we must always be aware of the airspace around a connector.

Everything Must Be Connectorized, Sometimes

I’m thinking back to the mistake I made on a program where these little modules were connected inside a housing using SMA connectors and semi-rigid coax. It was the RF amplifier for a portable radar system. It was one thing to create a drawing of all the sub-units put together. It was quite another to explain how we would get a wrench into each location to tighten the SMA cables.

A few of the modules had to be connected using semi-rigid cables before we could bolt them down inside the modular housing for the amplifier. Luckily, a semi-rigid cable can be bent a little without too much trouble. My manager and I had to sit down to work out a convoluted assembly process. From that debacle onward, I paid close attention to the entire life cycle when planning the connector scheme.

Figure 4. I count 16 connectors on the perimeter of this main logic board for a Chromebook. It's about a 50:50 mix of flex circuits and wired connectors. The USB type-C connector has through-hole pins. That and the SD card holder are the only user interfaces. (Source: Author)

Sometimes, it’s better to make the wire too long and include a service loop with the wire breakouts. The point here is to escape the two-dimensional mindset of our PCB design philosophy and to consider the process from fabrication through SMD, second op, and final test right from the beginning. Space must be provided for the hardware to meld into a producible product and come apart again if necessary.

When space is limited, the flex circuits thrive. They are an ecosystem of their own. Their specialty is the ability to conform to contours with minimal material thickness. They are used extensively in phones, extended reality headsets and laptops, to name a few. As an aside, the only time I’ve encountered a rigid-flex PCBA is in aerospace, when a loose connector could turn a satellite in orbit into space junk. We still use connectors for automotive applications, but they come with derated wiring and over-built clamps to keep them safe and secure.

Taking Interconnections to the Next Level

Data centers are the place where the leading edge of connector design plays out. Fiber optics are a key data pathway in data centers and are increasingly extending into enterprise IT environments. Those fiber optic connectors used to be limited to the edge of the board using transceivers in little metal cages. Connecting the fiber is one of those high-precision affairs where missing the lens by just a mite can cause signal-integrity issues.

A lot of copper is in between the chip and board edge, so mid-board optical connections have become more common. Board-to-board connections are supplemented with board-to-chip and chip-to-chip topology, along with the intersystem connections of the past. Some of those other attributes come into play here. We select for high data rates, better flammability ratings, and, of course, lots of pins.

John Burkhert, Jr. is a principle PCB designer in retirement. For the past several years, he has been sharing what he has learned for the sake of helping fresh and ambitious PCB designers. The knowledge is passed along through stories and lessons learned from three decades of design, including the most basic one-layer board up to the high-reliability rigid-flex HDI designs for aerospace and military applications. His well-earned free time is spent on a bike, or with a mic doing a karaoke jam.